> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think; a language model produces the former, not the latter.
It's been incredibly blackpilling seeing how many intelligent professionals and academics don't understand this, especially in education and academia.
They see work as the mere production of output, without ever thinking about how that work builds knowledge and skills and experience.
Students who know least of all and don't understand the purpose of writing or problem solving or the limitations of LLMs are currently wasting years of their lives letting LLMs pull them along as they cheat themselves out of an education, sometimes spending hundreds of thousands of dollars to let their brains atrophy only to get a piece of paper and face the real world where problems get massively more open-ended and LLMs massively decline in meeting the required quality of problem solving.
Anyone who actually struggles to solve problems and learn themselves is going to have massive advantages in the long term.
I’ve come across this analogy that I think works well:
Using an LLM to do schoolwork is like taking a forklift to the gym.
If all we were interested in was moving the weights around, you’d be right to use a tool to help you. But we’re doing this work for the effect it will have on you. The reason a teacher asks you a question is not because they don’t know the answer.
If students went to college only to learn, colleges wouldn't bother giving diplomas.
Compare: My piano teacher doesn't give diplomas because none of her students would care, her students actually want to learn. When my piano teacher cancels class, I am disappointed because I wanted to learn. My piano teacher doesn't need to threaten me with bad grades to get me to practice outside of class (analogous to homework), because I actually want to learn.
There are many college students for whom none of these tests would pass. They would not attend if there was no diploma, they're relieved when their professors cancel class, and they need to be bullied into studying outside of class.
What made us think these students were ever interested in learning in the first place? Instead, it seems more likely that they just want a degree because they believe that a degree will give them an advantage in the job market. Many people will never use the information that they supposedly learn in college, and they're aware of this when they enroll.
Personally, the fact that they can now get a degree with even less wasted effort than before doesn't bother me one bit. People who want to learn still have every opportunity to.
> It has value because a student can only get it by learning and problem-solving.
No, it has value because it gatekeeps class mobility. Degrees haven't signified learning or problem solving longer than I've been alive. The attitude is that if you pay for an education, you're entitled to a degree. The education aspect is optional.
Further, if a student can get a diploma without work, then the diploma does not have value anymore. If diplomas are no longer valuable, the signal they provide in the labor market will turn into noise.
If employers no longer look for the diploma-signal in an employee, what will be the reason an employer will hire an employee?
I think this story will become true, and society will radically shift into one where critical thinking skills will actually be the only skills employers look for in employees, since the grunt work can be automated.
What becomes the signal then? Will we shift back into apprenticeship based employment? How do potential laborers display their critical thinking skills apart from displaying them in person?
In a medieval guild, to be admitted as a master, an apprentice had to create a chef d'oevre, or masterpiece, so called for this reason.
In the computer engineering industry, you increasingly have to demonstrate the same: either as a part of your prior work for hire, or a side project, or a contribution to something open-source.
A diploma is still a useful signal, but not sufficient, except maybe for very junior positions straight from college. These are exactly the positions most under pressure by the automation.
I think software developers might be somewhat of an outlier. Industry wants good programmers but universities teach computer science which really should be called "computation science". Much of what we learn in university will hardly ever be used while many practical skills are at best learned as a side effect. Dijkstra favorite said that computer science is as much about computers as astronomy is about telescopes.
So degrees have been a weak signal for a long time. Several of the best developers I've worked with had no CS degree at all. As a result we have interview processes that are baffling to people from other industries. Imagine a surgeon having to do interview surgery, or an accountant having to solve accounting puzzles. AFAIK we are very unusual in this regard and I think it's because degrees are such a weak indicator and the same is true for certificates in our industry.
What becomes the signal then? Will we shift back into apprenticeship based employment? How do potential laborers display their critical thinking skills apart from displaying them in person?
This is already true to some extend. Not apprenticeship taking place of college, but the last couple of places I worked hiring generally happened based on: I already know this person from open source projects/working with them in a company/etc.
In certain companies, degrees were already unimportant even before LLMs because they generally do not provide a very good signal.
I might be a good thing. Colleges have become complacent and too expensive. Costs of an education have been increasing while employment opportunities decreasing for some degree categories. People have been sounding alarms for a while and colleges have not been listening. The student loan market is booming.
Now if students can shortcut the education process, they can spend less time in it and this may force colleges to reinvent themselves and actually rethink what education looks like in the new era.
Honestly, a future where diplomas are just noise and employers stop caring about them and thus young people stop wasting years of their lives "learning" something they don't care about sounds like a huge improvement.
LLM cheaters might incidentally be doing society a service.
They will only learn what's needed to "get the job done" for whatever it means at that moment, and we could potentially see more erosion in technical abilities and work quality. You don't know what you don't know, and without learning things that you don't care about, you loose the chance to expand your knowledge outside of your comfort zone.
> They will only learn what's needed to "get the job done" for whatever it means at that moment
I graduated university around the turn of the century, long before the current AI boom started, and the majority of my classmates were like that. Learning the bare minimum to escape a class isn't new especially if you're only taking that class because you have to because every adult in your life drilled into you that you'll be a homeless failure if you don't go to college and get a degree. The LLMs make that easier, but the university, if the goal wasn't just to take your tuition dollars to enrich a vast administrator class instead of cover the costs of employing the professors teaching you, could offset that with more rigorous testing or oral exams at the end of the class.
The real lesson I learned during my time in university is that the two real edges that elite universities give you (as a student) are 1) social connections to the children of the rich and leaders in the field that you can mine for recommendations and 2) a "wow" factor on your resume. You can't really get the first at a state school or community college, and you definitely can't get the second at a state school or community college, despite learning similar if not the same material in a given field of study.
It hasn't been about (just) the learning for a long time.
I don't think diplomas have mattered for decades, at least in tech. Let's not pretend anything improved with the introduction of chatbots.
Annyway, any advantage is entirely offset by having to live in a world with LLMs. I'd prefer the tradition of having to educate retarded college graduates. At least they grow into retarded adults. What are we gonna do about chatbots? You can't even educate them, let alone pinocchio them.
> If students find a way to get a diploma without doing the work, it will soon be worth less than the paper on which it is printed.
Amen.
I look forward to the era where we train professionals the old fashion way: apprenticeships. It sure worked for blacksmiths and artisans for hundreds of years.
That is how many professionals are training in many European countries, and they still want a diploma to go with it.
In many countries, regardless of how learning it was achieved, you still need a paper to prove that you actually did it.
And in countries like Germany, better keep all those job evaluations close at heart because they get asked for as part of many job interview processes, additionally have them reviewed by lawyers, as they legally can't say anything negative, there is an hidden language on how to express negativity which to the reader sound positive on first read.
Techbically even those hidden phrases arent permitted by law, and you have the right to get those phrases removed/have it be turned into a basic evaluation (just from when to when + what job)
The employers that do use those hidden phrases just hope they arent challenged/the employee doesnt notice.
Thats also why most evaluations are entirely written in the superlative.
Unfortunely they are common enough to keep lawyers and book publishing businesses going with dictionaries "evaluation => real meaning", exactly for that little help to go though it and validate its correctness.
Many employeers profit from foreigners that aren't well versed in these nuances, and have to be educated this is a thing.
An example for others not used to German work market,
"Missing" information is a huge red flag as well, though. If an Arbeitszeugnis only lists the dates and what the person worked on, and nothing about results or behavior, we would never invite them for an interview.
PhDs are apprenticeships. Students perform original research under the supervision of a doctoral advisor, and submit their work to a "guild" of external reviewers (peer reviewed journals or conferences). Once the student has three peer-reviewed publications, they graduate.
"Once the student has three peer-reviewed publications, they graduate."
That's not a standard at all. You usually can't graduate without at least one peer-reviewed publication, but beyond that, as far as number of publications goes, it varies a lot from institution to institution. The biggest standard is that you complete a dissertation and defend it.
An apprentice, after five years of work, can be a master, just by practcising.
That's why the universities of Oxford and Cambridge give Master's degree to everyone that gets a Bachelor's degree after five years, without further examination or coursework (note that these are MAs only, not MRes, MPhil or MBA degrees, which typically require 1-2 years of studies, exams and theses).
Historically, the academic Master was seen as equivalent to a Master in a craft (e.g. philosophy <=> carpentry).
> It has value because a student can only get it by learning and problem-solving.
No. That's how it should be, but the reality looks different: it has value because it shows that someone spent 3+ years doing what they were told to do, enduring all the absurdities they were subjected to in the meantime. Whatever means they used to cheat don't matter, since they still worked on what someone told them to and produced results satisfying the expectations.
There are, perhaps, institutions where learning and problem-solving are seen as the most important while "following orders" and "staying in line" are deemphasized. For the students of all the others, putting up with an utterly absurd environment is often one of the biggest barriers to learning. Yet, it's a requirement without fulfilling which you can forget about graduating. Hence my conclusion: the diploma from most learning institutions certifies you as a good corporate drone - and that's enough of a signal in many situations, so why bother trying to fix it?
This is a very bold statement, you should consider substantiating it at least a little, maybe give just one example on what those absurdities collee students have to endure are, that prove they're good corporate drones?
From what I remember it was 4 years of learning stuff I signed up to learn, occasionally being quizzed on said stuff, and then they gave me a paper that claims I know the stuff.
I hire people now and where they went to school means little to me. The first priority is “can they do the work?” which is a niche programming. After that is established, I barely take note of school.
I don’t personally count a CS degree as an indication the person is a good programmer, or thinks logically, or has good work ethic.
Problem is you most likely don’t have that big operation.
For company I work for we hire 1 dev per year and I believe last year we did not even hire a single person.
So I do have time to check up the candidate do 3 rounds each 1 hr so that candidates can see what company are we and what person is the candidate. We also are small company so we don’t get that many applicants anyway.
I cannot imagine how it goes when someone needs to hire 20 devs in one quarter. Especially for a company that is any way known and they can get 1000CVs for a single position. They need to filter somehow.
HR seems a lot like real estate appraisal as~in appraisers know nothing about construction.
One could in theory establish the value of a thing or person by comparing it with similar things but if everyone does that the process becomes senseless.
If a working car is worth 5000, the same car from the same year with a defect that costs 1000 to fix should be worth 4000. If the repair costs 6000 there is no car.
> It has value because a student can only get it by learning and problem-solving.
No. It has value, because companies value it. It sets the starting point for your first salary and then for every salary negotiation moving forward.
As someone who did not go to university, but has the same knowledge self-taught I can tell you that this piece of paper would have opened so many doors and made life so much easier. It took me 15 years to get a salary, that people get with the piece of paper after graduating. And not because it took me 15 years to reach that level of knowledge. I had that by the time the other graduated.
I had a friend who always cheated in school, and now he works for a big car company and earned a fuck ton of money.
Life is unfair and companies only care about the paper your diploma is printed on. If students would ask me for advice, I would tell them to cheat whenever possible.
You forgot one big reward with your approach. You are not 200k in debt which will take that kid with diploma to get that piece of paper. You are debt free.
I am not arguing on the merit of having a diploma is a bad thing - the colleges these days have turned their backs on the people as well.
Average $100k to get a piece of paper (LLMs are not the issue here the useless degrees that the colleges offer). Invest that $100k at an average 15% return - it becomes a lot of money 25 years from now.
Or get a piece of paper (if your major is useless) and pay the banks 100k + interest.
That’s a very American point of view though. In many European countries people don’t build up such big debts. I’m from NL and the upper bound among my friends was something like 50,000, but most people had much lower debts between 0 and 20,000. (Yay for lower enrollment fees + part of the money is given as a gift).
And even if you ramped up a 50k debt, it’s a government loan where the interest rates are low, your monthly pay off is based on your salary, and if you are not able to pay the debt within a certain period, it’s wiped away.
In reality, holding a piece of paper has nothing in common with relevant ability. Some graduates will exit university as competent people, some won't.
This isn't exactly news, or even a recent phenomenon. My peers who finished their CS degree circa 2004 ish were a broad mix of utter idiots who shouldn't get a job in that field, or damn sharp (if green) technically-minded people, or somewhere in between.
The industry will sort them out; the ones that can, do, the ones that can't, will do something else of use, or find another job.
There is a correlation, that is all you need for it to be valuable. Remove the correlation and value isn't there any more, see US masters in computer science for example that one doesn't correlate with skill so it isn't valued, but the bachelor is. In countries where a masters is correlated with skill they are values though.
I really think that they are used because it is easiest and cheapest fully legal filter. When you have significant number of candidates any filter that comes first to mind is selected. Degree is one such one that won't ever legally bite you.
And that won't be their problem, they'll be long gone having achieved what they wanted when they were cheating. It is a societal problem or institutional and industry problem. If the market and these universities want these degrees to mean anything as signaling mechanism they must stop these students from cheating some way somehow...the students are never going to stop themselves.
Some college students may be genuinely interested in one particular subject, but they're required to take a bunch of other courses, and consider those to just be hurdles.
I still think they're better off at least making an effort and trying to learn something, but I do think it's important to note that just because a student has no interest in one particular class, doesn't mean they have no interest in any class.
The course I'm interested in gets kinda hard, and my "just pull up an LLM" muscle is very, very strong, (and besides, I'm not used to struggling! and why should I get used to it in the classes i like?! I can't afford a C in my major!) so ... I use LLMs on my "I'm interested in it" class too and... we're back to the original argument.
> Similarly, vocational programs or apprenticeships are not a formal education either.
They are in some countries, you get at the vocational programs or apprenticeships alongside the highschool, and in the end you might get the opportunity to apply to the university or just carry on with your job.
That is how I did mine in the 1990's Portuguese education system, and how I was already coding and understanding the big boys computer world at 16y.
But if your goal is to be a musician, a music degree is basically useless. Whether you want to play in an orchestra, perform in a rock band, or compose video game soundtracks, nobody cares whether you have a degree or not - they want to hear you perform.
I won't say there are exactly zero self-taught professionals in classical music because I don't know for sure. But I will say I've never heard of one. If they exist at all they're exceptionally rare.
The music industry is built on the back of people with music degrees. They don't get the name recognition of headliners. But song writers, arrangers, and session musicians are all very likely to have formal training in theory and maybe performance.
Producers and engineers less so. Those are more of a track record who-you've-worked-with occupation.
Not only is that not true at all. There are many other jobs you can land other than performing with a formal music degree.
Of course with the right experience you might get away with not getting that formal education but you open so many doors by going through school and getting the "useless" piece of paper.
Case in point: Rachel “Raygun” Gunn. She had all the credentials in the world but single-handedly became the reason break dancing is no longer in the Olympics.
Slight overstatement… break dancing was one of the locally picked sports and the next Olympics had already selected different sports before she performed…
But she is a good example of degrees not equaling skill
It should also be noted that she doesn't have a degree in "performing," as far as I'm aware, she has a degree in "studying the culture" of break dancing. So, we (or at least any of us who haven't read her work) don't actually know if she's good at what her degree is in. We just know that she's not good at performing.
Most people don't go to college to learn for the sake of learning – that would be a very expensive luxury. They go for the opportunities that it opens up. The learning itself helps with that, but so does being able to prove that they've learned it. That what the diploma is meant to be for.
I don't think that's a good comparison. I went college to learn and was also relieved occasionally when professors canceled class and still had to force myself to study. I'm sure plenty of your piano teachers students don't enjoy practicing the same notes over and over but do so because they want to know how to play piano.
There are jobs that turns you down if you don't have that diploma. No amount of real world experience will get you past that bar. If it's not there, they wont even look at the rest of your application.
Absolutely. Also, the level of cheating in college, even pre-AI is often overlooked in these articles.
For the exact reasons you state, pre-AI homework was often copied and then individuals just crammed for the tests.
AI is not the problem for these students, it's that many students are only in it for the diploma.
If it wasn't AI it would just be copying the assignment from a classmate or previous grad.
And I imagine the students who really want to learn are still learning because they didn't cheat then, and they aren't letting AI do the thinking for them now.
This is, famously, why nobody gets a degree in fine arts. (/s)
Your piano teacher does not give a diploma because she is not offering a university education. If she worked with a few other experts and they designed a coordinated curriculum and shepherded students through it over the course of two to four years, and documented that process to the point where they could file with an accrediting agency, then she could issue a degree in piano.
You can get a BMus in Piano Performance, which is a standard music degree with a piano performance specialisation. It's not a special thing of its own kind, and it's normal for performers in music college to specialise in one instrument, usually with a second as an elective.
Being able to play moderately hard pieces from dots and sight-reading are the entry level requirement. It's taken for granted you can already do that.
The degree part means learning music history, theory, and performance styles, working on performance projects, solo and with other musicians.
The analogy with ChatGPT is that it's taking over the entry-level part of the process. You can't expect to get onto a music degree if you only know how to prompt ChatGPT to produce a MIDI file for your entrance exam.
And in CS, you can't produce good code if you barely know what a server is.
It's all very Dunning Kruger. If you use an LLM to produce course work to get your piece of paper at the end, you don't even know what prompts you should use to do an unfamiliar job, never mind having the skills to do it yourself.
Yes they did actually say that. They go into the process of accreditation that would result in a person or institute being able to give a degree "in piano"
The point of education isn't to actually learn though. It's to receive the credential.
This is much larger than a cultural problem with the students of today. They believe, rightfully and accurately, that the university degree is yet another part of the machine that they will become a cog in.
What should be alarming to everyone is that these students will graduate without having learned anything and then go into the workplace where they will continue to not use their atrophied critical thinking skills, to simply do yet more, as a cog in the machine.
This attitude is part of a more general cultural shift. Back in the 1960s, the majority of students said the primary motivation for going to college was to develop a philosophy of life, and a minority said the main goal was to be very financially successful. Somewhere around the 1980s this started to shift and the proportions are now inverted.
Arguably the cause of the cultural shift is economic. Wages began stagnating in the early 1970s, which caused increased demand for diplomas as a way to increase wages. This is most striking in the number of students seeking law degrees, which shoots off at around the same time.
I would also say some of the attitude shift is also contradictory. The amount of people I interact with who have a lot of bad things to say about the education who tell me universities should focus on education in a general meanwhile also say that schools to should focus on student getting jobs. Probably one that has been heard before, something along the lines of, "why don't high schools teach plumbers courses." I mean they can. While also, "colleges are too focused on checking the boxes so students can get jobs."
It's not inherently contradictory to hold both positions, if your overarching position is that the schools are in a weird in-between state that serves neither well. My time in school was marked with both forms of frustration.
I wanted to learn a new language and I wanted to take some history courses that covered regions and eras not well covered in my high school courses. But despite my university having a significant "elective" component to my degree path, none of those courses were on the list of allowed electives for my degree. So in this case, the university was failing at a focus on education by hindering my ability to branch out away from my core studies and requiring that I take "electives" that were more closely associated with the imagined career path my degree would provide.
On the flip side, the "core" courses for my degree were bogged down in academic minutia and exercises that bore only the most surface level resemblance to the things I've done in my actual career. Often the taught material was out of date relative to the state of the industry. Other times the material was presented with philosophical reasons for learning the material, but with no practical application backing it to help make that philosophy complete. And very little material (if any) covered the usage of tools of the industry. In this case, we're failing the goal of setting people up for their careers by not teaching the practical applications of the knowledge. And to be clear this isn't just a "learning examples are by necessity simplified examples" problem. I later went back to school at a local community college for different material and from day one those courses were more relevant and more up to date. They provided material that was immediately useful in real world applications of the underlying knowledge. And I think some of that was because many of the courses for that community college were taught by industry veterans, either part time or as a "retirement" gig.
In short, my experience at a large university was indeed a series of boxes that were to be checked, ostensively to provide me a "well rounded" education, but practically all narrowly focused on getting me a job in the field. Yet the boxes also failed at being relevant enough to the state of the industry to actually give me a foundation to work from when starting my career.
And there were fewer students going to university and so you had a higher proportion of people doing it for the love of learning. It was easier to get a job without a degree. It's not just a cultural shift, it's a change of supply and demand.
Notably, the late 70’s and early 80’s is when large scale social changes started to happen due to high inflation and major problems in the US economy. Also a lot of social unrest, and a pretty unhinged president (Tricky Dick).
The same issues are seen in other countries. It's not specific to US presidents or policy. The insistence on maximizing university attendance was a widespread idea coming out of the left, e.g. in Britain it was heavily pushed by Tony Blair. The Blair government also raised tuition fees considerably.
The stated rationale at the time was that degree holders earn more, so if everyone gets a degree, everyone will earn more. I am doubtful that was the true rationale but it's how the policy was sold to people.
We (America) made university educations super expensive. Causality is complicated, but it probably started as an effort to destroy the anti-war movement in the 70's.
"Most financial experts attribute he sudden increases that started in the 1970s with an influx of federal funding designed to make college more affordable."
> Somewhere around the 1980s this started to shift and the proportions are now inverted.
Yeah, that's when the great "push for education" came, as well as neoliberalism which preached continuous hustling and individuality. And in the 90s, the ADA
and other anti discrimination laws hit, and requiring a college degree was and still is a very useful pre-screening filter for HR to continue discrimination.
It's also as I recall when tuition and student costs started to spike in the US which is probably more directly related than some "philosophical" change in the zeitgeist[0]. When students and parents start racking up debt like this, you become very interested in the fastest way to pay it back.
For me the impact of the university administrators as they chased higher endowments for more buildings with naming rights and expanded their own bureaucracies with direct hires that did not directly contribute to the faculty mission did more to alter the university experience than anything else.
Me and most of my peers in college had the choice between two courses. Course A was interesting, yet vastly more challenging and therefore time consuming, with the additional downside of lower grade expectation. Course B was boring, a gentle breeze in comparison, yet with an almost guaranteed perfect grade.
Imagine which course most students choose?
Even if a student wants to take on the more interesting course, incentives matter, and the incentive is: better grades qualify for better compensated positions and prestigious degrees. Only students who didn't care about this or were confident enough in their ability did choose Course A. In the end, barely a handful of students out of hundreds went with A.
Well, the credentials defintely help to GET the first job. As a cog in the machine though you are most often valued for skills more than for credentials, at least in the U.S.
> The reason a teacher asks you a question is not because they don’t know the answer.
A decent amount of my professors don't know the answers because they bought the course, test questions, and lectures from Cengage. During exam review, they just regurgitate the answer justification that Cengage provided. During the lectures, they struggle to explain certain concepts since they didn't make the slides.
Professors automate themselves out of the teaching process and are upset when students automate themselves out of the learning process.
I can tell when the faculty views teaching as a checkbox that they officially have to devote 40% of their time to. I can tell when we are given busywork to waste our time instead of something challenging.
To use your analogy, I'm being told to move 1000 plush reproductions of barbells from Point A to B by hand because accreditation wants to see students "working out" and the school doesn't want high failure rates.
We are all pulling out the forklift. Some of us are happy because we don't have to work as hard. Others are using the forklift so we can get in a real workout at home, as school is not a good use of our time. Either way, none of us see value moving paperweights all day.
edit:
My favourite course during my Computer Engineering degree was Science Fiction because that professor graded us on substance instead of form. It was considered a hard class because one would get good marks on the essays by focusing on building substantive points instead of strict adherence to the form of a five-paragraph hamburger essay.
The call to action is to make courses harder and stop giving students plush barbells.
For example, University of Toronto Engineering Science (hardest program in Canada) gives first-year students a "vibe coding" lab in which students learn how to solve a problem that AI cannot.
There are many issues here. The lack of incentives is probably the most important one. For new professors (in research universities), good teaching is usually just a good thing to have, but it is not a deciding factor for their tenure. When they get their tenure, they probably have enough students, and they need to work hard to apply for funding and keep the students paid. Administrators care most about ranking, and teaching isn't really evaluated in the ranking. They just push the professors to do more research and apply for more funding.
It is also hard to evaluate university teaching because there are no benchmarks for that (compared with high school, for example), and it is hard to judge if teaching is good from student feedback. You can only know if someone fucked up or did really well, which are outliers.
There are other issues as well. Professor IMO is a ridiculous job, you are supposed to be an expert in the field, be a researcher, be a manager, be a teacher, be a salesman, all at the same time. There are people who can excel in all these, but these are probably just outliers. It doesn't help when PhD training doesn't train you to be a proper manager and teacher. While there are some teaching training, I think we are not really held to a high enough standard. E.g. One can pass the teaching course if they just show up and spend some time, even though their teaching is horrible.
Using an LLM to do schoolwork is like taking a forklift to a gym where you're told the goal is to be healthy and strong, but they can't really stop you from using a forklift, and jobs and compensation are given out according to how much you lifted irrespective of forklift use.
The tragedy is not that some students are going to college to get a diploma while learning as little as possible. It is that the boards of many private universities see their students' cash as more important than their education, and force the professors to pass everybody who went to higher education to buy a diploma.
This has a negative feedback loop where universities have to lower standards to bring dumber and lazier students to compete with other diploma mills.
> If all we were interested in was moving the weights around, you’d be right to use a tool to help you.
Does the use of a quantifiable metric like a GPA not exacerbate this? In a world where people take a GPA seriously, you'd have to be irrational to not consider cheating a viable option.
You could say the same about credit score and dating apps. These institutions assist the most predatory and harm the most vulnerable.
Why do some countries cheat in the Olympics? Because it is no longer a contest of human achievement, it's just about the medals as a symbol of national glory. Of course: once all countries are doping, the medals will become meaningless. College degrees will suffer the same fate if everyone cheats to get them.
> Using an LLM to do schoolwork is like taking a forklift to the gym.
I'm sure the time has come for college students to master using LLMs. It's just as important as grammar or basic math now. The software I build (and the entire tech industry) automates huge swaths of business processes with AI. Students need to be able to understand, work with, and manage swarms of AI agents doing work.
To stick to the analogy:
I need skilled forklift drivers, not big buff workers like I used to.
But the thing you are missing is that one needs a solid foundation of knowledge of the actual work to be able to manage it well.
Someone with years of coding experience is going to be able to laser guide an AI agent to the answer/result than someone who has muddled their way through comp sci 101 using an AI chatbot.
I use chatgpt in a socratic way from time to time because I don't want answers I want the joy of thinking and learning. I heard there were efforts to make educational LLMs (whatever that means). Maybe it will help multiply teachers leverage so that more kids get inspired without having the teacher spend 1-on-1 time with them.. I don't know.
I think we have to hold off a bit on the whole thing here.
Look, we have no idea what the feedback is like that this grad student gives, what the class sizes are like, what the cadence is, what the grade percentages are, etc. All we know is that Clayton Ramsey is a grad student at Rice in the robotics department and that he wrote a hot take here.
For me, the most important thing is if this grader is bothering to really grade at all. I think we've all had a harried grad student just dash off a few red lines on the week one HW about a week before the final exam. That's not a 2 way street, and if the feedback isn't as in-depth as he wants the work to be, well, he shouldn't be surprised. He can't be expecting students to put in the time unilaterally. But, we don't know any of that really.
Personally, I think that before the decade is out, we're not going to be talking about this at all. Because the students will be adept enough at using the LLMs to make it look like their own writing anyways. This is a problem that experience will solve for them.
And also, I think that the days of the massive lectures and essays are pretty much cooked. That 'cheap' model of education can't survive this LLM revolution. We obviously have to change what the heck higher education is trying to do.
My take is that we're going to go to smaller class sizes like those at St. John's or Oxbridge. Under 10 people, you have to have done the reading or look like a fool, all with a PhD in the subject as a guide/teacher. Large classes weren't cutting it for decades (ask any Frat about their test banks), and now the veil is just ripped off.
> The reason a teacher asks you a question is not because they don’t know the answer.
I remember illustrating a point to a class by posing a question and then calling on a student I figured wasn't smart enough to answer correctly so that everyone could see her make the mistake.
I think LLMs, if used correctly, can be useful for BOTH the credentialing and the human resource development (*cough*)
Essentially, since they are a summary of "the" state of knowledge, the teacher should be able to ask them to put a number on how novel a piece of text is.
Once LLMs are able to evaluate, independently, the soundness of an argument... (Hopefully, this will be achieved AFTER $5 H100s reach the average consumer)
If you took a forklift to the gym, you'd come out of the experience not only very good at "lifting weights", but having learned a whole lot more about the nature and physics of weightlifting from a very different angle.
Sure, you should lift them yourself too. But using an AI teaches you a shit-ton more about any field than your own tired brain was going to uncover. It's a very different but powerful educational experience.
> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think
I'm there for the degree. If I wanted to learn and engage with material, I could save $60,000 and do that online for free, probably more efficiently. The purpose of a class writing exercise is to get the university to give me the degree, which I cannot do by actually learning the material (and which, for classes I care about, I may have already done without those exercises), but can only do by going through the hoops that professors set up and paying the massive tuition cost. If there were a different system where I could just actually learn something (which probably wouldn't be through the inefficient and antiquated university system) and then get a valid certificate of employability for having actually learned it, that would be great. Unfortunately, however, as long as university professors are gatekeepers of the coveted certificate of employability, they're going to keep dealing with this incentive issue.
> I could save $60,000 and do that online for free, probably more efficiently.
Not to burst yours or anyone else's bubble, but no, probably not.
The hard part of learning isn't access to content, it's discipline and dedication. School provides structure, goals, timelines, and deliverables. The value of this cannot be understated.
I've heard from many people how they're going to learn programming online and then get a job as a developer. Almost all of them fail.
I'm teaching faculty at a university and, at least where it comes to lecture courses, I don't find this a bubble. There are definitely plenty of students who would do just fine, if not even better, without the external discipline and structure. And even more those who could do it with something like a MOOC or just a posted curriculum. And to be able to work without an external discipline is IMHO one of the main learning goals of university, and a decent rationale for why a university degree is taken as a signal for recruitment.
I learned programming online and got jobs as a developer (I did later study CS at a university though). In my experience the best developers are those who taught themselves. Admittedly this may have been more the case for my older generation where formal education for programming wasn't that great nor widely available.
Those students do exist, but there are an exceedingly small minority. Of course, they won't tell you this, because everyone likes to believe they're self-motivated. Most people just aren't.
The simple question to ask is, when you go home, what do you do? If the answer is learn how to sew or work on your project car you've had for 10 months, you can probably learn on your own. If your answer is watch TV, play video games, go on a walk - then you can't, and you should go to university. Some people have told me this question is unfair. I mean, they're so tired from work, of course they want to relax. Well, guess what - your life doesn't stop if you're learning how to code on your own or whatever. If that's all it takes for you to not do it, then you don't have what it takes.
How often are people picking up new and complex skills that takes years to get the hang of? Almost never. So there you go, most people require a formal, structured education to pull that off.
> There are definitely plenty of students who would do just fine, if not even better, without the external discipline and structure.
How do you know? It's easy enough to assert, but what kind of proof can there be for this assertion? Obviously the students are enrolled in university, and their accomplishments without it are only hypothetical.
>School provides structure, goals, timelines, and deliverables. The value of this cannot be understated.
That _may_ be true for the vast majority, but it's criminal to waste the time of bright young people by putting them though hoops. I would even speculate that that they phoney goals, timelines, and deliverables in school actually damage kids.
> I'm there for the degree. If I wanted to learn and engage with material, I could save $60,000
I would argue that if it costs $60,000, both your education system and the recruitment in those companies that require this degree are broken. It's not the case in all countries though.
Not that it is your fault, just stating the obvious.
It is broken. For every coveted job there are thousands of applicants. Employers will accept any signal that reliably predicts a modicum of intelligence, conscientiousness, and agreeability. University degrees cover all three.
But that's just the job market. The other elephants in the room are inflation and the housing market. People who don't have top-notch jobs (that require degrees) can't afford to buy a house. They can hardly afford rent. Cities don't want to build more housing because that will undermine the equity growth of homeowners.
>The other elephants in the room are inflation and the housing market.
I'm not even sure if people are aware of inflation/housing as a completely solvable issue by the govt. I guess it's because people most people are clueless on how it is to be solved.
>We are a society of ladder-pullers.
It's by design, to serve the rulers. It's an assembly line of slaves who are given some freedoms and are put through various stages of school, university, work and retirement. When most people retire they are left with little to nothing.
I don't disagree, but often we complain about people pulling up ladders and when faced with the same decision we follow suit. Ultimately we can't change this behavior if no one is willing to defect from "conventional wisdom"
We can't fix the problem by making better choices as individuals, and exhorting people to do so saps energy and distracts. The system interprets integrity as damage and routes around it.
> Ultimately we can't change this behavior if no one is willing to defect from "conventional wisdom"
To me it reads as if you're contradicting yourself. It can be plainly seen that almost nobody is willing to defect. So change is impossible, you claim. Also change is possible, you claim. Which is it? You think people will change their choices as a result of you writing this comment and "winning" the argument?
I agree that people are responsible for the current situation, since their choices brought it about. That does not constitute a solution, though.
One or a thousand individuals changing their choice by themselves will do nothing. For tens of millions to change some external factor is required. If it was not, then it would have already happened. You say no external factor is required, so why do you think it didn't happen already, and why would it happen in the future?
The problem of individuals making optimal choices on an individual scale being sub-optimal on a society scale is so widespread we have a specific phrase to describe it: Tragedy of the commons
I would love to live in a world where everyone was altruistic and made correct choices for the long term good of society, but I don't. And there are limits to how much I'm willing to act as if I do, when in practice it just means I'm giving away resources to people who are purely (thinkingly or unthinkingly) selfish.
The problem was created by individuals deliberately acting collectively, not simply choosing one way or another in their routine individual capacities. The solution will require the same.
This attitude willingly donates what agency you have to the people and things trying to take it from you.
Lead by example. Mimicry is real, we all do it whether or not we are aware. Every node in the graph influences others. If you must exhort, it works better if you follow your own advice.
Of course, be discerning as you do this, and don't expend your energy or goodwill where it will be wasted. Be like Gandalf.
Any dogmatic system is fault-tolerant, in that it will "route around" some amount of internal dissent, but this does not make it impregnable.
The ruts in our minds steer us just as much as those in the ground. But earth turns and so can we.
My tuition was $35k a year 25 years ago. Just checked, and now it is $60k a year. Before room and board.
Broken? Saddling individuals with a quarter million in debt when they are just starting life is absolutely broken. That they must indenture to be a modern professional (and buy hope for at least a middle class landing) is broken.
The notion that everything must return a (generally, near-term) accounting profit is on its face stupid.
Sure, the system is broken but what's the alternative? Employers have a surplus of applicants for entry-level technical positions. They need to filter the applicant pool down to those with some level of competence and discipline. Possession of a college degree is a reasonably accurate proxy for those attributes: lots of false negatives but good enough from the employer's perspective.
Ideally maybe employers ought to rely on more targeted selection mechanisms. But this would be extremely expensive (and potentially legally risky due to equal opportunity laws) so most don't bother.
>Sure, the system is broken but what's the alternative?
For a true solution, the entire taxation and monetary system will have to overhauled. It's of course not going to happen.
Transactions outside of the govt monetary system is effectively illegal or taxed so people are forced to participate by applying for jobs for their livelihood.
Are you saying the US is the only country that has an excess of applicants for entry-level positions? Or the only one for which credentialism is the solution to this problem? If the second, how does the place you're from solve it?
> I would argue that if it costs $60,000, both your education system and the recruitment in those companies that require this degree are broken.
Meh, academic degrees don't come for free, someone has to pay for universities, staff and other expenses. In the US it's everyone for themselves by student loans that can't be discharged in bankruptcies, in Europe it's the tax payers.
The problem is, the ones profiting from the gatekeeping (aka employers) aren't the ones paying for it in either system. If employers had to pay, say, 10.000$ for each job listing that requires an academic degree without an actual valid reason, guess how fast that incentive would lead to employers not requiring academic degrees for paper-pusher bullshit jobs.
In a sense, it still comes down to supply and demand — if applicants, upon graduating, all requested to be reimbursed for, say, 25% of their tuition up front and to the university they graduated from, we'd end up with reduced salaries compared to just distributing those 25% to the new employee over a number a years.
But how do you get all students to agree with this in principle when someone is in more rush to start earning an income than others?
However, employers would then look to only hire from universities that do good teaching, so maybe it's a win-win?
No, only for jobs that claim to require higher education but do not. Basically, an "abuse tax" to reimburse the government (or the students) for having to spend money on something clearly not needed.
There is also nobody cares about low prices but you can brag about how exclusive you are with high prices. One univertity near me automatically gives everyone a 40% scholarship - which is to say they have inflated their sticker price.
The work this degree will credential you for is so is so disconnected from the areas of study in your degree program - presumably in the same field as the job - that the majority of the things you might learn would not be valuable?
I can’t imagine this in my own life. I use concrete things and ways of thinking and working I learned in my CS degree _all the time_.
Heck my degree was in biochemistry, and now I’m a programmer, but I still feel like I am constantly using skills I developed in school. The scientific method and good test design transcend all sciences.
When you're in a position to hire or influence hiring, will you consider those without degrees?
I ask because I hear this sentiment a lot but we still have a system becoming more reliant on degrees. The universities may be the gatekeepers of those degrees but they're not the ones gatekeeping the jobs. They have no influence there. They were not the ones who decided degree = credentials. I ask because many people eventually grow in their jobs to a point where they have significant influence over hiring. So when that time comes will you perpetuate the system you criticize or push against it? Truthfully, this is a thing that can be done with little to no risk to your own employment.
A person is a idiot if he/she takes someone's competence at face value because of a degree.
( jobs aside - don't assume your doctor is competent because he has MD, it will cost you your life)
I see your point, but the issue is that it's quite futile to shame students for playing the game that people in the industry has set up. It doesn't help that in the past, college degrees were in fact more relevant than today, especially before the era of the internet and wikipedia, so if older people who are currently in charge of hiring aren't aware of these changes, they might just apply their outdated personal experience and just assume college degrees hold the same weight as they did in the past.
I'm pretty certain when kids these days eventually become responsible for hiring decisions, they probably will handle things differently since their experiences are different.
I do appreciate that you do that! People making change is the only way change can happen.
But I think the second part is amiss. I do agree that educators (being one) need to recognize the reality of the environment. But I think you also can't ignore that the reason the degrees are used to signal qualification is because the degree is intended to signal that some knowledge is held by a specific person. Yes, things have changed. It did used to be that interviews were much shorter, with a degree being a strong signal (at minimum, it is a signal that someone can sufficiently follow directions and manage their time). Should that mean those tasked with ensuring a student gets their education does not get their education? I get your point, but I think it is barking up the wrong tree. Asking academia to change only furthers the problem as it lets employers shift blame. It is like getting upset at a minimum wage worker for increased prices due to tariffs.
The really weird thing is we're now in this bizarroland setting where employers will filter for degree and then spend months and a lot of time and money putting candidates through testing rounds. Making them do tons of leetcode and other things to evaluate their performance. Where the candidates spend months studying to pass the interviews. And the complaints here are probably more apt, about how the material they need to study is not significantly relevant to the skills needed for a job[0]. Worse, it doesn't seem to be achieving the results it is setting out to seek.
As they say about the stock market "the market can stay irrational longer than you can stay solvent". My point is that some academics are making classes easier. Grade inflation is a measurable phenomena and it sure got worse with Covid. My point is that employers are acting irrationally.
[0] I'll give an example. I had an interview a few weeks ago and we had run a bit longer than expected prior to getting to the coding portion of the interview and where their coding software crashed a few times, giving us 10 minutes to do what was expected to be 30. Finally coding, the program crashed and I said "I don't use a lot of Juypter Notebooks, does the kernel crash when OOM?" I'm explaining my thought process aloud and frankly, I'd never hit this on the job. No answer. I quickly scroll through logs, say "I'm going to guess that's the issue, I'd normally guess and if wrong google it". Yes, it was the right answer. But there was no reason to do this, especially with me guessing right. This wasn't some dinky startup, it was a multi-trillion dollar company...
This seems to be the general feeling of students right now.
Academia put itself as a gateway and barrier to the middle class. Why would we be surprised when people with no interest in anything but the goal are not enthralled by the process?
How did academia do that? It doesn’t seem like universities would have the power to do that. More likely, either employers put academia as a gateway. Or even: the culture at large misunderstood what pathways existed to middle class life. Or even: pathways to middle class life became scarcer and more insecure, and the real gatekeepers (hiring managers) had no good ways to select which of the many people at the door to let through.
I agree with you (and upvoted) but consider this: why don’t fancy schools offer the same credentials for MOOCs if the information and assessment is on-par? I believe it is because they recognize the value of the credential and guard it closely.
I think you missed the parent's point. Maybe consider this:
Why don't employers recognize the credentials of a MOOC to the same degree that they would a university degree?
We could similarly ask why employers value the degrees of some universities more than others.
I think it's important to realize that ultimately the decisions come from the employers, not the universities. No one is making the employers do anything. But at least the second question might have a clearer partial answer. In part, there is a selection of a tribe, an implicit "culture fit" that's happening. It isn't uncommon to see employer bias towards specific universities. This is especially true with prestigious universities.
But it's not the universities that are making anyone do anything and that's an important distinction.
I brought up that I think the root problem is employers in a different comment, so I’m not missing that point.[1] The distinction is that I don’t think it’s a dichotomous problem. While colleges may not be the root problem, they can still be a contributor to the problem. (Besides, if an employer is to recognize a MOOC, that course completion has to be documented, which means it’s just another version of credential)
Academia clearly lost their monopoly on information. Since their last moat is a monopoly on credentials, I expect them to defend it intensely.
We could make it less meaningful if employers weren’t so keen on using credentials as their own gateway. That may have more of a chance of happening if the OPs perspective becomes more prevalent and the credential becomes an increasingly worse signal for meaningful skills.
They didn't have a monopoly on information anytime in the 20th century. If you wanted to have all the knowledge, it could be expensive to acquire a personal library, but nothing was stopping you from acquiring the same textbooks. Heck, until the information age it was pretty easy to forge a credential.
But it wasn't about credentials even. It was about inculcating a culture. You knew that someone had the knowledge and ability to reach university, and you knew that they had a shared common culture with you. Shared common culture and norms increases trust. Credentials mattered for doctors, but universities, in the end, were selling something far more intangible: culture.
High barriers to entry are a feature of monopolies. Saying I could start my own railroad by simply buying up billions of dollars of land and investing billions more in equipment isn’t a compelling argument against the idea of a monopoly.
There are accreditation bodies. So I don’t think your self-proclaimed engineering degree is going to help you get a job or a professional engineering license like one from an ABET accredited school.
I don't know where you live, but in the "developed" parts of the world this is illegal. There will either be some government agency or some council of credential-giving institutions and they will give you a license to issue degrees, or most likely they will not give it to you.
In the US, you can just make up degrees — but you have to be honest that you’re unaccredited.
Accreditation is regulated by NGOs who need government approval and without that you cant received financial aid (or participate in some programs) — but you can hand out pieces of paper for completing your program.
That’s a disingenuous argument. You don’t know what you don’t know. Literally. A completely self guided high school graduate following random online materials will not learn nearly as much on their own. Or they will go down rabbit holes and waste countless hours, and not having an expert unblock you or guide you down the right path would waste a lot of time.
Further, some high school graduates (like myself at the time) literally don’t know HOW to learn on their own. I thought I did but college humbled me, made me realize that suddenly i’m in the drivers seat and my teachers won’t be spoon feeding me knowledge step by step. it’s a really big shift.
If you were the perfect high school graduate, then congrats, you’re like the 0.01%! And you should be proud (no sarcasm). This doesn’t describe society at large though.
For the very few that are extremely motivated and know exactly what job they want, i do think we need something in between self guided and college? No BS - strictly focusing on job training. Like a boot camp, but one that’s not a scam haha.
The other aspect of college you ignore is, it is a way to build a network prior to entering the workforce. It’s also one of the best times to date, but that’s another story.
Completely agree that the cost of college in the US is ridiculous though.
>The other aspect of college you ignore is, it is a way to build a network prior to entering the workforce.
I don’t know how generalizable this is. I remember reading a few studies trying to assess if Ivy League education was really more valuable that a state school. The result (IIRC) was that it only matters for students who came from the lowest economic strata; the authors presumed it was due to the network effect. But that also means the network effect was negligible for the majority of students.
I used to work for a guy who had his engineering degree from Harvard. I was out of the Navy, got a job but also had a GI bill and was looking to go to school for engineering while working full time. I asked him his advise. Pretty much he said is to go to a state school. The only real benefit, at least for engineering from Ivy League, is the network, which can make getting the first job easier. Otherwise the course work is the pretty much the same.
>That’s a disingenuous argument. You don’t know what you don’t know. Literally. A completely self guided high school graduate following random online materials will not learn nearly as much on their own. Or they will go down rabbit holes and waste countless hours, and not having an expert unblock you or guide you down the right path would waste a lot of time.
Citation needed. There's great books out there that provide a lot of guidance down a particular path. I'd say a lot of them do, and I can't imagine online learning sources would be worse. There's online communities for learners for specific subjects that are full of people offering good advice.
I don’t have the time to provide a citation, my apologies.
I’ll leave you with this thought though. Of all the professions, tech is probably the one where this is the easiest to do. There are many companies that don’t require a bachelors. For most of the last 15-20 years tech was in a “boom” cycle, and yet the vast majority of software engineers I met DID have a bachelors degree.
Why? If it’s as easy as “pick up a book”, then why didn’t more people take that path? I think very few people have the drive and discipline to accomplish a full career prep on their own.
If your hypothesis was true, wouldn’t tech be mostly filled with self-taught engineers?
I would argue that while tech is the easiest to get into without a degree:
1) The cultural zeitgeist around higher education (at least in the US) is and has been "you must go to college and get a degree". It's been over 20 years since I dropped out of my first university. I'm doing just fine, and yet to this day, I will be asked by older members of my extended family if or when I'm going back to "finish" my degree, or whether my company will offer tuition assistance to help me go back. If you graduate high school these days (and for at least the past few decades), the expectation is that you go to college and get a degree. And if you're going to have to do that anyway, you might as well get the degree in the field you want to go into.
2) As a corollary to that, while new / younger companies might have set aside the degree requirements, the big tech houses definitely still preferred them and having the paper was still a leg up in the hiring and recruitment process. And even in companies without an explicit preference for a degree, it's often listed as a requirement on the job posting. "BS or Equivalent Experience" is easier to match both as a candidate and as an employer if the candidate has the BS as that is an objectively verifiable fact.
The purpose of the writing exercise is to produce a positive correlation between the possession of a degree and the skills that high-paying white collar jobs value. I don't blame students for not knowing that, or for not having the outside perspective to care. But positive signals of job competence are hard to come by, employers don't just blindly accept them despite what people like to say, and it's going to suck for new graduates if this one is eliminated.
It's especially damning since cognitive research is not at all ambiguous on the topic. Learning is the result of deliberate practice, intentionally placing yourself in situations where you need to solve problems. The output is irrelevant, building mental capacity is the game. The struggle to learn isn't just some unfortunate obstacle to be optimized away by technology. It is the cognitive mechanism by which we build knowledge.
This is somewhat reflected in how we value university degrees. You get very little additional salary from having finished all but the last semester of a degree. The big boost all comes from the last semester where you get the degree. You'd expect that the vast majority of the actual knowledge is already there at that point and late dropouts would be seen by employers as some great bargain. It shows that the signaling of the degree trump's the actual knowledge. Good discussion with Bryan Caplan on this: https://www.econtalk.org/bryan-caplan-on-college-signaling-a...
So, unfortunately the student's behavior is somewhat rational given the incentive structure they operate in.
When I was a kid and got an assignement for writing an essey about "why good forces prevailed in Lords of the Rings" as a gate check to see if I actually read the novel I had three choices:
(a) read the novel and write the essey myself
(b) find an already written essey - not an easy task in pre-internet era but we had books with esseys on most common topics you could "copy-paste" - and risk that the professor is familiar with the source or someone else used the same source
(c) ask class mate to give me their essey as a template and rephrase it as my own
A and C would let me learn about the novel and let me polish my writing skills.
Today I can ask ChatGPT to write me a 4 pages essay about a novel I've never heard of and call it a day. There's no value gained in the process.
That's a simple example. The problem is that the same applies to programming. Novice programmer will claim that LLM give them power to take on hard tasks and programm in languages they were not familiar before. But they are not gaining any skill nor knowledege from that experience.
If I ask google maps to plot me a directions from Prague to Brussels it will yield a list of turns that will guide me to my destinations, but by any means I can't claim I've learned topography of Germany in the process.
> Today I can ask ChatGPT to write me a 4 pages essay about a novel I've never heard of and call it a day. There's no value gained in the process.
If we take the original article at face value, no you can't do that. ChatGPT will apparently produce something that is obviously ChatGPT produced and fail to fool even the most absent minded of instructors that you have read the material. So even with a ChatGPT LLM to help you out, you're largely going to have to do a modified version of C, replacing your class mate with the LLM and adding in the need to do your own reading and validation to ensure that the text matches the actual book contents.
> If I ask google maps to plot me a directions from Prague to Brussels it will yield a list of turns that will guide me to my destinations, but by any means I can't claim I've learned topography of Germany in the process.
I would argue that even if you plotted a route by hand reading maps, you can't claim to have learned the topography of Germany either. "The map isn't the territory" after all.
> If I ask google maps to plot me a directions from Prague to Brussels it will yield a list of turns that will guide me to my destinations, but by any means I can't claim I've learned topography of Germany in the process.
There are multiple ways you can use such technology, too. If you use Google Maps with its out-of-the-box configuration for turn-by-turn directions, with it oriented in the direction of travel, you won’t learn so much; but if you change it to always display the map north-up, and look at the map it shows you—inferior though it be to good paper maps, in most cases—it’s easier to develop a feel for layouts and geography.
> Novice programmer will claim that LLM give them power to take on hard tasks and programm in languages they were not familiar before. But they are not gaining any skill nor knowledege from that experience.
Not true.
Using LLM to learn quickly a new programming language + being productive is best method ever. If you pay attention, you acquire rapidly new skill and knowledge, and those that are relevant to your job.
Using LLM is MUCH more efficient than reading a book going through all the minute details of the language prior telling how to use it. It's the same as learning a language from your parents compared to learning a language from a class. You might not know all the grammar rules, but you'll be way more proficient. And nothing prevents you from learning the grammar later on.
> Using LLM to learn quickly a new programming language + being productive is best method ever. If you pay attention, you acquire rapidly new skill and knowledge, and those that are relevant to your job.
I tend to disgree. maybe for you, but my mind does not work that way.
As a teacher at a university I come to see that students "learn" by asking an LLM, but they forget to understand the content the LLM produces because the LLM actually solves the assignment (mostly) for them. One may say that's the teachers task to produce better questions, but the thing I most "struggle" with that getting educated as a student seems to just be a play of "gaming the system". Yes, it was similar during my time (learning how to reach your goal with as less effort as possible is a good part of the "education" at a university IMHO), but we actually had to think and understand while today just seems like prompt-and-copy.
> Students who know least of all and don't understand the purpose of writing or problem solving or the limitations of LLMs are currently wasting years of their lives letting LLMs pull them along as they cheat themselves out of an education
My high-school age daughter told me how her small private school solved this problem:
They brought back oral exams.
There aren't a lot of other good options. Written take-home work and online tests have always been fertile ground for cheating. Another benefit of oral exams: you learn to communicate under stress.
I think these complaints will fall onto deaf ears for at least two reasons:
1. Students are a captive audience. They don't want to be there. It's the law that makes them be there. Even once you're beyond mandatory education this holds true: they were just carried into further education by momentum. They didn't realize they had a real choice or what alternatives were available.
2. A lot of the skills you build in classes aren't useful to you. I spent a lot of time in my English (second) language classes, but it was my use of the internet that really taught me the language. The later years of English classes was just busywork.
In my native language classes I had to write a fair number of essays. The only time this was useful was the final exam of that class. I haven't written a "real" essay since. Even if I did, it would probably be in English and use a different style - something taught to me by forum posts.
But this is exactly the problematic viewpoint. You thought that the point of letting you write essays in your native language was to enable you to write essays. It isn't, it never was.
During the final exam, a national exam (all the students did it at the same time in the country) you were given blank pieces of paper, a pen, 10 different topics to choose from, and 6 hours.
You pick a topic, write a draft, then write the essay. ~700 words.
Bonus points if you use relevant literature or science quotes in it. Quotes you had to memorize, without knowing the topics.
None of the topics resonated with you? Tough luck, buddy.
Don't have any 'novel' ideas? Figure something out!
Cited something the reader didn't know about? Should've known better!
Quoted someone the person grading doesn't know? Better hope they're in a good mood!
There's a reason they got rid of this style of exam a few years later. Doesn't change that this one thing dominated my native language and literature classes.
My biggest takeaway from it is how much I hate my native language class and literature class, but that might've happened even without the essay.
It's been obvious since ChatGPT blew up in early 2023 that educators had to rethink how they educate.
I agree that this situation that the author outlines is unsatisfactory but it's mostly the fault of the education system (and by extension the post author). With a class writing exercise like the author describes, of course the students are going to use an LLM, they would be stupid not to if their classmates are using it.
The onus should be on the educators to reframe how they teach and how they test. It's strange how the author can't see this.
Universities and schools must change how they do things with respect to AI, otherwise they are failing the students. I am aware that AI has many potential and actual problems for society but AI, if embraced correctly, also has the potential to transform the educational experience in positive ways.
> they would be stupid not to if their classmates are using it.
Why would they be stupid? Were people before LLMs stupid for not asking smarter classmate/parent/paid contractor to solve the homework for them?
Large part of education is learning about things that can be easily automated, because you can't learn hard things without learning easy things. Nothing conceptually changed in this regard, like Wolfram Alpha didn't change the way differentiation is taught.
Agreed, my bad choice of words. I really meant 'stupid' from a slightly ironic, competitive point of view. It's like the pressure to cheat in professional sport is obviously so intense, I'm sure a lot of a cheat's motivation is to remain competitive because their colleagues are cheating so they feel they have to as well, otherwise they lose.
YMMV, in my uni there is more or less zero impact of cheating in your homeworks in most classes: the homework has low weight in final grade and are graded very generously.
I agree that making assignments not designed with external sources in mind significantly impact the final grade is not ideal. I think this is minor and easily fixable point rather than some failure of the whole education system.
What's funny about this is if you are an athlete who cheats constantly while growing up - you'll never develop the skills to be able to make it as a pro. Interesting how students don't see this same situation with their homework
> Were people before LLMs stupid for not asking smarter classmate/parent/paid contractor to solve the homework for them?
In American universities where your GPA from your in-class assessments forms part of your final grade? Yes, absolutely.
Where I came from you do your learning in class and your assessment in a small, short set of exams (and perhaps one graded essay) at the end of each year. That seems far more conducive to learning things without having to juggle two competing objectives the whole time.
Amusingly, when I asked o3 to propose changes to the education system which address the author's complaints wrt writing assignments, one of the first things it suggested was transparent prompt logging (basically what the author proposes).
It seems to me that the little sister is acting rationally. Yes it's a bit lazy but then: when was the last time you used a calculator for a calculation that you could have done in your head? Do you actually need to be able to do mental arithmetic, do you actually really need to be able to work out what 24 + 7 is in your head?
I don't know what the answer is. I'm old school, if it was up to me I'd bring back slide rules and log tables, because that's such a visual and tactile way of getting to know mathematics and numbers.
It's interesting to consider how AI is affecting humans' cognition skills. Is it going to make us stupid or free us up to use our mental capacities for higher level activities? Or both?
One thing missing from the reddit story is how the author knows their sister is using ChatGPT to obtain the answer as opposed to validating work. I'll often punch a calculation I've already done in my head into a calculator to confirm for myself I landed on the right answer. Or come up with a solution to a problem, find myself wondering if there's a better solution or just feel like it doesn't quite look right and search for solutions online to compare and contrast. I'm not using the calculator or search to replace my work, I'm using it as the sniff test. But you couldn't determine that just by knowing my inputs to the calculator and google. You'd have to see the entire process.
24 plus 7 is 41 or 94 or 3 and 4/7 are all obviously wrong, but only if you have developed number sense. I've had students who wouldn't bat an eye at any of those solutions.
Only in a few fields. You can have a successful career publishing papers in the social sciences, or justifying decisions in middle-management, without being able to know whether an effect size passes the sniff test - actually not knowing will probably help you make convenient mistakes.
Why isn't both a valid option, can't one be at a university to both learn and get a degree+GPA showing they did well at doing so? In any case, why does a student selecting that they are there to learn negate the responsibility of the university to provide the best curricula for students to do so with?
The assumption you are making is that students consistently using llms is required by the best curriculum. That’s a big assumption. There may be value educationally in forcing students to learn how to do things on their own that an llm could do for them. And in that case, it’s the student failing their education if they circumvent it with an llm, not the institution’s.
At times there is value in that kind of approach, especially when the learning is specifically about those kinds of lower layers instead of higher concepts. As such, it's not that certain tools should be used in every assignment or never ever used, just whether they should be used commonly.
For most continued learning it's better if the university uses calculators, compilers, prepared learning materials, and other things that do stuff on behalf of the students instead of setting the bar permanently to "the student should want to engage everything at a base level or they must not be here to learn". It allows much more advanced learning to be done in the long run.
Universities do not provide knowledge, this is a romanticization of an ideal. They are about getting a passing grade and a certificate so that you can enter the workforce. The idea that you go to college to get an education is just a polite fiction to appease students and their parents
Seems like you missed the point of the article. The author is saying that if you treat the class/lesson as a means to an end only where the goal is to get a diploma then you’re not actually getting an education. If you’re using an LLM to do the work for you even if the other students are too then you’re just following all the other lemmings off the ledge.
> With a class writing exercise like the author describes, of course the students are going to use an LLM, they would be stupid not to if their classmates are using it.
Its only stupid if you try to optimize for the wrong things (finishing quickly, just getting a pass).
I'd say it's very smart if you don't rely on LLMs, copy the homework from someone else, or similar; because you're optimizing for learning, which will help you more than the various shortcuts.
Yeah I’ve been explaining something similar to people. Go ahead and spend your time using the Llm but if you think no one notices you’re wrong. Your colleagues/teammates notice and when it comes time to test your real mettle you will end up last in line because no one has confidence that you can do anything but fake your results. This will hurt you when you want a promotion or raise or recognition of some sort of your efforts.
> Students who know least of all and don't understand the purpose of writing or problem solving or the limitations of LLMs are currently wasting years of their lives
Exactly. I tend to think that the role of a teacher is to get the students to realise what learning is all about and why it matters. The older the students get, the more important it is.
The worst situation is a student finishing university without having had that realisation: they got through all of it with LLMs, and probably didn't learn how to learn or how to think critically. Those who did, on the other hand, didn't need the LLMs in the first place.
"Anyone who actually struggles to solve problems and learn themselves is going to have massive advantages in the long term."
Looking forward towards is, but I fear that might be wishful thinking in part.
Also pre LLMs I have seen too many deep thinkers fail and pretenders succed. I don't see how LLMs can change that. Unless we all collectivly grow tired if pretenders and fakers amd value deep understanding. I just see not many indication of that.
I agree wholeheartedly with your last point. Pretenders are so easily seen by word choice and phrasing alone and yet these valueless additions are accepted and not immediately called out because it's not polite to call out. For example: when minimizing or distancing themselves from their failure or associating themselves with someone else's success.
There should be zero tolerance for these types of behaviors in my opinion. I see zero evidence of these behavior even being identified by most, let alone any thought on calling them out or stopping them.
You've helped me realize that folks using llms in replace of learning to write themselves are almost certainly giving up all thought of nuance on a topic and are, without realizing it, letting the llm either ignore or add nuance based on its training data and random chance.
Hopefully the pendulum will swing the other way and there's a public epiphany but given the loss of nuance over the decades I'm not betting on it
> It's been incredibly blackpilling seeing how many intelligent professionals and academics don't understand this
I figured this out in high school. It can’t be all that uncommon of a thought that if you are already in school and paying and given time to learn, you might as well do so?
I think that figuring this out is a great achievement. Probably one of the goals of school. It depends on many factors and the sooner, the better.
Young kids don't get it, they just do what they're asked. That's okay. University students graduating without having figured it out is a problem. And somewhere in the middle is when the average student gets there, hopefully?
An analogy I've heard is that it's like using a forklift at the gym. The point is not to get an object from point A to point B, it's to develop skills.
A non-STEM friend got to study/research at really great places across India, Germany, UK, USA, and China for her UG, MA, MPhil, PhD, and postdoc (including lots of field/on ground work). She said that she noted one common theme among her seniors (guides, mentors, colleagues, professors etc) -> that some bad, shit, low quality, whatever published paper is >>>>>>>>>>> high quality work/research but not published or slow to publish (and often for good reason). And she added these were not run of the mill profs who needed those to survive professionally or academically but the likes whose names/reputation often matter more than their departments by many times and don’t need survival at all anymore.
I always get triggered when people argue against „rote memorization” - but it also is technique that builds up knowledge, skills and experience.
Even if one won’t need that specific know how after exams - just realization how much one can memorize and trying out some approaches to optimize it is where people grow/learn.
Coming from the other side of this argument: In my degree, rote memorization was required for a surprising amount of courses. It required students, me included, to memorize huge quantities of things we knew were utterly irrelevant to anything but being graded. (This prediction remained true). Committing irrelevant course work into memory over and over again almost burned me out, certainly made me lose all interest and fun in learning for over a decade afterwards. To be honest, I still feel slightly burned and that might never go away.
You might have attended a good degree, where the learned information was actually beneficial. But I'd bet for most degrees out there, rote memorization is the consequence of professors wanting easily gradable exams, existing for their benefit, not the students.
Which means the actual problem is low quality education and degrees and we might find common ground here.
Memorizing things is somewhat helpful but being able to parrot back answers to questions is not at all the same thing as knowledge, skills, or experience. Memorizing a bunch of facts is an adequate way to fool someone into thinking you have those things. Testing for memorized facts is a good way to misidentify useful skills.
And yet there is still value in facts. For example, learning the ideal gas law in high school has enabled me to understand in my adult life how gases will react when compressed or when released. Yet the ideal gas law is a simple fact that you can memorize in minutes. In this case, memorizing that small fact about physics enabled me to apply it to gain understanding of other situations.
I believe that the same holds true for other facts one might memorize. Yes, the fact may seem like meaningless trivia (and might even be so at times), but in the right situation knowing that fact can help with understanding. You can certainly spend too much time on memorization of facts, but that doesn't mean it has no place either.
A law is an infinite family of facts, it's already several steps up from memorising a concrete fact like the density of oxygen at room temperature and pressure.
I avree with starting from the third paragraph. However I disagree about what you said on academics. The few I interrogated on the topic already adapted their practice, notably changing the kind of homework they give, and changing exam format.
My students however don't understand that the importance is on the process, not the result. My colleagues do.
Using llm’s for papers does not mean your brain is atrophying though. There are lots of ways to challenge the mind even if you use llm’s to write some papers.
Sure. And there are new pedagogies that educators are trying out that help people learn even in the presence of these tools.
But a huge amount of "ugh I'm too smart for this assignment" complaining that students do is just kids being immature rather than an honest attempt at learning through other means.
> Using llm’s for papers does not mean your brain is atrophying though.
It means that you are losing your time. If you are a university student and use LLMs for your classes while "challenging your mind" for stuff outside of class, maybe you should just not be studying there in the first place.
Yes, I too enjoy that universities never send students to pointless filler courses. I really enjoyed writing long essays about ethics in software engineering for people who'd barely even read them. I especially benefited from being told that "tell your boss, tell HR, then make a police report and quit if that didn't fix it" is not an appropriate response to being asked to break the law at work.
While we're talking about things we're grateful for, I am so glad that we've structured the education and employment systems such that not having a degree puts you at significant risk of unemployment, prevents you from ever immigrating anywhere for the first decade of your working life, and generally marks you as a failure.
> Using llm’s for papers does not mean your brain is atrophying though. There are lots of ways to challenge the mind even if you use llm’s to write some papers.
Writing is hard. Sometimes it means sitting with yourself, for hours, without any progress. Leaning on an LLM to ease through those tough moments is 100% short circuiting the learning process.
To your point, maybe you're learning something else instead, like when/how to prompt an LLM or something. But you're definitely not learning how to write. Whether that's relevant is a separate discussion.
Lately I've been saying often the phrase "the process is the product." When you outsource the process, then the product will be fundamentally different from what you would have delivered on your own. In my own case of knowledge work, the value of the reports I write is not in the report itself (nobody ever reads them...) but rather the thinking that went into them and the hard-won wisdom and knowledge we created in our heads.
> Leaning on an LLM to ease through those tough moments is 100% short circuiting the learning process.
Sounds like "back in my days" type of complaining. Do you have any evidence of this "100% reduction" or is it just "AI bad" bandwagoning?
> But you're definitely not learning how to write.
How would you know? You've never tested him. You're making a far-reaching assumption about someone's learning based on using an aid. It's the equivalent of saying "you're definitely not learning how to ride a bicycle if you use training wheels".
If they use LLM for writing papers, they probably use it for other things as well. I have seen so many instances of adult actually skipping the step of "whys" and "whats" and go straight to "ask the LLM and we trim backwards".
Its basically adults producing texts of slop messages to each other. It is actually atrophying.
You might be in a circle of people that wants to know "why" things work. For example, when there's a bug, we go through several processes of:
There's a bug...why does it happen? What were they thinking when they wrote this? How to prevent this from happening?
This is true even for simple bugs, but nowadays you just vibe code your away into the solution, asking the AI to fix it over and over without ever understanding how it works.
Perhaps its just the way things are. I mean who uses their head to do calculations nowadays? Who knows how to create a blurring effect in physical drawing?
If you used a wheelchair every day, your legs would atrophy.
Regardless of the existence of other ways to exercise your legs which you also will not do, because you're a person with working legs who chooses to use a wheelchair.
You’re describing an extreme so let me counter with a different “what if” extreme: If
, every day, you use a wheel chair for five minutes and run for several hours followed by 200 squats then your legs will not atrophy. My point is that writing papers is only one way to work your mind and using llm’s for this does not indicate how you use your mind overall.
knowledge gap and consolidation just gets wider, like before. the internet rotted peoples brains but also brought that knowledge to people who would have never had the opportunity to learn.
those who use the tools to accelerate their learning will do so and others who use it just to get by will see their skills atrophy and become irrelevant.
You can’t reason through this problem. If there’s homework due, teenagers are going to use LLMs. The ones that don’t are going to work very hard just to end up with at par work with everyone else. The only solution is returning to a paradigm of heavy in class testing.
I think this is going to be a problem as long as the metrics by which we evaluate students are intolerant of human errors and prioritize right answers over thinking skills.
The current situation is that people need to pass exams, get certain GPA's, etc. to have opportunities unlocked to them. Education today is largely about collecting these "stamps" that open doors and not about actual learning.
Actually in real world AIs are doing great and continuosly accelerating and augmenting human efforts to solve open ended problems. Academic institutions should help students embrace AI tools and not just penalize them for using it.
Someday, LLMs are going to cost actual serious money to use and people will have “LLM bills” to pay every month just to have some semblance of a brain at work. Maybe a portion of salary can just be deducted directly from a paycheck to pay for LLM tools.
I can run the Qwen 3 0.6B model directly on my 3-year-old phone, and it can rewrite text and help me clearly explain my views. Even with new models possibly being less open and subsidized options drying up, we still have useful open models available for free on consumer hardware.
In real life, many education professionals are co-conspirators with students to produce work artifacts in lieu of actual work, ChatGPT has done wonders for both parties to help accelerate production of piles and piles of paper that have the appearance of schoolwork, paper piles > learning is the nash equilibrium of the incentives in our present education scheme in the United States.
But don't worry, worst case scenario, all of the kids growing up in this environment that are actually learning will build structures to exploit the prompters, I suspect the present situation where prompters can accidentally find themselves in real jobs is transient and building better filters will become survival imperative for businesses and institutions.
> how many intelligent professionals and academics don't understand this
Mastery of a discipline does not imply any pedagogical knowledge, despite anything one of my childhood heroes, Richard Feynman, might have claimed.
Despite frequent claims otherwise, in my experience and sampling of PhDs and Masters of different sorts and grad students working toward those degrees, an advanced degree does not teach anyone how to lead or teach. This is true of even some of the folks I knew studying Education itself who were a little too focused on their own research to understand anything "so simple."
> cheat themselves out of an education
What's "an education," though? For some people, education is focused on how to learn. For others, it's focused on some kind of certification to get a job. Some of us see value in both. And I'm sure there are other minority opinions as well. We, as a society, can't agree. The only thing we can seem to agree on in the US is that college should be expensive and saddle students with ridiculous debt.
> I believe that the main reason a human should write is to communicate original thoughts.
in fairness to the students, how does the above apply to school work?
why does a student write, anyway? to pass an assignment, which has nothing to do with communicating original thoughts-- and whose fault is that, really?
education is a lot of paperwork to get certified in the hopes you'll get a job. it's as bereft of intelectual life as the civil service examinations in imperial china. original thought doesn't enter the frame.
This is equivalent to students using AI to complete computer programming assignments. They misconstrue the purpose of an assignment as just one of generating output instead something to teach the principles and techniques they'll require later if they want a job in the profession. While they may believe they're fooling the teacher, all they're really doing is fooling, and cheating, themselves.
Whether it be writing or computer programming, or exercising, for that matter, if you aren't willing to put in the work to achieve your goals, why bother?
> I have never seen any form of create generative model output (be that image, text, audio, or video) which I would rather see than the original prompt.
I've used LLM before to document command-line tools and APIs I've made; they aren't the final product since I also tweaked the writing and fixed misunderstandings from the LLM. I don't think the author would appreciate the original prompts, where I essentially just dump a lot of code and give instructions in bullet point form on what to output.
These generated documentation are immensely useful, and I use them all the time for myself. I prefer the documentation to reading the code because finding what I need at a glance is not trivial nor is remembering all the conditions, prerequisites, etc.
That being said, the article seems to focus on a use case where LLM is ill-suited. It's not suited for writing papers to pretend you wrote a paper.
> I say this because I believe that your original thoughts are far more interesting
Looking at the example posted, I'm not convinced that most people's original thoughts on gimbal lock will be more interesting than a succinct summary by an LLM.
I've already asked a number of colleagues at work producing insane amount of gibberish with LLMs to just pass me the prompt instead: if LLM can produce verbose text with limited input, I just need that concise input too (the rest is simply made up crap).
I'm far from the first to make this observation but LLMs are like anti-compression algorithms when used like that, a simple idea gets expanded into a bloated mess by an LLM, then sent to someone else who runs it through another LLM to summarize it back to something approximating the original prompt. Nobody benefits aside from Sam Altman and co, who get to pocket a cool $0.000000001 for enabling this pointless exercise.
There is one other very useful form of "expansion" that LLMs do.
If you aren't aware: (high-parameter-count) LLMs can be used pretty reliably to teach yourself things.
LLM base models "know things" to about the same degree that the Internet itself "knows" those things. For well-understood topics — i.e. subjects where the Internet contains all sorts of open-source textbooks and treatments of the subject — LLMs really do "know their shit": they won't hallucinate, they will correct you when you're misunderstanding the subject, they will calibrate to your own degree of expertise on the subject, they will make valid analogies between domains, etc.
Because of this, you can use an LLM as an infinitely-patient tutor, to learn-through-conversation any (again, well-understood) topic you want — and especially, to shore up any holes in your understanding.
(I wouldn't recommend relying solely on the LLM — but I've found "ChatGPT in one tab, Wikipedia open in another, switching back and forth" to be a very useful learning mode.)
See this much-longer rambling https://news.ycombinator.com/item?id=43797121 for details on why exactly this can be better (sometimes) than just reading one of those open-source textbooks.
I use LLMs a huge amount in my work as a senior software engineer to flesh out the background information required to make my actual contributions understandable to those without the same background as me. eg, if I want to write a proposal on using SLO's and error budgets to make data driven decisions about which errors need addressing and which don't, inside a hybrid kubernetes and serverless environment, I could do a few things:
* Not provide background information and let people figure it out for themselves. This will not help me achieve my goals.
* Link them to Google's SRE book and hope they read it. Still not achieving my goals, because they won't.
* Spend 3 hours writing the relevant background information out for them to read as part of my proposal. This will achieve my goals, but take an extra 3 hours.
* Tell the LLM what I'm looking for and why, then let it write it for me in 2 minutes, instead of 3 hours. I can check it over, make sure it's got everything, refine it a little, and I've still saved 2.5 hours.
So for me, I think the author has missed a primary reason people use LLMs. It saves a bunch of time.
For a teacher, they want to see you spend those 3h to see what you come up with, and if there's something they should direct your attention to, or something they should change in their instruction.
But ultimately, getting the concise summary for a complex topic (like SLIs and SLOs are) is brilliant, but would be even better if it was full of back-links to deeper dives around the Internet and the SRE book.
It feels like the information is there strewn across the internet, in forums, Reddit posts, stack overflow, specs, books. But to trawl though it all was so time consuming. With an LLM you can quickly distill it down to just the information you need.
Saying that, I do feel like reading the full spec for something is a valuable exercise. There may be unknown unknowns that you can't even ask the LLM about. I was able to become a subject expert in different fields just but sitting down and reading through the specs / RFCs, while other colleagues continued to struggle and guess.
This is how I currently am relearning upper high school math. It’s tremendously helpful as I am a why guy.
Why is the angle called m? Why is a combination nPr * (1/r)? What is 1/r doing there?
I use mathacademy.com as my source of practice. Usually that’s enough but I tend to fall over if small details aren’t explained and I can’t figure out why those details are there.
In high school this was punished. With state of the art LLMs, I have a good tutor.
Also it’s satisfying to just upload a page in my own handwriting and it understands what I did, and is able to correct me there.
The fact that we see lower grades as "punishment" is the root of the problem: grades are an assessment of our understanding level and competence on a topic.
Now, I know psychologically it's not as simple, and both society and ourselves equate academic (and professional, later on) success with personal worth, but that's a deeper, harder topic.
They are also useful for association. Imagine an LLM trained on documentation. Then you can retrieve info associated with your question.
This can go beyond just specific documentation but also include things like "common knowledge" which is what the other poster meant when they talked about "teaching you things".
In my experience, if the information I need is in the documentation, then I don't need the LLM. If it is not in the documentation, then the LLM will invent stuff that could be there but that isn't, and it's actually a loss of time.
> In my experience, if the information I need is in the documentation, then I don't need the LLM
need is a strong word. Do you need to be able to do ctrl+F? Not really, you can just read it all. But maybe it's easier to do ctrl+F. Same with LLM. Just imagine it as a fuzzy ctrl+F. Can be useful.
Rituals are significant because they are long. A ritual that consisted of the words “Rain please” wouldn’t convince the gods, much less their human followers.
Depends on what you are looking for. I’ve turned half baked ideas into white papers for plenty of praise. I’ve used them to make my Jira tickets seem complicated and complete. I’ve used them to get praised for writing comprehensive documentation.
Part of my performance review is indirectly using bloat to seem sophisticated and thorough.
Documentation is an interesting use case. There are various kinds of documentation (reference, tutorial, architecture, etc.) and LLMs might be useful for things like
- repetitive formatting and summarization of APIs for reference
- tutorials which repeat the same information verbosely in an additive, logical sequence (though probably a human would be better)
- sample code (though human-written would probably be better)
The tasks that I expect might work well involve repetitive reformatting, repetitive expansion, and reduction.
I think they also might be useful for systems analysis, boiling down a large code base into various kinds of summaries and diagrams to describe data flow, computational structure, signaling, etc.
Still, there is probably no substitute for a Caroline Rose[1] type tech writer who carefully thinks about each API call and uses that understanding to identify design flaws.
Yes, but none of the current LLMs are even remotely useful doing that kind of work for even something moderately complex. I have a 2k LOC project that no LLM even "understands" *. They can't grasp what it is: It's a mostly react-compatible implementation of "hooks" to be used for a non-DOM application. Every code assistant thinks it's a project using React.
Any documentation they write at best re-states what is immediately obvious from the surrounding code (Useless: I need to explain why), or is some hallucination trying to pretend it's a React app.
To their credit they've slowly gotten better now that a lot of documentation already exists, but that was me doing the work for them. What I needed them to do was understand the project from existing code, then write documentation for me.
Though I guess once we're at the point AI is that good, we don't need to write any documentation anymore, since every dev can just generate it for themselves with their favorite AI and in the way they prefer to consume it.
* They'll pretend they understand by re-stating what is written in the README, then proceed to produce nonsense.
This kind of "perf review" hacking works ~everywhere; how well it works correlates with how entrenched the organization is (i.e., how hard it is for new players to disrupt).
You don't have to play the game the same way to work there. But it helps to accept that others will play it, and manage your own expectations accordingly.
Which big, well-paying companies do not have "those bullshit metrics"? I know for a fact that Meta, Google, Stripe, Airbnb, and Oracle all lean heavily on performance review cycles based entirely on ridiculous metrics. Getting ahead there requires you to play the stupid games GP is suggesting.
Feel for you or anyone surrounded by such others but it is most definitely not everywhere - that is used to justify your presence in a place of work you should not be
> creating bloat until it becomes so unproductive it kills the host
Maybe we should let/encourage this to happen. Maybe letting bloated zombie-like organisations bloat themselves to death would thin the herd somewhat, to make space for organisations that are less “broken”.
I am on the mind that every organization should eventually die before it becomes a monster. I am also not a huge fan of inheritance for the same reason.
I fully believe you and I am saddened by the reality of your situation.
At the same time, I strive really hard to influence the environment I am in so it does not value content bloat as a unit of productivity, so hopefully there are at least some places where people can have their sanity back!
If your organisation is such that you have to do this even though you are competent for your job, then they deserve it. They lose money because they do it wrong.
If your organisation is functional and you are abusing it by doing that, then you deserve to get fired.
...thinking about it, there are probably situations where making something more verbose makes it take less effort to read. I can see how an LLM might be useful in that situation.
Something I’ve found very helpful is when I have a murky idea in my head that would take a long time for me to articulate concisely, and I use an LLM to compress what I’m trying to say. So I type (or even dictate) a stream of consciousness with lots of parentheticals and semi-structured thoughts and ask it to summarize. I find it often does a great job at saying what I want to say, but better.
(See also the famous Pascal quote “This would have been a shorter letter if I had the time”).
P.s. for reference I’ve asked an LLM to compress what I wrote above. Here is the output:
When I have a murky idea that’s hard to articulate, I find it helpful to ramble—typing or dictating a stream of semi-structured thoughts—and then ask an LLM to summarize. It often captures what I mean, but more clearly and effectively.
Agreed, the messiness of the original text has character and humanity that is stripped from the summarized text. The first text is an original thought, exchanged in a dialogue, imperfectly.
Elsewhere in this comment section, it's discussed about the importance of having original thought, which the summarized text specifically isn't, and has leeched away.
The parent comment has actually made the case against the summarized text being "better" (if we're measuring anything that isn't word count).
Learning to articulate your thoughts is pretty vital in learning to think though.
An LLM could make something sound articulate even if your input is useless rambling containing the keywords you want to think about. Having someone validate a lack of thought as something useful doesn't seem good for you in the long term
Your original here is distinctly better! It shows your voice and thought patterns. All character is stripped away in the "compressed" version, which unsurprisingly is longer, too.
The one case where this doesn't work, is if the prompt is, say 3 ideas, which the LLM expand to 20, and the colleague then trimmed down to 10.
Ideally there's some selection done, and the fact you're receiving it means it's better than a mean answer. But sometimes they haven't even read the LLM output themselves :-(
This is how I've felt about using LLMs for things like writing resumes and such. It can't possibly give you more than the prompt since it doesn't know anything more about you than you gave it in the prompt.
It's much more useful for answering questions that are public knowledge since it can pull from external sources to add new info.
Recently I wasted half a day trying to make sense of story requirements given to me by a BA that were contradictory and far more elaborate than we had previously discussed. When I finally got ahold of him he confessed that he had run the actual requirements through ChatGPT and "didn't have time to proofread the results". Absolutely infuriating.
Chatgpt very useful for adding softness and politeness to my sentences. Would you like more straight forward text which probably will be rude for regular american?
If we can detach content and presentation, then the reader can choose tone and length.
At some point we will stop making decisions about what future readers want. We will just capture the concrete inputs and the reader's LLM will explain it.
I don't think form and function can be separated so cleanly in natural language. However you encode what's between your ears into text, you've made (re)presentational choices.
A piece of text does not have a single inherently correct interpretation. Its meaning is a relation constructed at run- (i.e. read-)time between the reader, the writer, and (possibly) the things the text refers to, that is if both sides are well enough aligned to agree on what those are.
My LLM workflow involves a back-and-forth clarification (verbose input -> LLM asks questions) that results in a rich context representing my intent. Generating comments/docs from this feels lossy.
What if we could persist this final LLM context state? Think of it like a queryable snapshot of the 'why' behind code or docs. Instead of just reading a comment, you could load the associated context and ask an LLM questions based on the original author's reasoning state.
Yes, context is model-specific, a major hurdle. And there are many other challenges. But ignoring the technical implementation for a moment, is this concept – capturing intent via persistent, queryable LLM context – a valuable problem to solve? I feel it would be.
I like the author's take: it isn't a value judgement on the individual using ChatGPT (or Gemini or whichever LLM you like this week), it's that the thought that went into making the prompt is, inevitably, more interesting/original/human than the output the LLM generates afterwards.
In my experiments with LLMs for writing code, I find that the code is objectively garbage if my prompt is garbage. If I don't know what I want, if I don't have any ideas, and I don't have a structure or plan, that's the sort of code I get out.
I'd love to hear any counterpoints from folks who have used LLMs lately to get academic or creative writing done, as I haven't tried using any models lately for anything beyond helping me punch through boilerplate/scaffolding on personal programming projects.
This is the CRUX of the issue. Even with SOTA models (Sonnet 3.5, etc) - the more open-ended your prompt - the more banal and generic the response. It's GIGO turtles all the way down.
I pointed this out a few weeks ago with respect to why the current state of LLMs will never make great campaign creators in Dungeons and Dragons.
We as humans don't need to be "constrained" - ask any competent writer to sit quietly and come up with a novel story plot and they can just do it.
That being said - they can still make AMAZING soundboards.
And if you still need some proof, crank the temperature up to 1.0 and pose the following prompt to ANY LLM:
Come up with a self-contained single room of a dungeon that involves an
unusual puzzle for use with a DND campaign. Be specific in terms of the
puzzle, the solution, layout of the dungeon room, etc. It should be totally
different from anything that already exists. Be imaginative.
I guarantee 99% of the returns will return a very formulaic physics-based puzzle response like "The Resonant Hourglass", or "The Mirror of Acoustic Symmetry", etc.
When using Claude Sonnet 3.7 for coding, I often find that constraints I add to the prompt, end up producing unintended side effects.
Some examples:
- "Don't include pointless comments." - The model doesn't keep track of what it's doing as well, I generally just do another pass after it writes the code to simplify things.
- "Keep things simple" - The model cuts corners(often unnecessarily) on things like type safety.
- "Allow exceptions to bubble up" - Claude deletes existing error handling logic. I found that Claude seems to prefer just swallowing errors and adding some logging, instead of fixing the underlying cause of the error, but adding this to the prompt just caused it to remove the error handling that I had added myself.
> I guarantee 99% of the returns will return a very formulaic physics-based puzzle response like "The Resonant Hourglass"
Haha, I was suspicious, so I tried this, and I indeed got an hourglass themed puzzle! Though it wasn't physics-based - characters were supposed to share memories to evoke emotions, and different emotions would ring different bells, and then you were supposed to evoke a certain type of story. Honestly, I don't know what the hourglass had to do with it.
The room is a simple 30-foot square with a single exit door that's currently sealed. In the center sits a large stone cube (roughly 5 feet on each side) covered in various textured surfaces - some rough like sandpaper, others smooth as glass, some with ridged patterns, and others with soft fabric-like textures.
Around the room, six distinct scent emitters are positioned, each releasing a different aroma (pine, cinnamon, ocean breeze, smoke, floral, and citrus). The room is otherwise empty except for a small stone pedestal near the entrance with a simple lever.
## The Puzzle Concept
This puzzle operates on "synesthetic translation" - converting sensory experiences across different senses. The core concept is entirely verbal and tactile, making it fully accessible without visual components.
## How It Works
When players pull the lever, one of the scent emitters activates strongly, filling the room with that particular aroma. Players must then approach the central cube and touch the texture that corresponds to that smell according to a hidden synesthetic logic.
The connection between smells and textures follows this pattern:
- Pine scent → ridged texture (like tree bark)
- Cinnamon → rough, granular texture (like spice)
- Ocean → smooth, undulating surface (like waves)
- Smoke → soft, cloudy texture (like mist)
- Floral → velvet-like texture (like petals)
- Citrus → bumpy, pitted texture (like orange peel)
After correctly matching three smell-texture pairs in sequence, the door unlocks. However, an incorrect match causes the lever to reset and a new random smell to emerge.
## Communication & Accessibility
The DM describes the smells verbally when they're activated and can describe the various textures when players explore the cube by touch. The entire puzzle can be solved through verbal description, touch, and smell without requiring sight.
For extra accessibility, the DM can add:
- Distinct sounds that play when each scent is released
- Textured surfaces that have subtle temperature differences
- Verbal clues discovered through successful matches
## What Makes This Unique
This puzzle uniquely relies on cross-sensory associations that aren't commonly used in dungeons. It:
- Doesn't rely on visuals at all
- Uses smell as a primary puzzle component (rare in D&D)
- Creates unusual connections between different senses
- Has no mathematical, musical, or traditional riddle elements
- Can be experienced fully regardless of vision status
- Creates interesting roleplaying opportunities as players discuss how different scents "feel" texturally
For the DM, it's easy to describe and implement while still being conceptually unique. Players solve it through discussion, exploration, and experimentation rather than recalling common puzzle patterns.
It is totally different from anything that exists. It fulfils the prompt, I suppose! It has to be crazy so you can be more certain it's unique. The prompt didn't say anything about it being good.
Yeah this was the part I found a little silly, mostly because I just couldn't visualize what that mesh looked like or how I would describe how to operate it.
I have mixed feelings. Generally I don’t think that LLM output should be used to create anything that a human is supposed to read, but I do carve out a big exception for people using LLMs for translation/writing in a second language.
At the same time, however, the people who need to use an LLM for this are going to be the worst at identifying the output’s weaknesses, eg just as I couldn’t write Spanish text, I also couldn’t evaluate the quality of a Spanish translation that an LLM produced. Taken to an extreme, then, students today could rely on LLMs, trust them without knowing any better, and grow to trust them for everything without knowing anything, never even able to evaluate their quality or performance.
The one area that I do disagree with the author, though, is coding. As much as I like algorithms code is written to be read by computers and I see nothing wrong with computers writing it. LLMs have saved me tons of time writing simple functions so I can speed through a lot of the boring legwork in projects and focus on the interesting stuff.
I think Miyazaki said it best: “I feel… humans have lost confidence“. I believe that LLMs can be a great tool for automating a lot of boring and repetitive work that people do every day, but thinking that they can replace the unique perspectives of people is sad.
I actually feel very strongly that code is very much written for us humans. Sure, it's a set of instructions that is intended to be machine read and executed but so much of _how_ code is written is very much focused on the human element that's been a part of software development. OOP, design patterns, etc. don't exist because there is some great benefit to the machines running the code. We humans benefit as the ones maintaining and extending the functionality of the application.
I'm not making a judgement about the use of LLMs for writing code, just that I do think that code serves the purpose of expressing meaning to machines as well as humans.
In my experience Gemini can be really good at creative writing, but yes you have to prompt and edit it very carefully (feeding ideas, deleting ideas, setting tone, conciseness, multiple drafts, etc).
I use Gemini pretty much exclusively for creative writing largely because the long context lets you fit an entire manuscript plus ancillary materials, so it can serve as a solid beta reader, and when you ask it to outline a chapter it is very good at taking the events preceding and following into account. It's hard to overstate the value of having a decent beta reader that can iteratively review your entire work in seconds.
As a side note, I find the way that you interact with a LLM when doing creative writing is generally more important than the model. I have been having great results with LLMs for creative writing since ChatGPT 3.5, in part because I approach the model with a nucleus of a chapter and a concise summary of relevant details, then have it ask me a long list of questions to flesh out details, then when the questions stop being relevant I have have it create a narrative outline or rough draft which I can finish.
Interesting. I think I'm a better editor so I use it as a writer, but it makes sense that it works the other way too for strong writers. Your way might even be better, since evaluating a text is likely easier than constructing a good text (Which is why your process worked even back with 3.5).
I have a horrible time editing my own work. Decision paralysis and what not, but I did have the idea that a good way to practice would be editing the content of LLM generated fictional narratives. I think the point that many are making that LLMs are useful as cognitive aids that augment thinking rather than replacements for thinking. They can be used to train your mind by inspiring thoughts you wouldn't have came up with on your own.
Sometimes, good writing is like an NP-complete problem, hard to create, but easy to verify. If you have enough skill to distinguish good output from garbage, you can produce reasonably good results.
> Sometimes, good writing is like an NP-complete problem, hard to create, but easy to verify.
Doesn’t this match pretty much all human creation? It’s easier to judge a book that to write it, it’s easier to watch a rocket going up in the space than to build it, it’s easier to appreciate some Renaissance painting or sculpture than to actually make it.
For creative and professional writing, I found them useful for grammar and syntax review, or finding words from a fuzzy description.
For the structure, they are barely useful: Writing is about having such a clear understanding, that the meaning remains when reduced to words, so that others may grasp it. The LLM won't help much with that, as you say yourself.
LLMs may seem like magic buy they aren't. They operate within the confines of the context they're given. The more abstract the context, the more abstract the results.
I expect to need to give a model at least as much context as a decent intern would require.
Often asking the model "what information could I provide to help you produce better code" and then providing said information leads to vastly improved responses. Claude 3.7 sonnet in Cline is fairly decent at asking for this itself in plan mode.
More and more I find that context engineering is the most important aspect of prompt engineering.
> I'd love to hear any counterpoints from folks who have used LLMs lately to get academic or creative writing done
They’re great at proofreading. They’re also good at writing conclusions and abstracts for articles, which is basically synthesising the results of the article and making it sexy (a task most scientists are hopelessly terrible at). With caveats:
- all the information needs to be in the prompt, or they will hallucinate;
- the result is not good enough to submit without some re-writing, but more than enough to get started and iterate instead of staring at a blank screen.
I want to use them to write methods sections, because that is basically the exact same information repeated in every article, but the actual sentences need to be different each time. But so far I don’t trust them to be accurate with technical details. They’re language models, they have no knowledge or understanding.
Point two is critical. I have found that the best way for me is to avoid using copy-and-paste. Instead, I put the browser on the right corner of the screen and my text editor on the left, then transcribe the text word by word by typing it using the keyboard. In this way, my natural laziness is less likely to accept words, expressions, and sentences that are perhaps okay-ish but not 100% following my taste.
I use an LLM to brainstorm for a creative writing project. Mostly I ignore its suggestions! but, somehow having the chatter helps me see what I am trying to say
I think the author has a fair take on the types of LLM output he has experience with, but may be overgeneralizing his conclusion. As shown by his example, he seems to be narrowly focusing on the use case of giving the AI some small snippet of text and asking it to stretch that into something less information-dense — like the stereotypical "write a response to this email that says X", and sending that output instead of just directly saying X.
I personally tend not to use AI this way. When it comes to writing, that's actually the exact inverse of how I most often use AI, which is to throw a ton of information at it in a large prompt, and/or use a preexisting chat with substantial relevant context, possibly have it perform some relevant searches and/or calculations, and then iterate on that over successive prompts before landing on a version that's close enough to what I want for me to touch up by hand. Of course the end result is clearly shaped by my original thoughts, with the writing being a mix of my own words and a reasonable approximation of what I might have written by hand anyway given more time allocated to the task, and not clearly identifiable as AI-assisted. When working with AI this way, asking to "read the prompt" instead of my final output is obviously a little ridiculous; you might as well also ask to read my browser history, some sort of transcript of my mental stream of consciousness, and whatever notes I might have scribbled down at any point.
> the exact inverse of how I most often use AI, which is to throw a ton of information at it in a large prompt
It sounds to me that you don't make the effort to absorb the information. You cherry-pick stuff that pops in your head or that you find online, throw that into an LLM and let it convince you that it created something sound.
To me it confirms what the article says: it's not worth reading what you produce this way. I am not interested in that eloquent text that your LLM produced (and that you modify just enough to feel good saying it's your work); it won't bring me anything I couldn't get by quickly thinking about it or quickly making a web search. I don't need to talk to you, you are not interesting.
But if you spend the time to actually absorb that information, realise that you need to read even more, actually make your own opinion and get to a point where we could have an actual discussion about that topic, then I'm interested. An LLM will not get you there, and getting there is not done in 2 minutes. That's precisely why it is interesting.
You're making a weirdly uncharitable assumption. I'm referring to information which I largely or entirely wrote myself, or which I otherwise have proprietary access to, not which I randomly cherry-picked from scattershot Google results.
Synthesizing large amounts of information into smaller more focused outputs is something LLMs happen to excel at. Doing the exact same work more slowly by hand just to prove a point to someone on HN isn't a productive way to deliver business value.
> Doing the exact same work more slowly by hand just to prove a point to someone on HN isn't a productive way to deliver business value.
You prove my point again: it's not "just to prove a point". It's about internalising the information, improving your ability to synthesise and be critical.
Sure, if your only objective is to "deliver business value", maybe you make more money by being uninteresting with an LLM. My point is that if you get good at doing all that without an LLM, then you become a more interesting person. You will be able to have an actual discussion with a real human and be interesting.
Understanding or being interesting has nothing to do with it. We use calculators and computers for a reason. No one hires people to respond to API requests by hand; we run the code on servers. Using the right tool for the job is just doing my job well.
> We use calculators and computers for a reason. No one hires people to respond to API requests by hand; we run the code on servers
We were talking about writing, not about vibe coding. We don't use calculators for writing. We don't use API requests for writing (except when we make an LLM write for us).
> Using the right tool for the job is just doing my job well.
I don't know what your job is. But if your job is to produce text that is meant to be read by humans, then it feels like not being able to synthesise your ideas yourself doesn't make you excellent at doing your job.
Again maybe it makes you productive. Many developers, for instance, get paid for writing bad code (either because those who pay don't care about quality or can't make a difference, or something else). Vibe coding obviously makes those developers more productive. But I don't believe it will make them learn how to produce good code. Good for them if they make money like this, of course.
> We were talking about writing, not about vibe coding. We don't use calculators for writing. We don't use API requests for writing (except when we make an LLM write for us).
We do however use them to summarize and transform data all the time. Consider the ever present spreadsheet. Huge amounts of data are thrown into spreadsheets and formulas are applied to that data to present us with graphs and statistics. You could do all of that by hand, and you'd probably have a much better "internalization" about what the data is. But most of the time, hand crafting graphs from raw data and internalizing it isn't useful or necessary to accomplish what you actually want to accomplish with the data.
You seem to not make the difference between maths and, say, literature or history.
Do you actually think that an LLM can take, say, a Harry Potter book as an input, and give it a grade in such a way that everybody will always agree on?
And to go further, do you actually use LLMs to generate graphs and statistics from spreadsheet? Because that is probably a bad idea given that there are tools that actually do it right.
> Do you actually think that an LLM can take, say, a Harry Potter book as an input, and give it a grade in such a way that everybody will always agree on?
No, but I also don't think a human can do that either. Subjective things are subjective. I'm not sure I understand how this connects to the idea you expressed that doing various tasks with automation tools like LLMs prevent you from "internalizing" the data, or why not "internalizing" data is necessarily a bad thing. Am I just misunderstanding your concern?
Many of the posts I find here defending the use of LLMs focus on "profitability". "You ask me to give you 3 pages about X? I'll give you 3 pages about X and you may not even realise that I did not write them". I completely agree that it can happen and that LLMs, right now, are useful to hack the system. But if you specialise in being efficient at getting an LLM to generate 3 pages, you may become useless faster than you think. Still, I don't think that this is the point of the article, and it is most definitely not my point.
My point is that while you specialise in hacking the system with an LLM, you don't learn about the material that goes into those 3 pages.
* If you are a student, it means that you are losing your time. Your role as a student is to learn, not to hack.
* More generally as a person, "I am a professional in summarising stuff I don't understand in a way that convinces me and other people who don't understand it either" is not exactly very sexy to me.
If you want to get actual knowledge about something, you have to actually work on getting that knowledge. Moving it from an LLM to a word document is not it. Being knowledgeable requires "internalising" it. Such that you can talk about it at dinner. And have an opinion about it that is worth something to others. If your opinion is "ChatGPT says this, but with my expertise in prompting I can get it to say that", it's pretty much worthless IMHO. Except for tricking the system, in a way similar to "oh my salary depends on the number of bugs I fix? Let me introduce tons of easy-to-fix bugs then".
We were talking about writing, not about vibe coding.
No one said anything about vibe coding. Using tools appropriately to accomplish tasks more quickly is just common sense. Deliberately choosing to pay 10x the cost for the same or equivalent output isn't a rational business decision, regardless of whether the task happens to be writing, long division, or anything else.
Just to be clear, I'm not arguing against doing things manually as a learning exercise or creative outlet. Sometimes the journey is the point; sometimes the destination is the point. Both are valid.
I don't know what your job is.
Here's one: prepping first drafts of legal docs with AI assistance before handing them off to lawyers for revision has objectively saved significant amounts of time and money. Without AI this would have been too time-consuming to be worthwhile, but with AI I've saved not only my own time but the costs of billable hours on phone calls to discuss requirements, lawyers writing first drafts on their own, and additional Q&A and revisions over email. Using AI makes it practical to skip the first two parts and cut down on the third significantly.
Here's another one: doing security audits of customer code bases for a company that currently advertises its use of AI as a cost-saving/productivity-enhancing mechanism. Before they'd integrated AI into their platform, I would frequently get rave reviews for the quality and professionalism of my issue reports. After they added AI writing assistance, nothing changed other than my ability to generate a greater number of reports in the same number of billable hours. What you're suggesting effectively amounts to choosing to deliver less value out of ego. I still have to understand my own work product, or I wouldn't be able to produce it even with AI assistance. If someone thinks that somehow makes the product less "interesting", well then I guess it's a good thing my job isn't entertainment.
Don't get me wrong: I don't deny that LLMs can help tricking other humans into believing that the text is more professional than it actually is. LLMs are engineered exactly for that.
I'd be curious to know whether your legal documents are as good as without LLMs. I wouldn't be surprised at all if they were worse, but cheaper. Talking about security audits, that's actually a problem I've seen: LLMs makes it harder to detect bad audits, and in my experience I have been more often confronted to bad security audits than to good ones.
For both examples, you say "LLMs are useful to make more money". I say "I believe that LLMs lower the quality of the work". It's not incompatible.
If you present your AI-powered work to me, and I suspect you employed AI to do any of the heavy lifting, I will automatically discount any role you claim to have had in that work.
Fairly or unfairly, people (including you) will inexorably come to see anything done with AI as ONLY done with AI, and automatically assume that anyone could have done it.
In such a world, someone could write the next Harry Potter and it will be lost in a sea of one million mediocre works that roughly similar. Hidden in plain sight forever. There would no point in reading it, because it is probably the same slop I could get by writing a one paragraph prompt. It would be too expensive to discover otherwise.
To be clear, I'm not a student, nor do I disagree with academic honor codes that forbid LLM assistance. For anything that I apply AI assistance to, the extent to which I could personally "claim credit" is essentially immaterial; my goal is to get a task done at the highest quality and lowest cost possible, not to cheat on my homework. AI performs busywork that would cost me time or cost money to delegate to another human, and that makes it valuable.
I'm expanding on the author's point that the hard part is the input, not the output. Sure someone else could produce the same output as an LLM given the same input and sufficient time, but they don't have the same input. The author is saying "well then just show me the input"; my counterpoint is that the input can often be vastly longer and less organized or cohesive than the output, and thus less useful to share.
> In such a world, someone could write the next Harry Potter and it will be lost in a sea of one million mediocre works that roughly similar. Hidden in plain sight forever. There would no point in reading it, because it is probably the same slop I could get by writing a one paragraph prompt. It would be too expensive to discover otherwise.
This has already been the case for decades. There are probably brilliant works sitting out there on AO3 or whatnot. But you'll never find them because it's not worth wading through the junk. AI merely accelerates what was already happening.
> someone could write the next Harry Potter and it will be lost in a sea of one million mediocre works that roughly similar.
To be fair, the first Harry Potter is a kinda average British boarding school story. Rowling is barely an adequate writer (and it shows badly in some of the later books). There was a reason she got rejected by so many publishers.
However, Netscape was going nuts and the Internet was taking off. Anime was going nuts and produced some of the all time best anime. MTV animation went from Beavis and Butthead to Daria in this time frame. Authors were engaging with audiences on Usenet (see: Wheel of Time and Babylon 5). Fantasy had moved from counterculture for hardcore nerd boys to something that the bookish female nerds would engage with.
Harry Potter dropped onto that tinder and absolutely caught fire.
I was surprised to find how not true that is when I eventually read the books for myself, long after they became a phenomenon. The books are well-crafted mystery stories that don't cheat the reader. All the clues are there, more or less, for you to figure out what's happening, yet she still surprises.
The world-building is meh at best. The magic system is perfunctory. But the characters are strong and the plot is interesting from beginning to end.
I think the answer to the professor's dismay is quite simple. Many people are in university to survive a brutal social darwinist economic system, not to learn and cultivate their minds. Only a very small handful of them were ever there to study Euler angles earnestly. The rest view it as a hoop they have to jump through to hopefully get a job that might as well be automated away by AI anyway.
Also viewed from a conditional reinforcement perspective, all the professor has to do is start docking grade points from students who are obviously cheating. Theory predicts they will either stop doing it, or get so good at it that it becomes undetectable-possibly an in-demand skill for the future.
Nit pick: He's not a professor, just a grad student at the same place he got his undergrad, and he's mostly gone to university during covid. At least per his page here: https://claytonwramsey.com/about/
Its not like professors get real training either, but the guy doesn't seem to have gotten any real pedagogy.
I guess that I'm driving at that this guy is awfully young and the essay was a hot take. We should judge it accordingly.
Ever since AI came out I’ve been talking about the prompt to output ratio. We naturally assume that the prompt will be smaller than the output just because of the particulars of the systems we use, but as you get more and more particular of what you want, the prompt grows while the output stays the same size. This is logical. If instead of writing an essay, I just describe what I want the essay to say, the description is necessarily gonna be a larger amount of text than the essay itself. It’s more text to describe what’s said, than to just say it. The fact that we expect to do less effort and get back more effort indicates exactly what we’re getting here: a bunch of filler.
In that way, the prompt is more interesting, and I can’t tell you how many times I’ve gone to go write a prompt because I dunno how to write what I wanna say, and then suddenly writing the prompt makes that shit clear to me.
In general, I’d say that AI is way more useful to compress complex ideas into simple ones than to expand simplistic ideas in to complex ones.
This is why it’s unlikely these systems will effectively replace software development. By the time you’ve specified the novel system you want to build well enough in English such that you get exactly the system you want you might as well have written the code.
Yep. To put it another way: In a scenario where you want to say something, you can’t outsource what you want to say to anyone. It doesn’t matter whether or not you want to say it in code or if you want to say it in English.
Well… you’re forgetting the part where you can cut out the middleman. Currently a leader has to ask an engineer to build a system, and has to communicate effectively with the engineer until all of the novel details have been ironed out in the specification, and _y_ the engineer builds it.
In a world where the LLM can do the building, the engineer is no longer required.
More often than not, the building is the easy part once the specifics are ironed out.
In my experience, an ideas leader (you know the type) will fail at telling a machine exactly what to do and get bored with the inevitable edge cases, computers saying no, and non-ideas drudgery. This is where I believe every no-code and low-code and WYSIWYG platform and now LLMs fall apart.
A major aspect of programming is translating the messy meatspace to something an extremely fast moron (a computer and I wish I coined this term) understands. And as much of a step change LLMs for writing code are, I have yet to see them take this step.
They've tried this before and they made COBOL. Turns out you still need programmers to write COBOL because it's still programming even if the program looks Englishy.
I don't think prompt / output ratio, is the speed to get the output that matters.
If I spend 1 hour and write 500 words of prompt to then attach X additional rows of data (e.g. Rows from a table) and the LLM returns X rows of perfect answers. It shouldn't matter that the output ratio is worse than if I had typed those characters myself.
The important thing is whether within that 1 hour (+ few minutes of LLM processing) I managed to get the job done quicker or not.
It's similar to programming, using LLMs is not necessarily to write better code than I personally could but to write good enough code much faster than I ever would.
The issue, IMO, is that some people throw in a one-shot, short prompt, and get a generic, boring output. "Garbage in, generic out."
Here's how I actually use LLMs:
- To dump my thoughts and get help organizing them.
- To get feedback on phrasing and transitions (I'm not a native speaker).
- To improve tone, style (while trying to keep it personal!), or just to simplify messy sentences.
- To identify issues, missing information, etc. in my text.
It’s usually an iterative process, and the combined prompt length ends up longer than the final result. And I incorporate the feedback manually.
So sure, if someone types "write a blog post about X" and hits go, the prompt is more interesting than the output. But when there are five rounds of edits and context, would you really rather read all the prompts and drafts instead of the final version?
I couldn't agree more, this 'polished' style the finished comment comes in is super boring to read. It's hard to put the finger on it, but overall flow is just too... Samesame? I guess it's perfectly _expected_ to be predictable to read ;)
And then one of the iterations was asking additional ways llms could be used and then adding some of those as content which seems odd but plausibly helpful brainstorming.
Just the phrasing of the original question makes it sound like things the user isn't actually doing but wants in their comment if that makes sense.
Thanks for the example chat it was a valuable learning for me!
Exactly it is a tool which needs skill to use. I would add extra use of mine:
- To "Translate to language XYZ", and that is not sometimes strightforward and needs iterating like "Translate to language <LANGUAGE> used by <PERSON ROLE> living in <CITY>" and so on.
And the author is right, I use it as 2nd-language user, thus LLM produces better text than myself.
However I am not going to share the prompt as it is useless (foreign language) and too messy (bits of draft text) to the reader.
I would compare it to passing a book draft thru editor and translator.
For what it's worth, I think that sending a message translated to a foreign language you don't master is the worst thing you can do.
You speak English? Write and send your message in English. The receiver can copy-paste it in a translator. This way, they will know that they are not reading the original. So if your translated message sounds inaccurate, offensive or anything like that, they can go back to your original message.
> would you really rather read all the prompts and drafts instead of the final version?
I think you missed the point of the article. They did not mean it literally: it's a way to say that they are interested in what you have to say.
And that is the point that is extremely difficult to make students understand. When a teacher asks a student to write about a historical event, it's not just some kind of ceremony on the way to a degree. The end goal is to make the student improve in a number of skills: gathering information, making sense of it, absorbing it, being critical about what they read, eventually building an opinion about it.
When you say "I use an LLM to dump my thoughts and get help organising them", what you say is that you are not interested in improving your ability to actually absorb information. To me, it says that you are not interested in becoming interesting. I would think that it is a maturity issue: some day you will understand.
And that's what the article says: I am interested in hearing what you have to say about a topic that you care about. I am not interested into anything you can do to pretend that you care or know about it. If you can't organise your thoughts yourself, I don't believe that you have reached a point where you are interesting. Not that you will never get there; it just takes practice. But if you don't practice (and use LLMs instead), my concern is that you will never become interesting. This time is wasted, I don't want to read what your LLM generated from that stuff you didn't care to absorb in the first place.
I have copilot turned off for markdown files. Cursor has this built in now. I’d never want AI to help write docs (except for narrow cases, repetitive references).
> I’ll now cover the opposite case: my peers who see generative models as superior to their own output. I see this most often in professional communication, typically to produce fluff or fix the tone of their original prompts. Every single time, the model obscures the original meaning and adds layers of superfluous nonsense to even the simplest of ideas.
I'm going to call out what I see as the elephant in the room.
This is brand new technology and 99% of people are still pretty clueless at properly using it. This is completely normal and expected. It's like the early days of the personal computer. Or Geocities and <blink> tags and under construction images.
Even in those days, incredible things were already possible by those who knew how to achieve them. The end result didn't have to be blinking text and auto-playing music. But for 99% it was.
Similarly, with current LLMs, it's already more than possible to use them in effective ways, without obscuring meaning or adding superfluous nonsense. In ways whose results have none of the author's criticisms apply. People just don't know how to do it yet. Many never will, just like many never learnt how to actually use a PC past Word and Excel. But many others will learn.
That's because the instructor is asking questions that merely require the student to regurgitate the instructor's text.
To actually teach this, you do something like this:
"Here's a little dummy robot arm made out of Tinkertoys. There are three angular joints, a rotating base, a shoulder, and an elbow. Each one has a protractor so you can see the angle.
1. Figure out where the end of the arm will be based on those three angles. Those are Euler angles in action. This isn't too hard.
2. Figure out what the angles should be to touch a specific point on the table. For this robot geometry, there's a simple solution, for which look up "two link kinematics". You don't have to derive it, just be able to work out how to get the arm where you want it. Is the solution unambiguous? (Hint: there may be more than one solution, but not a large number.)
3. Extra credit. Add another link to the robot, a wrist. Now figure out what the angles should be to touch a specific point on the table. Three joints are a lot harder than two joints. There are
infinitely many solutions. Look up "N-link kinematics". Come up with a simple solution that works, but don't try too hard to make it optimal. That's for the optimal controls course.
This will give some real understanding of the problems of doing this.
Very well said. It’s a bad assignment! Is 1 student does something like this maybe they’re wrong, but if 90% of students are doing this, then IMO the assignment is wrong.
Or maybe 90% of students are destined for mediocrity.
One of the most fun classes I took in undergrad had people complaining about the professor’s teaching capabilities because it was too hard. We shouldn’t cater to the poor performers.
I fully support the author’s point but it’s hard to argue with the economics and hurdles around obtaining degrees. Most people do view obtaining a degree as just a hurdle to getting a decent job, that’s just the economics of it. And unfortunately the employers these days are encouraging this kind of copy/paste work. Look at how Meta and Google claim the majority of the new code written there is AI created?
You get what you measure, and you should expect people to game your metric.
Once upon a time only the brightest (and / or richest) went to college. So a college degree becomes a proxy for clever.
Now since college graduates get the good jobs, the way to give everyone a good job is to give everyone a degree.
And since most people are only interested in the job, not the learning that underpins the degree, well, you get a bunch of students that care only for the pass mark and the certificate at the end.
When people are only there to play the game, then you can't expect them to learn.
However, while 90% will miss the opportunity right there in front of them, 10% will grab it and suck the marrow. If you are in college I recommend you take advantage of the chance to interact with the knowledge on offer. College may be offered to all, but only a lucky few see the gold on offer, and really learn.
That's the thing about the game. It's not just about the final score. There's so much more on offer.
> However, while 90% will miss the opportunity right there in front of them, 10% will grab it and suck the marrow.
Learning is not just a function of aptitude and/or effort. Interest is a huge factor as well, and even for a single person, what they find interesting changes over time.
I don't think it's really possible to have a large cohort of people pass thru a liberal arts education, with everyone learning the same stuff at the same time, and have a majority of them "suck the marrow" out of the opportunity.
I did a comp science degree, so I can't speak for the liberal arts. However I imagine the same experience could apply.
For us the curriculum was the start of the learning, not the end. We'd get a weekly assignment that could be done in an afternoon. Most of the class did the assignments, and that was enough.
There was a small group of us that lived (pretty much) in the lab. We'd take the assignment and run with it, for days, nights, spare periods, whatever. That 10 line assignment? We turned it into 1000 lines every week.
For example the class on sorting might specify a specific algorithm. We'd do all of them. Compete against each other to make the fastest one. Compare one dataset to another. Investigate data distributions. You know, suck the marrow.
(Our professors would also swing by the lab from time to time to see how things were going, drop the odd hint, or prod the bear in a direction and so on. And this is all still undergrad.
I can imagine a History major doing the same. Researching beyond the curriculum. Going down rabbit holes.
My point is though is that you're right. You need to be interested. You need to have this compulsion. You can't tell a person "go, learn". All you can do is offer the environment, sit back, and see who grabs the opportunity.
I get that you cant imagine this playing out. To those interested only in the degree, it's unimaginable. And no, as long as burning-desire is not on the entry requirements, it most certainly will not be the majority.
In truth the lab resources eoild never have coped if the majority did what we did.
> I did a comp science degree, so I can't speak for the liberal arts.
By 'liberal arts' I meant the common 4 year, non-vocational education. My major was CS too, but well over half of the time was spent on other subjects.
> I get that you cant imagine this playing out. To those interested only in the degree, it's unimaginable
I can easily imagine what you describe playing out. I just wouldn't call it 'sucking the marrow' (unless you were equally avid in all your classes, which time likely would not permit).
But as you allude to in your last point, the system isn't really designed for that. It's nice when it does effectively support the few who have developed the interest, and have extra time to devote to it, as it did for you.
I'd rather see systems that were designed for it though.
That's entirely the point. If you see the degree only as a stepping stone to the company job, then that's all you see and that's all you get.
If you accept that the degree/job relationship is the start, not end, of the reason for being there, then you see other things too.
There are opportunities around the student which are for them, not for their degree, not for their job. There are things you can learn, and never be graded. There are toys to play with you'll never see again. There are whole departments of living experts happy to answer questions.
For example, (this is pre google) I wrote a program and so needed to understand international copyright. I could have gone to the library and read about it. Instead I went to the law faculty, knocked on the door, and found their professor who specialized in intellectual property.
Since the program I wrote was in the medical space, I went to the medical campus, to the medical research library, and found tomes that listed researchers who might benefit. I basically learned about marketing.
If all you care about is the company job, then all you'll see is the degree.
right and getting a family is also just a box to check and eating food is a box to check and brushing my teeth is just a box to check and on it goes for every single thing in life. If we all just checked boxes then we'd not be human anymore.
> Most people do view obtaining a degree as just a hurdle to getting a decent job
Then fail to actually learn anything and apply for jobs and try to cheat the interviewers using the same AI that helped them graduate. I fear that LLMs have already fostered the first batch of developers who cannot function without it. I don't even mind that you use an LLM for parts of your job, but you need to be able to function without it. Not all data is allowed to go into an AI prompt, some problems aren't solvable with the LLMs and you're not building your own skills if you rely on generated code/configuration for the simpler issues.
I think, rather than saying they can’t do their job without an LLM, we should just say some can’t do their jobs.
That is, the job of a professional programmer includes having produced code that they understand the behavior of. Otherwise you’ve failed to do your due diligence.
If people are using LLMs to generate code, and then actually doing the work of understanding how that code works… that’s fine! Who cares!
If people are just vibe coding and pushing the results to customers without understanding it—they are wildly unethical and irresponsible. (People have been doing this for decades, they didn’t have the AI to optimize the situation, but they managed to do it by copy-pasting from stack overflow).
> That is, the job of a professional programmer includes having produced code that they understand the behavior of.
I have met maybe two people who truly understood the behaviour of their code and both employed formal methods. Everyone else, including myself, are at varying levels of confusion.
> I fear that LLMs have already fostered the first batch of developers who cannot function without it.
Playing the contrarian here, but I'm from a batch of developers that can't function without a compiler, and I'm at 10% of what I can do without an IDE and static analysis.
That's really curious: I've never felt that much empowered by an IDE or static analysis.
Sure, there's a huge jump from a line editor like `ed` to a screen editor like `vi` or `emacs`, but from there on, it was diminishing returns really (a good debugger was usually the biggest benefit next) — I've also had the "pleasure" of having to use `echo`, `cat` and `sed` to edit complex code in a restricted, embedded environment, and while it made iterations slower, not that much more slower than if I had a full IDE at my disposal.
In general, if I am in a good mood (and thus not annoyed at having to do so many things "manually"), I am probably only 20% slower than with my fully configured IDE at coding things up, which translates to less than 5% of slow down on actually delivering the thing I am working on.
I think there’s a factor of speed there, not a factor of insight or knowledge. If all you have is ‘ed’ and a printer, then I think most of the time you will spend is with the printout. ‘vi’ eliminates the printout and the tediousness of going back and forth.
Same with more advanced editors and IDEs. They help with tediousness, which can hinders insight, but does not help it if you do not have the foundation.
I've seen this comparison a few times already, but IMHO it's totally wrong.
A compiler translates _what you have already implemented_ into another computer runnable language. There is an actual grammar that defines the rules. It does not generate new business logic or assumptions. You have already done the work and taken all the decisions that needed critical thought, it's just being translated _instruction by instruction_. (btw you should check how compilers work, it's fun)
Using an LLM is more akin to copying from Stackoverflow than using a compiler/transpiler.
In the same way, I see org charts that put developers above AI managers, which are above AI developers. This is just smoke. You can't have LLMs generating thousands of lines of code independently. Unless you want a dumpster fire very quickly...
Yeah ok. I was viewing AI as "a tool to help you code better", not as "you literally can't do anything without it generating everything for you". I could do some assembly if I really had to, but it would not be efficient at all. I wonder if there's actually "developers" who are only prompting an LLM and not understanding anything in the output ? Must be generating dumpster fires as you said.
Lots and lots of developers can't program at all. As in literally - can't write a simple function like "fizzbuzz" even if you let them use reference documentation. Many don't even know what a "function" even is.
(Yes, these are people with developer jobs, often at "serious" companies.)
This is half the point of interviewing. I've been at places that just skip interviewing is the person comes highly recommended, has a great CV, or whatever.
Predictably they end up with some people on the range from "can't code at all" to "newbie coder without talent"
I've never met someone like that and don't believe the claim.
Maybe you mean people who are bad at interviews? Or people whose job isn't actually programming? Or maybe "lots" means "at least one"? Or maybe they can strictly speaking do fizzbuzz, but are "in any case bad programmers"? If your claim is true, what do these people do all day (or, let's say, did before LLMs were a thing...)?
I've definitely worked with a person who struggled to write if statements (let alone anything more complex). This was just one guy, so I wouldn't say "lots and lots" like the other poster did, but they do exist.
Yeah I’ve been doing this for a while now and I’ve never met an employed developer who didn’t know what a function is or couldn’t write a basic program.
I’ve met some really terrible programmers, and some programmers who freeze during interviews.
By "lots" I estimate about 40 percent of the software developer workforce. (Not a scientific estimate.)
> Maybe you mean people who are bad at interviews?
No, the opposite. These developers learn the relevant buzzwords and can string them together convincingly, but fail to actually understand what they're regurgitating. (Very similar to an LLM, actually.)
E.g., these people will throw words like "Dunder method" around with great confidence, but then will completely melt down for fifteen minutes if a function argument has the same name as a module.
When on the job these people just copy-paste existing code from the "serious company" monorepo all day, every day. They call it "teamwork".
> Most people do view obtaining a degree as just a hurdle to getting a decent job, that’s just the economics of it.
Because those who recruit based on the degree aren't worth more than those who get a degree by using LLMs.
Maybe it will force a big change in the way students are graded. Maybe, after they have handed in their essay, the teacher should just have a discussion about it, to see how much they actually absorbed from the topic.
Or not, and LLMs will just make everything worse. That's more likely IMO.
I don't. I think the world is falling into two camps with these tools and models.
> I now circle back to my main point: I have never seen any form of create generative model output (be that image, text, audio, or video) which I would rather see than the original prompt. The resulting output has less substance than the prompt and lacks any human vision in its creation. The whole point of making creative work is to share one’s own experience
Strong disagree with Clayton's conclusion.
We just made this with AI, and I'm pretty sure you don't want to see the raw inputs unless you're a creator:
> We just made this with AI, and I'm pretty sure you don't want to see the raw inputs unless you're a creator:
I am not a creator but I am interested in generative AI capabilities and their limits, and I even suffered through the entire video which tries to be funny, but really isn't (and it'd be easier to skim through as a script than the full video).
So even in this case, I would be more interested in the prompt than in this video.
I’m learning C programming at the moment, originally I was doing it to understand security vulnerabilities more deeply, but I’ve found that I really enjoy the mental exercise of it (and the benefits of that exercise in my career, life etc.) Hopefully the ideas in this article will get to a lot of people eventually, otherwise I feel that people are going to dig themselves in a hole with using LLMs and not thinking for themselves.
"No worthy use of an LLM involves other human beings reading its output."
If you use a model to generate code, let it be code nobody has to read: one-off scripts, demos, etc. If you want an LLM to prove a theorem, have it generate some Coq and then verify the proof mechanically. If you ask a model to write you a poem, enjoy the poem, and then graciously erase it.
I used to teach, years before LLMs, and got lots of copy-pasted crap submitted. I always marked it zero, never mentioning plagiarism (which would require some university administration) and just commenting that I asked for X and instead got some pasted together nonsense.
As long as LLM output is what it is, there is little threat of it actually being competitive on assignments. If students are attentive enough to paraphrase it into their own voice I'd call it a win; if they just submit the crap that some data labeling outsourcer has RLHF'd into a LLM, I'd just mark it zero.
Yeah, the author here is as much a part of the problem. If you let students get away with submitting ChatGPT nonsense, of course they’re going to do that - they don’t care about the 3000 words appeal to emotion on your blog, they take the path of least resistance.
If you’re not willing to cross out an entire assignment and return it to the student who handed it in with “ChatGPT nonsense, 0” written in big red letters at the top of it, you should ask yourself what is the point of your assignments in the first place.
But I get it, university has become a pay-to-win-a-degree scheme for students, and professors have become powerless to enforce any standards or discipline in the face of administrators.
So all they can do is give the ChatGPT BS the minimum passing grade and then philosophize about it on their blog (which the students will never read).
Yeah this is what I did the one time I invigilated/marked a Matlab exam. Very obvious cheating (e.g. getting the right answer with incorrect code). But no way was I going through the admin of accusing them of cheating. They just got a 0.
If it’s copy pasted it’s obvious, and the assignment isn’t to turn in a correct solution, but to turn in evidence that you are able to determine a correct solution. Automated answers deserve 0 credit.
When I was kid in school I would write original essays, and I mean truly original creative ideas. But of course any new idea has a chance of failure, so these essays were mostly bad and got bad grades. At a loss for what to do I quickly stopped reading the books I was assigned basing my essays on Wikipedia summaries and other people’s reviews. I saw my first few As and even A+s and I realized if I write something original of even just average intelligence roughly 50% of people will be too dumb to understand it. For an idea to truly be considered intelligent in literature it has to be appealing to people have no actual memory of the things they’ve read. Even for a knowledgeable intelligent person they have a sea of similar information clouding their view.
Relatedly, there was a major controversy at work recently over the propriety of adding something like this to a lengthy email discussion:
> Since this is a long thread and we're including a wider audience, I thought I'd add Copilot's summary...
Someone called them out for it, several others defended it. It was brought up in one team's retro and the opinions were divided and very contentious, ranging from, "the summary helped make sure everyone had the same understanding and the person who did it was being conscientious" to "the summary was a pointless distraction and including it was an embarrassing admission of incompetence."
Some people wanted to adopt a practice of not posting summaries in the future but we couldn't agree and had to table it.
I think the attribution itself is a certain form of cowardice. If one is actually confident that a summary is correct they'd incorporate it directly. Leaving in the "Copilot says" is an implicit attempt to weasel out of taking responsibility for it.
I'd push back on this: I wouldn't share something I haven't checked, yet my natural writing style is nothing like Gemini, so having a whole paragraph in a different style in my mail requires a disclaimer (I don't want people double guessing what I'm quoting or if I had some different intent writing that paragraph)
Rewriting it in my own words would clear the issue, but then why am I even using an AI in the first place ?
LLMs aren't even that good at summarizing poorly structured text, like email discussions. They can certainly cherry-pick bits and pieces and make a guess at the overall topic, but my experience has been that they're poor at identifying what's most salient. They get particularly confused when the input is internally inconsistent, like when participants on a mailing list disagree about a topic or submit competing proposals.
It is an admission of incompetence. If you need a summary, why don't you add it yourself? Moreover, any person nowadays can easily create a chatGPT summary if necessary. It is just like adding a page of google search results to your writing.
I've noticed that even on here, which is generally extremely bullish on LLMs and AI in general, people get instantly downvoted into oblivion for LLM copypasta in comments. Nobody wants to read someone else's slop.
I often find Copilot summaries to be more or less an attempt at mainsplaining a simple change. If my tiny PR with a one line description requires Copilot to output a paragraph of text about it it’s not a summary, it’s simply time wasted on someone who loves to talk.
How is mansplaining related to this? And saying that summaries of already short information are a waste of time is not really relevant to someone talking about summaries of long and probably repetitive/hard to read discussions.
I will often write a bunch of stuff and then use an LLM to pre-process it a little bit and suggest some improvements. I will then work through the differences and consider them individually and either accept them, or use them to write my own improvements. This is kind of like having an okay editor working for you. No substitute for a real editor, but it means that a: what I intended to say is preserved, b: there's no additional waffle (the prompt includes instructions not to expand on any topic, but only ever to summarise where possible), c: everything still goes by me in the end, and if it doesn't feel like something I would actually write then it doesn't get used.
I believe that it has improved my writing productivity somewhat, especially when I'm tired and not completely on the ball. Although I don't usually reach for this most of the time (e.g. not for this comment).
I think prompt should be treated just like source code, as in the actual "craft" that you're paid to produce. Source code, if computer generated, feels more like artifacts (as in binaries).
If you use AI all that is important is your ability to specify the problem, of course, as it always has been, you can just reiterate faster.
Writing sharpens your thoughts. Working math problems sharpens your ability to do math. In the context of the education system though, where grades are a signal of future ability, there's a strong incentive to engage in rent-seeking by either searching for a solutions manual or "refining" an LLM's output. The little I've done when I've taught math-based econ is to make it clear that in-class tests have very high weight on your final grade, and out-of-class problem sets have very low weight. I can only mouth words as to why it's vital for students to struggle independently on the problem sets as a tool for learning.
I just want to point out that AI generated material is naturally a confirmation bias machine. When the output is obviously AI, you confirm that you can easily spot AI output. When the output is human-level, you just pass through it without a second thought. There is almost no regular scenario where you are retroactively made aware something is AI.
The vast majority of the time people question whether or not an image or writing is "AI", they're really just calling it bad and somehow not realizing that you could just call the output bad and have the same effect.
Every day I'm made more aware of how terrible people are at identifying AI-generated output, but also how obsessed with GenAI-vestigating things they don't like or wouldn't buy because they're bad.
> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think; a language model produces the former, not the latter.
This so much. A writing exercise sharpens your mind, it forces you to think clearly through problems, gives you practice in both letting your thoughts flow onto paper, and in post-editing those thoughts into a coherent structure that communicates better. You can throw it away afterwards, you'll still be a better writer and thinker than before the exercise.
Is the author forgetting a baby in the bathwater he is throwing out?
Especially on coding. He points out that vibe coding is bad, and then concluding that any program written through the use of an AI is bad.
For example if you already have a theory of your code, and you want to make some stuff that is verbose but trivial. It is just more efficient to explain the theory to an LLM and extract the code.
I do like the idea of storing the underlying prompt in a comment.
Same for writing. If you truly copy paste output, it's obviously bad. But if you workshop a paragraph 5 or 6 times that can really get you unstuck.
Even the euler angles example. That output would be a good starting point for an investigation.
A lot of criticism of LLMs is certainly legit and relevant, but are targeting very naive uses of the tools. Post-processing the output with an improvement prompt isn't even mentioned here. I'd rather read a criticism of agents, systems that produce SEO optimized articles etc. The debates would be more interesting.
While I agree with the thrust of the article being that students are cheating themselves by relying on LLMs, it's important to reflect on ways in which educators have encouraged this behavior. Anyone who has been to college in the age of the internet knows that many professors, particularly in the humanities, lazily pad out their class work with short menial writing assignments, often in the form of a "discussion board", that are rarely even graded on content. For students already swamped with work, or having to complete these assignments for general ed courses unrelated to their major/actual interests, it is totally understandable why they would outsource this work to a machine. This is a totally fixable issue: in-person discussions and longer writing assignments with well structured progress reports/check-ins and rounds of peer review are a couple ways that I can think of off the top of my head. Professors need to be held accountable for creating course loads that are actually intellectually interesting and are at least somewhat challenging to use LLMs to complete. When professors are constantly handing out an excess of low-effort assignments, using shortcuts becomes a learned behavior of students.
Hate the game not the player. For the moment we continue to live in a world where the form and tone of communication matters and where foregoing the use of AI tools can put you at a disadvantage.
There are countless homework assignments where teachers will give better grades to LLM outputs.
An LLM can quickly generate targeted cover letters dramatically increasing efficiency while job hunting.
Getting a paper accepted requires you to adhere to an academic writing style. LLMs can get you there.
Maybe society just needs a few more years to adjust and shift expectations.
In the meantime you should probably continue to use AI.
Surely this just makes a mockery of the same tone and style that indicates someone put effort and thought into producing something. This just seems in net to waste everyone's time with no benefit to us.
> They are invariably verbose, interminably waffly, and insipidly fixated on the bullet-points-with-bold style.
No, this is just the de-facto "house style" of ChatGPT / GPT models, in much the same way that that that particular Thomas Kinkade-like style is the de-facto "house style" of Stable Diffusion models.
You can very easily tell an LLM in your prompt to respond using a different style. (Or you can set it up to do so by telling it that it "is" or "is roleplaying" a specific type-of-person — e.g. an OP-ED writer for the New York Times, a textbook author, etc.)
I was surprised to see such world-weary criticism of the bullet-points-with-bold style in TFA— it's long been what I've reached for when writing for a technical audience, whether that's in a wiki page, a design doc, a README, a PR, or even a whole book.
I feel like for most of my audiences it provides the proper anchor points for effective skimming while still giving me room to include further detail and explanation so that it's there as desired by the reader.
(And responding to my sibling comment, I also use em dashes and semicolons all the time. Has my brain secretly always been an LLM??)
One of my issues with LLMs is how much they match the academic, technical, and corporate styles of speaking Ive learned over the years. Now when I write people ignore me because they assume I'm just pasting LLM output.
You are not alone.
Nowadays, I'm ashamed of using words like "moreover", "firstly", "furthermore". Pre-LLM, people used to compliment me on my writing style
That was not a good attempt at changing the style.
You can't just say "don't sound like an LLM." The LLM does not in fact know that it is "speaking like an LLM"; it just thinks that it's speaking the way the "average person" speaks, according to everything it's ever been shown. If you told it "just speak like a human being"... that's what it already thought it was doing!
You have to tell the LLM a specific way to speak. Like directing an image generator to use a specific visual style.
You can say "ape the style of [some person who has a lot of public writing in the base model's web training corpus — Paul Graham, maybe?]". But that coverage will be spotty, and it's also questionably ethical (just like style-aping in image generation.)
But an LLM will do even better if you tell it to speak the in some "common mode" of speech: e.g. "an email from HR", or "a shitpost rant on Reddit" or "an article in a pop-science magazine."
LLM cheating detection is an interesting case of the toupee fallacy.
The most obvious ChatGPT cheating, like that mentioned in this article, is pretty easy to detect.
However, a decent cheater will quickly discover ways to conduce their LLM into producing text that is very difficult to detect.
I think if I was in the teaching profession I'd just leave, to be honest. The joy of reviewing student work will inevitably be ruined by this: there is 0 way of telling if the work is real or not, at which point why bother?
You assume that the teachers job is to catch when someone is cheating; its not. The teachers job is to teach, and if the kids don't learn because their parents allow them to cheat, don't check them at all, and let them behave like shitheads, then the kids will fail in life.
Quite the assertion. If anything the evidence is in favor of the other direction.
It was eye opening to see that most students cheat. By the same token, most students end up successful. It’s why everyone wants their kids to go to college.
Or, bad money chases out good. Idiots that cheat will get the recommendations for jobs where by maxing the grade. The person that actually works gets set back. Even worse society at large loses and actually educated person. And lastly a school is going to attempt to protect their name by preventing cheating.
On reviewing students' work: people exchange copies, get their hands on past similar assignments, get friends to do their homework , potentially each of them shadow the other in fields they're good at etc.
There always was a bunch of realistic options to not actually do your submitted work, and AI is merely makes it easier, more detectable and more scalable.
I think it moves the needle from 40 to 75, which is not great, but you'd already be holding your nose at student work half of the time before AI, so teaching had to be about more than that (and TBH it was, when I was in school teachers gave no fuck about submitted work if they didn't validate it by some additional face to face or test time)
This only came to light after the study had already been running for a few months. That proves that we can no longer tell for certain unless it's literal GPT-speak the author was too lazy to edit themselves.
Teachers will lament the rise of AI-generated answers, but they will only ever complain about the blatantly obvious responses that are 100% copy-pasted. This is only an emerging phenomenon, and the next wave of prompters will learn from the mistakes of the past. From now on, unless you can proctor a room full of students writing their answers with nothing but pencil and paper, there will be no way to know for certain how much was AI and how much was original/rewritten.
> This only came to light after the study had already been running for a few months. That proves that we can no longer tell for certain unless it's literal GPT-speak the author was too lazy to edit themselves.
Rule 3 of the subreddit quite literally bars people from accusing posts of being AI-generated. I have only visited it a few times in recent times, but I noticed quite a few GPT-speak posts with comments calling it out getting removed and punished.
Maybe it will get us to rethink the grading system. Do we need to grade them, or do we need students to learn things? After all, if they grow up to be incompetent, they will be the ones suffering from it.
But I know it's easier said than done: if you get a student to realise that the time they spend at school is a unique opportunity for them to learn and grow, then you're job is almost done already.
> there is 0 way of telling if the work is real or not, at which point why bother?
I might argue you couldn't really tell if it was "real" before LLMs, either. But also, reviewing work without some accompanying dialogue is probably rarely considered a joy anyway.
> there is 0 way of telling if the work is real or not
Talk to the student, maybe?
I have been an interviewer in some startups. I was not asking leetcode questions or anything like that. My method was this: I would pretend that the interviewee is a new colleague and that I am having coffee with them for the first time. I am generally interested in my colleagues: who are they, what do they like, where do they come from? And then more specifically, what do they know that relates to my work? I want to know if that colleague is interested in a topic that I know better, so that I could help them. And I want to know if that colleague is an expert in a topic where they could help me.
I just have a natural discussion. If the candidate says "I love compilers", I find this interesting and ask questions about compilers. If the person is bullshitting me, they won't manage to maintain an interesting discussion about compilers for 15 minutes, will they?
It was a startup, and the "standard" process became some kind of cargo culting of whatever they thought the interviews at TooBigTech were like: leetcode, system design and whatnot. Multiple times, I could obviously tell in advance that even if this person was really good at passing the test, I didn't think it would be a good fit for the position (both for the company and for them). But our stupid interviews got them hired anyway and guess what? It wasn't a good match.
We underestimate how much we can learn by just having a discussion with a person and actually being interested in whatever they have to say. As opposed to asking them to answer standard questions.
I've found really saddening to see students submit written-by-ChatGPT arguments to the department council when their university spot was on the line (for failing grades). This was their ultimate chance to prove their worth and they left it to ChatGPT.
At first, I thought they didn't care. However, it was so pervasive that it couldn't be the only explanation. I was forced to conclude they trusted ChatGPT more than themselves to argue their case... (Some students did not care, obviously.)
The problem isn't LLM, it's how universities are designed. With short terms and high pressure, students develop 'knowledge bulimia' (in lack of a better term). They have to study highly complex fields in short amounts of time, then move on to often unrelated fields quickly thereafter with no emphasize on persistent learning: the knowledge learned in previous exam can be mostly discarded. They may need to 're-learn' it for another exam, but that's fine, they are very good at learning new things which later on can get discarded.
Using LLMs to achieve this is just another step in the evolution of a broken education system. The fix? IMO, make the exams for the courses delayed by one semester. So during the exam study-period, the students have to 'catch up' on the lectures they had a few months ago.
Yeah, to recycle a comment [0] from a few months back:
> Yeah, one of their most "effective" uses is to counterfeit signals that we have relied on--wisely or not--to estimate deeper practical truths. Stuff like "did this person invest some time into this" or "does this person have knowledge of a field" or "can they even think straight." [...]we might have to cope by saying stuff like: "Fuck it, personal essays and cover letters are meaningless now, just put down the raw bullet-points."
In other words, when the presentation means nothing, why bother?
Not to mention that if the teacher would rather read the prompt… then ask for the prompt then? This genuinely reads to me like “I asked for an apple and got an apple instead of an orange, hurr durr I’m so annoyed”.
Asking students for regurgitated info and then being annoyed because they supplied generic regurgitated info is somewhat telling an attitude no?
You're confusing the artifact with the purpose. Teachers across the nation are not trying to accumulate the largest corpus of distinct human-written reviews of The Great Gatby.
The goal is to elicit some kind of mental practice, and the classic request is for something that helps prove it occurred. The issue is that such proofs are now being counterfeited with unprecedented scale and ease.
When those indicators become debased and meaningless, we need to look for other ways of motivating and validating.
Or how about we actually collectively learn a lesson from this - if your assignments just ask people to generically regurgitate info, don’t be surprised that 90% of students lose interest and see it for what it is, pointless busywork.
I genuinely believe I had many excellent learning experiences at university, and I can assure you none of them were the times I had to re-write course info and hand it back to them in order to check off a box.
Maybe, if one student does something they might be wrong, but if 90% of students do something, perhaps the assignment is wrong? Doubling down and saying “we’ll force them to do it by hand then!” Is rather blindly missing the point here no?
It's challenging. Assignments (and particularly programming assignments) were by far the larger and more difficult part of my CS degree, and also the place where I learned the most. I cannot imagine losing that portion of my education and just replacing it with a few exams.
That's how it works in Germany. Usually assignments are either optional or you just have to get 50% of the total assignment marks over the semester to be admitted to the exam (written or often oral, in person). Then your grade is entirely based on the exam. Hand-holding throughout the semester assignment-to-assignment, checking attendance etc. is more an Anglo-specific thing where students are treated as kids instead of adults.
I don’t think there’s anything childish about assigning large-scale programming assignments. I had to do everything from implement a compiler to building a ray tracer as an undergrad, and it set me up very well for an independent career as a software developer at a research lab.
As a professor today, assignments are the place where I’m happy to throw my students into the “deep end” (go learn a new language and a set of library toolkits while also learning this skill.) Exams just don’t provide that experience. Worse, students tend to cram for exams which is the worst way to retain information. I can’t even imagine thinking that the two are comparable in terms of retention and skill-building.
As someone who did CS 15 years ago, assignments show you can actually understand and build programs. Exams show you bothered to re-read the lecture notes the day before and can regurgitate information on command that you could just as easily forget the next day. Not to mention they’re hugely stressful and some students can’t cope with them well even if they’d make excellent programmers.
I can’t imagine disincentivising actually getting stuck into programming and incentivising being good at regurgitating info in an exam room being a good thing for CS students.
If the exam is made well, it's not just regurgitation. It's not multiple choice, but I agree this doesn't test programming skills. But programming is also a quite small slice of the CS curriculum. It has so many other things like linear algebra, real analysis, formal logic, graph theory, automata, coding theory like Reed Solomon, compression, complexity, operating systems like scheduling and virtual memory, how flip-flops and adders and CPUs work. A self taught web developer who can program well still wouldn't know most of these. Programming knowledge is neither fully necessary nor sufficient for a CS degree. I hear American colleges are more vocational, but in most of Europe it's understood that you gain practical experience either in internships, side jobs, doing hobby projects or simply after graduation at your first job.
It is challenging. In my CS degree grading for programming questions fell into two areas
1. Take home projects where we programmed solutions to big problems.
2. Tests where we had to write programs in the exam on paper during the test.
I think the take home projects are likely a lot harder to grade without AI being used. I'd be disappointed if schools have stopped doing the programming live during tests though. Being able to write a program in a time constrained environment is similar to interviewing, and requires knowledge of the language and being able to code algorithms. It also forces you to think through the program and detect if there will be bugs, without being able to actually run the program (great practice for debugging).
Pretty sure you could just give an LLM the given docs / course material and it would be able to write the language to a reasonable standard. Especially if it had sensible error messages.
Honestly I think we'll get back there. I remember ... fondly(?) exams from my history courses in undergrad in the mid 90s. 3-4 questions, 3 hours, anything less than what would amount to a pretty decent length and moderately thorough term paper would fail and have to be made up with an absolutely BRUTAL multiple choice + fill in the blank exam at the end of the term.
Those classes are what taught me how to study and really internalize the material. Helped me so much later in college too. I really can't imagine how kids these days are doing it.
> [Not a student’s real answer, but my handmade synthesis of the style and content of many answers]
> You only have to read one or two of these answers to know exactly what’s up: the students just copy-pasted the output from a large language model, most likely ChatGPT. They are invariably
This is validating. Your imitation completely fooled me (I thought it really was ChatGPT and expected to be told as much in an entirely unsurprising "reveal") and the subsequent description of the style is very much in agreement with how I'd characterize it.
In previous discussions here, people have tried to convince me that I can't actually notice these obvious signs, or that I'm not justified in detecting LLM output this way. Well, it may be the case that all these quirks derive from the definitely-human training data in some way, but that really doesn't make them Turing-test-passing. I can remember a few times that other people showed me LLM prose they thought was very impressive and I was... very much not impressed.
> When someone comments under a Reddit post with a computer-generated summary of the original text, I honestly believe that everyone in the world would be better off had they not done so. Either the article is so vapid that a summary provides all of its value, in which case, it does not merit the engagement of a comment, or it demands a real reading by a real human for comprehension, in which case the summary is pointless. In essence, writing such a comment wastes everyone’s time.
I think you've overlooked some meta-level value here. By supplying such a comment, one signals that the article is vapid to other readers who might otherwise have to waste time reading a considerable part of the article to come to that conclusion. But while it isn't as direct as saying "this article is utterly vapid", it's more socially acceptable, and also more credible than a bald assertion.
I’ve used ChatGPT as an editor and had very good results. I’ll write the whole thing myself and then feed it into ChatGPT for editing. And then review its output to manually decide which pieces I want to incorporate. The thoughts are my own, but sometimes ChatGPT is capable of finding more succinct ways of making the points.
I generally make sure I use diff tools for that type of task, because LLMs are really good at making subtle changes you don't easily notice that are wrong.
> You only have to read one or two of these answers to know exactly what’s up: the students just copy-pasted the output from a large language model.
I don't understand this either. I use it a lot, but I never just use what an LLM says verbatim. It's so incredibly obvious it's not written by a human.
Most of the time I write an initial draft, ask Claude to check it and improve it, and then I might touch up a few sentences here and there.
> Vibe coding; that is, writing programs almost exclusively by language-model generation; produces an artifact with no theory behind it. The result is simple: with no theory, the produced code is practically useless.
Maybe I still don't know what vibe coding is, but for the few times when I _can_ use an LLM to write code for me, I write a pretty elaborate instruction on what I want, how it should be written, ... Most of the time I use it for writing things I know it can do and seem tedious to me.
Am a student, the main message I have taken from this article, I should love to write and be comfortable with my thoughts no matter the situation. Thanks for this amazing writing.
So it seems like the future is people to write in a command prompt style for llms to better parse and repeat back our information. God I hope that isn't the future of the internet.
How about an emoji like library designed exclusively for LLMs, so we can quickly condense context and mood without having to write a bunch of paragraphs, or the next iteration of "txt" speech for LLMs. What does the next step of users optimising for LLMs look like?
AI has changed how we learn by making the process of improving work much easier. Normally, learning involves writing a draft, finding mistakes, and fixing them over time. This helps build critical thinking. AI, trained on tons of refined data, can create polished work right away. While this seems helpful, it can skip the important step of learning through trial and error.
The question is: Should we limit AI to keep the old way of learning, or use AI to make the process better? Instead of fixing small errors like grammar, students can focus on bigger ideas like making arguments clearer or connecting with readers. We need to teach students to use AI for deeper thinking by asking better questions.
We need to teach students that asking the right questions is key. By teaching students to question well, we can help them use AI to improve their work in smarter ways. The goal isn’t to go back to old methods for iterating but change how we iterate altogether.
I would argue that if you are losing a consequent amount of your time fixing grammar, then it sounds like you need to spend that time to improve your grammar skills.
> We need to teach students to use AI for deeper thinking by asking better questions.
Same thing here: the whole point of learning critical thinking is that you don't need to ask someone/something else. Teaching you how to ask the LLM to do it for you is not the same as teaching you how to actually do it.
In my opinion, we need to make students realise that their goal is to learn how to do it themselves (whatever it is). If they need an LLM to do it, then they are not learning. And if they are not learning, there is no point in going to school, they can go work in a field.
You’re getting to the crux of the argument, knowing when to use AI. Doing or learning “It” in 2025 means using AI whether to understand it better or use it to get better grades.
My take is teach them to get better at asking questions and then teach them when to use their own understanding to change their answer for the better. How many times has an AI’s answer been 5/10 and with a few fixes it’s a 9/10. That comes with time. Getting them asking questions and learning the “when” later is better at least to me.
> Doing or learning “It” in 2025 means using AI whether to understand it better or use it to get better grades.
Depends, I think. If we are talking about writing an essay (and I believe we are), then the LLM is somewhere between useless and counter-productive.
Of course, if the LLM is used to understand an RFC (I would debate how useful it is for that, but that's another discussion), then it's different. The goal was to understand the RFC, it doesn't really matter how you did it. But the goal of writing an essay is not to end up with a written essay at all. Nobody cares about it, you can burn it right after it's graded.
>I believe that the main reason a human should write is to communicate original thoughts
More than communicate, I would say to induce thoughts.
I write poetry here and there (on paper, just for me). I like how exploration through lexical and syntactic spaces can be intertwined with semantics and pragnatic matters. More importantly, I appreciate how careful thoughts are playing with attention and other uncharted thoughts. The invisible side effects on mental structures happening in the creation of expression can largely outweight the importance of what is left as an artefact publicly visible.
For a far more trivial example, we can think about how notes in the margin of a book can radically change the way we engage with the reading. Even a careful spare word highlight can be a world of difference in how we engage with the topic. It's the very opposite of "reading" a few pages before realizing that not a single thought percolated into consciousness as it was wandering on something else.
By now I consider LLM text a double insult. It says “I couldn’t be bothered to spend time writing this myself,” yet it makes _me_ waste time reading through the generated fluff! I agree with the article, I'd rather read the prompt.
If you outsource that to a model, you often end up with words but shallow or no understanding. Writing forces you to clarify your ideas.
LLMs substitute genuine thinking with surface-level prose, which might sound alright but often lacks depth behind it.
You mean to tell me even anti-AI people are glazing my unseen prompts now! The solve for slop is easy for teachers and communicators alike: stop asking sorry questions and you stop getting sorry responses. Or stay the easy course, you will cede the game to discreet cheaters just to make honest people jump through antiquated hoops.
An exception to test the rule with: people are generating lifelike video based on the pixel graphics from old video games. I have no interest in seeing a prompt that says "Show me a creature from Heroes of Might and Magic 3, with influences from such and so", but it's incredible to see the monsters I've spent so much time with coming to life. https://www.youtube.com/watch?v=EcITgZgN8nw&lc=UgxrBrdz4BdEE...
Maybe the problem is that the professor doesn't want to read the student work anyway, since it's all stuff he already knows. If they managed to use their prompts to generate interesting things, he'd stop wanting to see the prompts.
> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think; a language model produces the former, not the latter.
Personally, I’ve been enjoying using ChatGPT to explore different themes of writing. It’s fun. In my case the goal is specifically to produce artifacts of text that’s different from what I’d normally produce.
> A typical belief among students is that classes are a series of hurdles to be overcome; at the end of this obstacle course, they shall receive a degree as testament to their completion of these assignments.
I agree with the broader point of the article in principle. We should be writing to edify ourselves and take education seriously because of how deep interaction with the subject matter will transform us.
But in reality, the mindset the author cites is more common. Most accounting majors probably don't have a deep passion for GAAP, but they believe accounting degrees get good jobs.
And when your degree is utilitarian like that, it just becomes a problem of minimizing time spent to obtain the reward.
I think the author is conflating assignments with learning. The assignments are, by definition, hurdles to be overcome, and they will always be treated as such. Learning can happen in other better ways, and perhaps if your assignment design is to get students to generically regurgitate course material you shouldn’t be surprised when that’s exactly what they do.
I can’t be the only student who had both the experience of wonderful learning moments, AND could see a badly designed assignment a mile off and wasn’t motivated to give such a thing my full attention no?
As a side note, if you want the prompt, simply ask for it in the assignment. Asking students for one thing and then complaining when you don’t get another is insanity.
I have a very similar experience. Some students who want to get involved in contributing to open source will try to contribute to Zulip by taking whatever they wanted to say and asking ChatGPT to write it better for them, and posting the result.
Even when no errors are introduced in the process, the outcome is always bad: 3 full paragraphs of text with bullets and everything where the actual information is just the original 1-2 sentences that the model was prompted with.
I never am happy reading one of those; it's just a waste of time. A lot of the folks doing it are not native English speakers. But for their use case, older tools like Grammarly that help improve the English writing are effective without the problematic decompression downsides of this class of LLM use.
Regardless of how much LLMs can be an impactful tool for someone who knows how to use one well, definitely one of the impacts of LLMs on society today is that a lot of people think that they can improve their work by having an LLM edit it, and are very wrong.
(Sometimes, just telling the LLM to be concise can improve the output considerably. But clearly many people using LLMs think the overly verbose style it produces is good.)
I occasionally pair write with LLMs, but I give them my piece and then say, "I don't want your edits, just your feedback" and ask them some simple questions about the content and different angles on it. When the LLM says what I want it to say, I consider the piece good enough. That is to say, if a machine understands what you're saying and a human doesn't, that human's criticism might be below engaging with.
> [...] but not so distinctive to be worth passing along to an honor council. Even if I did, I’m not sure the marginal gains in the integrity of the class would be worth the hours spent litigating the issue.
The school should be drilling into students, at orientation, what some school-wide hard rules are regarding AI.
One of the hard rules is probably that you have to write your own text and code, never copy&paste. (And on occasions when copy&paste is appropriate, like in a quote, or to reuse an off-the-shelf function, it's always cited/credited clearly and unambiguously.)
And no instructors should be contradicting those hard rules.
(That one instructor who tells the class on the first day, "I don't care if you copy&paste from AI for your assignments, as if it's your own work; that just means you went through the learning exercise of interacting with AI, which is what I care about"... is confusing the students, for all their other classes.)
Much of society is telling students that everything is BS, and that their job is to churn BS to get what they want. Early "AI' usage popular practices so far looks to be accelerating that. Schools should be dropping a brick wall in front of that. Well, a padded wall, for the students who can still be saved.
Back in the 90’s I remember similar sorts of kickback from traditional media about the internet and from academia about students not using the library on campuses anymore.
Libraries are still in every campus, often with internet access.
Traditional media have transitioned to become online content media farms. The NYT Crossword puzzle is now online. Millions of people do Wordle every day online.
This is just kickback. Every paradigm shift needs kickback in order to let the dust settle and for society to readjust and find equilibrium again.
The output is so convenient that these students seem like they don't even change bits of it to make it their own.
Since there is no "interdiction" to use any LLM, perhaps it should be mandatory to include the prompt as well when used. Feels like that could be the seed that sparks the curiosity..
Is bringing up Naur's paper and arguing that theory of program is all that matters and LLMs cannot do that, just a 2025 version of calling LLMs stochastic parrots and claiming they don't model or work in terms of concepts? Feels like it.
EDIT: Not a jab at the author per se, more that it's a third or fourth time I see this particular argument in the last few weeks, and I don't recall seeing it even once before.
As someone who is an immigrant that had to go to high school in English speaking country and who struggled a lot and couldn’t do anything about improving essay writing no matter what I did, I say all these English teachers deserve this. I wish ChatGPT existed during my school years, I would’ve at least had someone(thing) explain me how to write better.
> I would’ve at least had someone(thing) explain me how to write better.
I actually don't think that it is good at that. I have heard of language teachers trying to use it to teach the language (it's a model language, it should be good at it, right?) and realised that it isn't good at that.
Of course I understand the point of your message, which is that you feel your teachers were not helpful and I have empathy for that.
> The model produces better work. Some of my peers believe that large language models produce strictly better writing than they could produce on their own. Anecdotally, this phenomenon seems more common among English-as-a-second-language speakers. I also see it a lot with first-time programmers, for whom programming is a set of mysterious incantations to be memorized and recited.
AI usage is a lot higher in my work experience among people who no longer code and are now in business/management roles or engineers who are very new and didn't study engineering. My manager and skip level both use it for all sorts of things that seem pointless and the bootcamp/nontraditional engineers use it heavily. Our college hires we have who went through a CS program don't use it because they are better and faster than it for most tasks. I haven't found it to be useful without an enormous prompt at which point I'd rather just implement the feature myself.
The "aha" moment for me came when I started writing a ticket for a junior engineer to work on. However, to satisfy my own curiosity, I gave the ticket to Cursor, and was able to get 90% of the way there (implementing a small feature based on a Figma design).
As it turns out, a well written ticket makes a pretty good input into an LLM. However, it has the added benefit of having my original thought process well documented, so sometimes I go through the process of writing a ticket / subtask, even if I ended up giving it to an AI tool in the end.
I mostly use LLMs as a more convenient Google and to automate annoying code transformations with a conveniency of a natural language interface.
Sometimes, I use it to "improve" my writing style.
I have to admit I was a bit surprised how bad LLMs are at the continue this essay task. When I read it in the blog I suspected this might have been a problem with the prompt or the using one of the smaller variants of Gemini. So I tried it with Gemini 2.5 Pro and iterated quite a bit providing generic feedback without offering solutions. I could not get the model to form a coherent well reasoned argument. Maybe I need to recalibrate my expectations of what LLMs are capable, but I also suspect that current models have heavy guardrails, use a low temperature and have been specifically tuned for problem solving and avoid hallucinations as much as possible.
Although I agree with the OP that copy pasting verbatim from the LLM is a meaningless exercise, I just like to say that LLMs can be a fantastic study tool. I am using ChatGPT to help me with learning German, and it has had a profound impact on my learning.
> Anecdotally, this phenomenon seems more common among English-as-a-second-language speakers
That part caught my attention. As an English-as-a-second-language speaker myself, I find it so difficult to develop any form of "taste" in English the same way I have in my mother tongue. A badly written sentence in my mother tongue feels painful in a sort of physical way, while bad English usually sound OK to me, especially when asserted in the confident tone LLMs are trained in. I wish I could find a way to develop such sense for the foreign languages I currently use.
Good that you brought that up: that works pretty well to me in my mother tongue. I still learn and absorb beautiful and useful patterns reading good authors. But that doesn't seem to work in other languages as well. I somehow don't manage to appropriate the new patterns, or maybe I do, but very slowly.
Interestingly, when it comes to spoken English, I can learn by imitation way faster.
I think half the population of the world wants just the facts, the other half wants long flowing beautiful content like on apple.com, and neither group knows the other exists. Of course this is the right way to do it!
Spoken like a true "just the facts" guy. If this is a competition then Hemmingway is losing hands down. Serious well-funded marketing departments came up with that long-form style, because that's what the vast majority of people respond to. I'm just happy they have a tiny "tech specs" section for me (and presumably you).
I just went on apple.com and tried for 8 minutes to find "long beautiful flowing content". Most of the actual text I came across there were not-so-beautiful legal disclaimers...
For math and writing, we still have in-class exams as an LLM-free evaluation tool.
I wish there was some way to do the same for programming. Imagine a classroom full of machines with no internet connection, just a compiler and some offline HTML/PDF documentation of languages and libraries.
+-12 years ago I took a course on concepts of programming languages. The exam consisted out of programming exercises in a classroom without internet access.
Grading was based on the number of (hidden) unit tests that succeeded.
I have a lot of sympathy for the author's position but I may have missed the point in the article where he explained why clarity of writing and genuineness of human expression was so vital to a robotics class. It's one thing for an instructor to appreciate those things; another for them to confound their own didactic purpose with them. This point seems obvious enough that I feel like I must have missed something.
As always, I reject wholeheartedly what this skeptical article has to say about LLMs and programming. It takes the (common) perspective of "vibe coders", people who literally don't care what code says as long as something that runs comes out the other side. But smart, professional programmers use LLMs in different ways; in particular, they review and demand alterations to the output, the same way you would doing code review on a team.
I think they summed it up well in the section "Why do we write, anyway?" — they nowhere claimed it was vital for students' success in a robotic class. On the contrary as they title a subsection there with "If it’s worth doing, it’s worth doing *badly*" (emphasis mine) — so what they are looking for is to peer into the author's mind and their original thoughts.
The implication there is that this is acceptable to pass a robotics class, and potentially this gives them more information about students' comprehension to further improve their instruction and teaching ("...that they have some kind of internal understanding to share").
On that second point, I have yet to see someone demonstrate a "smart, professional programmer use LLMs" in a way where it produces high quality output in their area of expertise, while improving their efficiency and thus saving time for them (compared to them just using a good, old IDE)!
A couple of examples from influential open source developers: Adam Wathan (Tailwind) agreeing with Mitchell Hashimoto that LLMs are making them more productive. "in their area of expertise" is not obvious from this post alone, but I am pretty confident from the way they talk about it that they are not exclusively using LLMs on side projects where they're inexpert.
Armin Ronacher also talks about using LLMs quite a bit, but I don't have as good of an example from his tweets of him straightforwardly saying "yes, they are useful to me!"
Is "what has been my experience" not implied in what I am still waiting for — "someone using an LLM to produce high quality code in the field of their expertise in less time than without an LLM"?
So, observing a couple of my colleagues (I am an engineering manager, but have switched back and forth between management and IC roles for the last ~20 years), I've seen them either produce crap, or spend so much time tuning the prompts that it would have been faster to do it without an LLM. They mostly used Github Copilot or ChatGPT (most recent versions as of last few months ago).
Again, I am not saying it's not being done, but I have struggled to find someone who would demonstrate it happen in a convincing enough fashion — I am really trying to imagine how I would best incorporate this into my daily non-work programming activities, so I'd love to see a few examples of someone using it effectively.
If you don't want to engage in an honest discussion, please refrain from making assumptions: nobody mentioned "merging crap". I stopped clearly at guiding an LLM to make a code change, which is at best, a pull request.
Preach about the bullet points. I was grading some assignments a while ago and by some mysterious coincidence like a third of the answers were written in this strange bullet point format listing the the same 3 ideas.
The punchline? Bullet point 3 was wrong (it was a PL assignment and I'm 99% sure the AI was picking up on the word macro and regurgitating facts abut LISP). 0 points all around, better luck next time.
As long as they provide the prompt and output combo, and the output solves the requirements of the assignment, I don’t see the difference. Half the students are probably asking the llm for prompt help also anyway.
I’d argue that making students give generic regurgitated info as an assignment is the actual issue. Make a good assignment…
the solution is obvious. stop grading the result, and start grading the process.
if you can one-shot an answer to some problem, the problem is not interesting.
the result is necessary, but not sufficient. how did you get there? how did you iterate? what were the twists and turns? what was the pacing? what was the vibe?
no matter if with encyclopedia, google, or ai, the medium is the message. the medium is you interacting with the tools at your disposal.
record that as a video with obs, and submit it along with the result.
for high stakes environments, add facecam and other information sources.
reviewers are scrubbing through video in an editor. evaluating the journey, not the destination.
Unfortunately, the video is a far cry from carrying all the representative information: there is no way you can capture your full emotions as you are working through a problem, and where did you get your "eureka" moments unless you are particularly good at verbalising your through process as you go through multiple dead-ends and recognize how they lead you in the right direction.
Not really: I love reading fiction where I can imagine characters the way I want to based on their written depictions. When I see a book cover replaced with a recent movie adaptation actor, it usually reduces the creative space for the reader instead of enlarging it.
Video in itself is not more information by definition. Just look at those automatically generated videos when you try finding a review on an unusual product.
Here’s something I sometimes do to avoid boring content when using LLMs: I type out what it gives me and tweak it as I go instead of copy/pasting directly.
It helps me spot the bits that feel flat or don’t add much, so I can cut or rework them—while still getting the benefit of the LLM’s idea generation.
I think you're missing the point. You should be providing the ideas in any piece of text you write. The boring prose style is not the major objection being expressed here.
Where I especially hold this viewpoint is for end-of-year peer performance reviews.
People say “I saved so much time on perf this year with the aid of ChatGPT,” but ChatGPT doesn’t know anything about your working relationship with your coworker… everything interesting is contained in the prompt. If you’re brain dumping bullet points into an LLM prompt, just make those bullets your feedback and be done with it? Then it’ll be clear what the kernel of feedback is and what’s useless fluff.
I am personally proud of my use of AI because for anything non-trivial it is generally a conversation where each recommendation needs to be altered by an imaginative suggestion. So ultimately it’s the entire conversation that needs to be considered, not just the “final” prompt.
Except I have never met anyone that likes fluffed up emails, nor considers them better in any way than your 4 points. Take that from a long-email writer (people beg me for a few bullets, somehow I always feel that it does not accurately convey my message, I am aware that this is a shortcoming) ;)
We had a town hall type of thing recently with some of our investors, and some of the answers they gave were hilarious, if it weren't so depressing that those people are the ones with most of the money in this world.
The worst was the answer to the question "How can we utilize AI to greater effect in our work?". A nice open-ended question where they had a beautiful opportunity to show off how knowledgeable and forward thinking they are, right? Especially considering they're the ones behind the massive AI push our product has gone with as of late.
"You can ask it to write emails for you!" Was the one and only thing these multi-milli/billionaires could come up with. Our core product itself is literally an email interface, and we have an AI email generation feature built in...
I had to turn my webcam off because I genuinely laughed out loud at that response for how insanely elementary and useless it was as an answer. It also showed me these people do literally nothing other than answer emails - and even then they're too bloody lazy and give so little of a shit they can't even do that part themselves.
Well of course. But well thought out and well written communication (which admittedly is rare) is an opportunity to actually think through what you’re telling people. How does point A relate to point B? As you read through what you’ve written do you realize there should be a point A.1 to bridge a gap?
It’s like math homework, you always had to show your working not just give the answer. AI gives us an answer without the journey of arriving at one, which removes the purpose of doing it in the first place.
This is a lot of words of complaining instead of just changing the assignment specification to ask students to provide a prompt that they’d use, instead of asking them to regurgitate information. You get what you ask for mate, I don’t understand what’s so hard about this.
If your assignment can be easily performed by an LLM, it’s a bad assignment. Teachers are just now finding out the hard way that these assignments always sucked and were always about regurgitating information pointlessly and weren’t helpful tools for learning lol. I did heaps of these assignments before the existence of LLMs, and I can assure you that the busywork was mostly a waste of time back then too.
People using LLMs is just proof they don’t respect your assignment - and you know what, if one person doesn’t respect your assignment, they’re probably wrong. But if 90% of people don’t respect your assignment? Maybe you should consider whether the assignment is the problem. It’s not rocket science.
Yeah. I wrote so many hamburger essays that 1) were terrible & 2) got As, I wouldn't blame a student for generating that crap. I have written approximately 0 hamburger essays post-high school. They don't want good, they want their made-up rules to be followed
> you're not always going to have a calculator in your pocket was the old fib, now it's s/calculator/llama/
So much of education when I was growing up was pointless box-checking, I'm a little satisfied that, finally, LLMs might swing the pendulum back towards valuable work. If any assignment an LLM could handle is limited to 10%, and 90% of the final grade is determined by oral exams, that seems positive.
Is "2x4" equal to 4 + 4, or 2 + 2 + 2 + 2 ? There's only one correct answer! "You ... pick up that can" wow such learning
If LLMs love to produce 3-bullet-point-bold-font copypasta, it's probably because it's the exact varietal of crap that garnered so many upvotes on Quora. Why would I be asked to write a 500-word essay when 50 words would suffice? Maybe let's move beyond regurgitation & rote drudgery
As someone who doesn't have to grade student assignments, I'd rather read the bullet points. I've always liked using bullet points even in grade school, and also used words like dwelve, so I'd definitely fail the AI sniff test if I ever went back to school.
Just paste the full text and ask a LLM to summarize it for you.
It feels like we are getting to this weird situation where we just use LLMs as proxies, and the long, boring text is just for LLMs to talk to each other.
The very first time I enjoyed talking to someone in another language, I was 21. Then an exchange student, I had a pleasant and interesting discussion with someone in that foreign language. On the next day, I realised that I wouldn't have been able to do that without that foreign language. I felt totally stupid: I had been getting very good grades in languages for years at school without ever caring about actually learning the language. And now, it was obvious, but all that time was lost; I couldn't go back and do it better.
A few years earlier, I had this great history teacher in high school. Instead of making us learn facts and dates by heart, she wanted us to actually get an general understanding of a historical event. Actually internalise, absorb the information in such a way that we could think and talk about it. And eventually develop our critical thinking. It was confusing at first, because when we asked "what will the exam be about", she wouldn't say "the material in those pages". She'd be like "well, we've been talking about X for 2 months, it will be about that".
Her exams were weird at first: she would give us articles from newspapers and essentially ask what we could say about them. Stuff like "Who said what, and why? And why does this other article disagree with the first one? And who is right?". At first I was confused, and eventually it clicked and I started getting really good at this. Many students got there as well, of course. Some students never understood and hated her: their way was to learn the material by heart and prove it to get a good grade. And I eventually realised this: those students who were not good at this were actually less interesting when they talked about history. They lacked this critical thinking, they couldn't make their own opinion or actually internalise the material. So whatever they would say in this topic was uninteresting: I had been following the same course, I knew which events happened and in which order. With the other students were it "clicked" as well, I could have interesting discussion: "Why do you think this guy did this? Was it in good faith or not? Did he know about that when he did it? etc".
She was one of my best teachers. Not only she got me interested in history (which had never been my thing), but she got me to understand how to think critically, and how important it is to internalise information in order to do that. I forgot a lot of what we studied in her class. I never lost the critical thinking. LLMs cannot replace that.
Yeah this article misses a big point and you’ve highlighted it well. If you just ask students to regurgitate course material generically, it doesn’t (and wasn’t) leading to good learning outcomes whether LLMs existed or not. All the LLM is doing here is signalling to the teacher that their assignment design is bad, but it seems they’re learning the wrong lesson.
There’s a lot of “no, it is the children who are wrong” going on in academia right now and it’s an issue.
Crucially: if you just send me the prompt, and for some reason I would rather have read the model output, I can just paste the prompt into the model. However, theres no way to go the other way
There's also strong inferiority complex. When you read and find out the output your motivation to at least paraphrase the prompt output instantly dive because it looks so good and proper whereas your original writing looks so dumb in comparison
1. “When copying another person’s words, one doesn’t communicate their own original thoughts, but at least they are communicating a human’s thoughts. A language model, by construction, has no original thoughts of its own; publishing its output is a pointless exercise.”
LLMs, having being trained using the corpus of the web, I would argue communicate other human’s thoughts particularly well. Only in exercising an avoidance of plagiarism are the thoughts of other human’s evolved into something closer to “original thought” for the would-be plagarizer. But yes, at least a straight copy/paste retains the same rhetoric as the original human.
2. I’ve seen a few advertisements recently leverage “the prompt” as a means to resonate visual appeal.
i.e a new fast food delivery service starting their add with some upbeat music and a visual presentation of somebody typing into a LLM interface, “Where’s the best sushi around me?” And then cue the advertisement for the product they offer.
I like reading and writing stories. Last month, I compared the ability of various LLMs to rewrite Saki's "The Open Window" from a given prompt.[1] The prompt follows the 13-odd attempts. I am pretty sure in this case that you'd rather read the story than the prompt.
I find the disdain that some people have for LLMs and diffusion models to be rather bizarre. They are tools that are democratizing some trades.
Very few people (basically, those who can afford it) write to "communicate original thoughts." They write because they want to get paid. People who can afford to concentrate on the "art" of writing/painting are pretty rare. Most people are doing these things as a profession with deadlines to meet. Unlike you are GRRM, you cannot spend decades on a single book waiting for inspiration to strike. You need to work on it. Also, authors writing crap/gold at a per-page rate is hardly something new.
LLMs are probably the most interesting thing I have encountered since I did the computer. These puritans should get off of their high horse (or down from their ivory tower) and join the plebes.
You can train an LLM to maximize the information content bitrate. I just think most companies want to maximize "customer satisfaction" or w/e, which is why we get the verbose, bold, bullet points.
All it takes is to provide a slightly better prompt (“write the answer in a natural prose style, no bullet points, no boring style, perhaps introduce a small error). It’s not that difficult.
I am thinking about creating a proof-of-writing signature. Basically an editor with an "anti-cheat", you can't paste text into it. It signs your text with a public key.
There is no way to design such a system that is not cheatable. At the very least, someone could simply type out text from another window or device. On any normal operating system or browser, the user will be able to bypass whatever mechanism you have in place anyway.
For everyone pointing out that this idea can be cheated by just typing AI-generated text into the editor - add an AI-detector to the editor. Gamify the whole thing by making a leaderboard of people with the lowest AI-detector-similarity score across things that they have "written"
In a class setting, maybe make the AI-detection an element of take-home assignments - whoever gets the lowest AI-similarity score gets a few points of extra credit or something
As for computer science courses, I'm guessing it's hard to not write simple code that appears AI-generated...so maybe that kind of work needs a written summary to go along with the code as well
You can still just type the Ai response.
Often when I generate larger code I type it instead of copy paste, that helps me understand it and spot issues faster
> Don’t let a computer write for you! I say this not for reasons of intellectual honesty, or for the spirit of fairness. I say this because I believe that your original thoughts are far more interesting, meaningful, and valuable than whatever a large language model can transform them into.
Having spent about two decades reading other humans' "original thoughts", I have nothing else to say here other than: doubt.
> A typical belief among students is that classes are a series of hurdles to be overcome; at the end of this obstacle course, they shall receive a degree as testament to their completion of these assignments.
IMO the core problem is that in many cases this typical belief holds true.
I went to university to get a degree for a particular field of jobs. I'd generously estimate that about half of my classes actually applied to that field or jobs. The other half were required to make me a more "well rounded student" or something like that. But of course they were just fluff to maximize my tuition fees.
There was no university that offered a more affordable program without the fluff. After all, the fluff is a core part of the business model. But there isn't much economic opportunity without a diploma so students optimize around the fluff.
I think people who don’t like writing shouldn’t be forced to write, just like people who don’t like music shouldn’t be forced to play music. Ditto for math.
Forcing people to do these things supposedly results in a better, more competitive society. But does it really? Would you rather have someone on your team who did math because it let them solve problems efficiently, or did math because it’s the trick to get the right answer?
Writing is in a similar boat as math now. We’ll have to decide whether we want to force future generations to write against their will.
I was forced to study history against my will. The tests were awful trivia. I hated history for nearly a decade before rediscovering that I love it.
History doesn’t have much economical value. Math does. Writing does. But is forcing students to do these things the best way to extract that value? Or is it just the tradition we inherited and replicate just because our parents did?
Many of the things we teach in school aren’t just for the direct knowledge or skill. We largely don’t need to do arithmetic any more, but gaining the skill at doing it really improves our ability to deal with symbolic manipulation and abstraction.
I remember another parent ranting about their 3rd grade kids “stupid homework” since it had kids learning different ways of summing numbers. I took a look at the homework and replied “wow, the basics out set theory are in here!” We then had a productive discussion of how that arithmetic exercise led to higher math and ways of framing problems.
Similarly, writing produces a different form of thought than oral communication does.
History is a bit different, but a goal of history and literature is (or it least should be) to socialize students and give them a common frame of reference in society.
Finally there is the “you don’t know when you’ll need it defense.” I have a friend who spent most of the last 20 years as a roofer, but his body is starting to hurt. He’s pivoting to CAD drafting and he’s brushing off a some of those math skills he hated learning in school. And now arguing with his son about why it’s important.
Those are the fundamental defenses- that we are seeking not skills but ways of viewing the world + you don’t know what you’ll need. There are obviously limits and tradeoffs to be made, but to some degree yes, we should be forcing students (who are generally children or at least inexperienced in a ___domain) to things they don’t like now for benefits later.
Then your friend spent 20 years not needing math skills. If someone spent years doing something useless to them for two decades, we wouldn’t call them efficient. But for some bizarre reason, we celebrate it as a point of honor in academia.
One counter argument to yours is that when you do need the skills, you can learn them later. It’s arguably easier than it has been at any point in human history. In that context, why front load people with something they hate doing, just because their parents think it’s a good idea? Let them wait and learn it when they need it.
People just have to want to like things. If they don't like something enough then a near-ubiquitous form of outsourcing is now available for them to get carried away with.
The "wanting to like things" is a highly undervalued skill/trait. It comes down to building a habit through repetition - not necessarily having fun or getting results, but training your mind like a muscle to think putting in effort isn't that bad an activity.
For those growing up I think this is not something that is taught - usually it is already there as a childlike sense of wonder that gets pruned by controlling interests. If education forcing you to do math removes any enthusiasm you had for math, that's largely determined by circumstance. You'd need someone else to tell you the actual joys of X to offset that (and I'd guess most parents/teachers don't practice math for fun), or just spontaneously figuring out how interesting X is totally on one's own which is even rarer.
I didn't have either so I'm a mathophobe, but I'm alright with that since I have other interests to focus on.
"Forcing" is a bit strong IMHO — I believe we've instead lost track of what is "passable", and everyone in higher education should be able to reach that and score a passing grade (D? C?).
Maybe professors are too stringent with their evaluation, or maybe they are not good at teaching people what a passable writing style is, or maybe students simply don't want to accept that if they don't excel at writing, a D or a C is perfectly fine. Perhaps teachers that look for good writing should have separate tests which evaluate students in both scenarios: with and without LLM help.
The same holds true for math: not everybody needs to know how to deduce a proof for every theorem, but in technical sciences, showing that ability and capability will demonstrate how much they are able to think and operate with precision on abstract concepts, very much like in programming. Even if coursework is a bit repetitive, practice does turn shallow knowledge into operational knowledge.
In most schools a D is not passing or at least doesn’t count as credit towards graduation. I’m not really sure what the point of that grade is to be honest.
Reading, writing and math have been the constants utilized throughout life and as such have been core subjects carried through educational systems. I'm not quite sure what subjects and topics we would be teaching future generations that didn't include reading, writing, math and science. At the very least writing should be included in more subjects. The hidden feature of including writing in all subjects, as you might have seen in your history endeavor's, is improvements in critical thinking, formulating cohesive arguments and a clearer understanding of topics.
There are greater difficulties that people will have to do in their daily lives than being "forced" to learn how to read, write and do arithmetic. Maybe learning the lesson of overcoming smaller, difficult tasks will allow them to adapt to greater difficulties in the future.
To quote Seneca:
A gem can not be polished with friction, nor a man perfected without trials.
People defending this are wrong in an additional, more pathetic way:
Even if you insist on “cheating” and using an LLM to communicate, you are using it badly. You manage to be obviously incompetent at using the tool you are evangelizing.
Maybe ridiculous from my part - but I think if they at least had to go hand written with these it could transfer a bit more knowledge back to the brain...
Simply blaming models is an easy way out and creates little value - Maybe changing the medium and exercise to which it transfers could be a thing?
I found that the book "Writing to Learn" by William Zinsser was excellent in convening this process. As noted in the book the author advocated for more writing to be included in all subjects.
Personally, I've used LLM to help me better structure my blog post after I write it. Meaning I've already written it, then it enhances it. Most of the time, I'm happy with the results at the time of editing. But when I come back a week or two to re-read it, it looks just like the example the author shared.
The goal is to make something legible, but the reality is we are producing slop. I'm back to writing before my brain becomes lazy.
[Edit: I agree] I've also grown to dislike even this use case. I did this back in 2023 but as AI text is spreading, the style - yes, even with prompt adjustments it leaks through - is recognized by more and more people and it's a very very bad look. If I see AI-like text from someone, I take it as an insult. It means they don't feel that it's worth their time to brush up the text themselves. And sure, it may well be that they don't value our interaction enough to spend the time on it. But that fact is indeed by itself insulting. So I only send AI touched up text to orgs that are so faceless or bureaucratic that I don't mind "offending" them.
I've grown to respect typos and slightly misconstructed sentences. It's an interesting dynamic that now what appeared lazy to 2021 eyes actually indicates effort and what appeared polished and effortful in 2021 now indicates laziness.
An example is how the admins of my local compute cluster communicate about downtimes and upgrades etc and they are clearly using AI and it's so damn annoying, it feels like biting into cotton candy fluff. Just send the bullet points! I don't need emojis, I don't need the fake politeness. It's no longer polite to be polite. It doesn't signal any effort.
> A typical belief among students is that classes are a series of hurdles to be overcome; at the end of this obstacle course, they shall receive a degree
Yes, totally. Unfortunately, it takes time and maturity to understand how this is completely wrong, but I feel like most students go through that belief.
Not sure how relevant it is, but it makes me think of two movies with Robin Williams: Dead Poet's Society and Will Hunting. In the former, Robin's character manages to get students interested in stuff instead of "just passing the exams". In the later, I will just quote this part:
> Personally, I don’t give a shit about all that, because you know what? I can’t learn anything from you I can’t read in some fuckin’ book. Unless you wanna talk about you, who you are. And I’m fascinated. I’m in.
I don't give a shit about whether a student can learn the book by heart or not. I want the student to be able to think on their own; I want to be able to have an interesting discussion with them. I want them to think critically. LLMs fundamentally cannot solve that.
LLMs and AI use create new dichotomies we don’t have language for.
Exploring a concept-space with LLM as tutor is a brilliant way to educate yourself. Whereas pasting the output verbatim, passing it as one’s own work, is tragic: skipping the only part that matters.
Vibe coding is fun right up to the point it isn’t. (Better models get you further.) But there’s still no substitute for guiding an LLM as it codes for you, incrementally working and layering code, committing to version control along the way, then putting the result through both AI and human peer code reviews.
Yet these all qualify as “using AI”.
We cannot get new language for discussing emerging distinctions soon enough. Without them we only have platitudes like “AI is a powerful tool with both appropriate and inappropriate uses and determining which is which depends on context”.
I hear you. But in this case, it seems like the author was mostly referencing academic uses of LLMs for either writing assignments or reviewing (academic) papers. Enterprise communications have their own carefully set requirements, but often they aren't meant to be instructive to the person writing them (assignments) or enhancing an existing corpus of knowledge (academic papers, optimistically).
The level of cheating in college, pre-AI, is often overlooked in these articles.
Pre-AI, homework was often copied and then individuals just crammed for the tests.
AI is not the problem for these students, it's that many students are only in it for the diploma.
If it wasn't AI it would just be copying the assignment from a classmate or previous grad.
And I imagine the students who really want to learn are still learning because they didn't cheat then, and they aren't letting AI do the thinking for them now.
Because everyone needs to finish university for jobs that 50-100 years ago were done by people with basic literacy and 4 years of primary school. And even some advanced jobs are basically trades, so trade schools to churn out paper pushers and coders, and so on, but leave the universities for scientists and academics.
Write a troll response in the style of Hacker News troll who wants to troll the Hacker news users who just read the article titled "I'd rather read the prompt". Make it ironic
Teachers lose much if not all of their time teaching while people applying what they've learned spend all of their time applying and advancing the practical side of it. The later don't even know how to use LLM's.
I teach a university class in which I ask the students to submit writing each week, and I have also seen obviously LLM-produced writing. Yes, it’s boring and doesn’t show the students’ thinking, and the students are not getting any wiser by doing assignments that way. Just last week, I told my students that, while they can use LLMs any way they like for the class, their writing will be more interesting if they write it themselves and use LLMs only sparingly, such as for fixing grammatical mistakes (most of the students are not native speakers of English). It helps, I think, that in this class the students’ writing is shared among the students, and during class I often refer to interesting comments from student writing. The students themselves, I hope, will come to understand the value of reading human-written writing.
That said, I myself am increasingly reading long texts written by LLMs and learning from them. I have been comparing the output of the Deep Research products from various companies, often prompting for topics that I want to understand more deeply for projects I am working on. I have found those reports very helpful for deepening my knowledge and understanding and for enabling me to make better decisions about how to move forward with my projects.
I tested Gemini and ChatGPT on “utilizing Euler angles for rotation representation,” the example topic used by the author in the linked article. I first ran the following metaprompt through Claude:
Please prepare a prompt that I can give to a reasoning LLM that has web search and “deep research” capability. The prompt should be to ask for a report of the type mentioned by the sample “student paper” given at the beginning of the following blog post: https://claytonwramsey.com/blog/prompt/ Your prompt should ask for a tightly written and incisive report with complete and accurate references. When preparing the prompt, also refer to the following discussion about the above blog post on Hacker News: https://news.ycombinator.com/item?id=43888803
I put the the full prompt written by Claude at the end of the Gemini report, which has some LaTex display issues that I couldn’t get it to fix:
I know nothing about this topic, so I cannot evaluate the accuracy or appropriateness of the above reports. But when I have had these two Deep Research models produce similar reports on topics I understand better, they have indeed deepened my understanding and, I hope, made me a bit wiser.
The challenge for higher education is trying to decide when to stick to the traditional methods of teaching—in this case, having the students learn through the process of writing on their own—and when to use these powerful new AI tools to promote learning in other ways.
Teachers say they would rather read the prompt but the truth is plain that they wouldn't
It's the old joke of the teacher who wants students to tried their best and that failure doesn't matter. But when the student follows the process to the best of their ability and fails they are punished while the student who mostly follows the process and then fudges their answer to the correct one is rewarded.
> Either the article is so vapid that a summary provides all of its value, in which case, it does not merit the engagement of a comment, or it demands a real reading by a real human for comprehension, in which case the summary is pointless.
There's so much bad writing of valuable information out there. The major sins being: burying the lede, no or poor sectioning, and just generally verbose.
In some cases, like in EULAs and patents that's intentional.
I wish the author had state out right that they were not using LLMs much, since their opinion on them and their output has no value (its a new technology, and different enough that you do have to spend some time with them in order to be able to find out what value they have for your particluar work[0].
The is especially the case when you are about to complain about style, since that can easily be adjusted, by simply telling the model what you want.
But I think there is a final point that the author is also wrong about, but that is far more interesting: why we write. Personally I write for 3 reasons: to remember, to share and to structure my thoughts.
If an LLM is better then me at writing (and it is) then there is no reason for me to write to communicate - it is not only slower, it is counterproductive.
If the AI is better at wrangling my ideas into some coherent thread, then there is no reason for me to do it. This one I am least convinced about.
AI is already much better than me at strictly remembering, but computers have been that since forever, the issue is mostly convinient input/output. AIs makes this easier thanks to speech to text input.
Quite likely, further progress will lead to LLMs writing "better" than at least 99% of humans.
I think this will be no more of a contest than playing chess has been: humans don't stand a chance, but it also doesn't matter because being better or worse than the AI is besides the point.
> ... their opinion on them and their output has no value
This is ridiculous. Even if the author has never typed a single character into a prompt box, he can still come to perfectly valid conclusions about the technology just by observing patterns in the outputs that are shoved into his face.
"I wish these astrophysicists had stated up front that they've never created a galaxy. How can they have a well-formed opinion on cosmic structures if they only ever observe them?"
I don’t know anything about the subject area, so I don’t know if this captures enough to get a good grade. But I’m curious if anyone could tell whether the last answer were AI generated if I copied and pasted. These are the iterations I go through when writing long requirement documents/assessments/statements of work (consulting).
Yes I know the subject area for which I write assessments and know if what is generated is factually correct. If I’m not sure, I ask for web references using the web search tool.
I mean this is not blind obviously, but it feels unnaturally enthusiastic/conversational to me. Maybe for like an video script it would fit, but for a requirements document or something it is a little oddly 'sauced up', as if someone put an extra pass through it to try to make it entertaining to read.
I was trying to make it sound like a college student with no strong writing experience.
I use to work at AWS (Professional Services) and there are a few different writing styles depending on what your audience was. I learned how to write in the different “house styles” before LLMs were a thing. So I know when something doesn’t sound right.
I use LLMs all of the time to write. I’m 99% certain that no one can tell the difference between my writing 100% without an LLM to my writing with one
> I didn’t realize how much that could throw things off until I saw an example where the object started moving in a strange way when it hit that point.
Would feel off, because why change the person? And even if it's intented, then I'd say it's not formal to do in an assignement.
These are art students not English writers. If I were a teacher I would think this is more authentic. LLMs don’t make this kind of mistake in its default house style.
100%. Students aren’t stupid, they can tell the difference between a lazily designed assignment that doesn’t deserve their full attention, and actually engaging learning moments/environments that will spark something inside them. It’s not their fault that the latter is so incredibly rare in course design.
Sounds to me like they asked the students to just regurgitate genetic course info and then complained when that’s what they received. This wasn’t going to lead to an excellent learning moment for these students whether an LLM was used or not.
> I should hope that the purpose of a class writing exercise is not to create an artifact of text but force the student to think
Back in HS literature class, I had to produce countless essays on a number of authors and their works. It never once occurred to me that it was anything BUT an exercise in producing a reasonably well written piece of text, recounting rote-memorized talking points.
Through-and-through, it was an exercise in memorization. You had to recall the fanciful phrases, the countless asinine professional interpretations, brief bios of the people involved, a bit of the historical and cultural context, and even insert a few verses and quotes here and there. You had to make the word count, and structure your writing properly. There was never any platform for sharing our own thoughts per se, which was sometimes acknowledged explicitly, and this was most likely because the writing was on the wall: nobody cared about these authors or their works, much less enjoyed or took interest in anything about them.
I cannot recount a single thought I memorized for these assignments back then. Passed these with flying colors most usually, but even for me, this was just pure and utter misery. Even in hindsight, the sheer notion that this was supposed to make me think about the subject matter at hand borders on laughable. It took astronomical efforts to even retain all the information required - where would I have found the power in me to go above and beyond, and meaningfully evaluate what was being "taught" to me in addition to all this? How would it have mattered (in specifically the context of the class)? Me actually understanding these topics and pondering about them deeply is completely inobservable through essay writing, which was the sole method of grading. If anything, it made me biased against doing so, as it takes a potentially infinite extra time and effort. And since there was approximately no way for our teacher to make me interested in literature either, he had no chance at achieving such lofty goals with me, if he ever actually aimed for them.
On the other side of the desk, he also had literal checklists. Pretty sure that you do too. Is that any environment for an honest exchange of thoughts? Really?
If you want to read people's original thoughts, maybe you should begin with not trying to coerce them into producing some for you on demand. But that runs contrary to the overarching goal here, so really, maybe it's the type of assignment that needs changing. Or the framework around it. But then academia is set in its ways, so really, there's likely nothing you can specifically do. You don't deserve to have to sift through copious amounts of LLM generated submissions; but the task of essay writing does, and you're now the one forced to carry this novel burden.
LLMs caught incumbent pedagogical practices with their pants down, and it's horrifying to see people still being in denial of it, desperately trying to reason and bargain their ways out of it, spurred on by the institutionally ingrained mutual-hostage scenario that is academia. *
* Naturally, I have absolutely zero formal relation to the field of pedagogy (just like the everyday practice of it in academia to my knowledge). This of course doesn't stop me from having an unreasonably self-confident idea on how to achieve what you think essay writing is supposed to achieve though, so if you want a terrible idea or two, do let me know.
The suggestion that an artificial intelligence follows a specific kind of writing style is a trap.
Relying on that to automatically detect their use makes no sense.
From a teaching perspective, if there is any expectation that artificial intelligence is going to stick, we need better teachers. Ones that can come up with exercises that an artificial intelligence can't solve, but are easy for humans.
But I don't expect that to happen. I expect instead text to become more irrelevant. It already has lost a lot of its relevancy.
Can handwriting save us? Partially. It won't prevent anyone from copying artificial intelligence output, but it will make anyone that does so think about what is being written. Maybe think "do I need to be so verborragic?".
> I say this because I believe that your original thoughts are far more interesting, meaningful, and valuable than whatever a large language model can transform them into.
Really? The example used was for a school test. Is there really much original thought in the answer? Do you really want to read the students original thought?
I think the answer is no in this case. The point of the test is to assess whether the student has learned the topic or not. It isn’t meant to share actual creative thoughts.
Of course, using AI to write the answer is contrary to the actual purpose, too, but it isn’t because you want to hear the students creativity, but because it is failing to serve its purpose as a demonstration of knowledge.
Because you want to pass on knowledge? I am not saying there aren't ANY situations where a teacher cares about what their students think, but the example given isn't really one of those times. The question is not one that has many opportunities for original thought; it is a basic question that everyone who knows the answer will answer similarly. The entire purpose is to ascertain if the person understands what was taught, it isn't meant to engender a novel response.
Arguably, that's not what teachers mainly do (to an ever increasing proportion).
Most knowledge is easily available. A teacher is teaching students to think in productive ways, communicate their thoughts and understand what others are trying to tell them. For this task, it's essential that the teacher has some idea what the students are thinking, especially when it's something original.
Sure, but my contention was more with the word “original”, because they aren’t really original thoughts. The teacher just wants to make sure the student’s thoughts contain the information they are teaching. The teacher isn’t looking for actual original thought in this test.
Teaching isn't blatting someone's mind with the Correct Answers and checking whether the overwrite took. Instead, you empirically build a model of the student's own world model, then examine the model to see where it fails to conform to reality, and construct ways to fix it. The example from TFA was consistent with "this is university coursework", which is fertile ground for a tutor to identify such errors! It's not like the student should be receiving a yes/no response and nothing else; there will presumably be comments on the incorrect answers.
It's not a "test", it's an "assignment". Assignment is a way to practice what you've learned, and a (good) teacher would want to get your original thoughts so they could adjust their instruction and teaching material to what they believe you missed in order to improve your mental model around the topic (or in other words, to teach you something).
Perhaps the problem is that they are "graded", but this is to motivate the student, and runs against the age-old problem of gamification.
It's actually doesn't matter. I hated this hassle of writing various texts while studying so much. Like does it really matter whether student would generate this text or just go google and copy paste some paragraphs from somewhere? And don't even hope for them to genuinely write all that stuff themselves because it's a huge waste of time even for those who actually cares and interested in the subject.
Yup, people like to get high and mighty but they’ve obviously forgotten how much of school was pointless assignments designed so the teacher could tick off the boxes to their employer that they’d done something rather than having any interest in that assignment actually inspiring learning. Students aren’t stupid, they spot this behaviour a mile off and simply respond in kind.
It's been incredibly blackpilling seeing how many intelligent professionals and academics don't understand this, especially in education and academia.
They see work as the mere production of output, without ever thinking about how that work builds knowledge and skills and experience.
Students who know least of all and don't understand the purpose of writing or problem solving or the limitations of LLMs are currently wasting years of their lives letting LLMs pull them along as they cheat themselves out of an education, sometimes spending hundreds of thousands of dollars to let their brains atrophy only to get a piece of paper and face the real world where problems get massively more open-ended and LLMs massively decline in meeting the required quality of problem solving.
Anyone who actually struggles to solve problems and learn themselves is going to have massive advantages in the long term.
reply