I find this way of looking at LLMs to be odd. Surely we all are aware that AI has always been probabilistic in nature. Very few people seem to go around talking about how their binary classifier is always hallucinating, but just sometimes happens to be right.
Just like every other form of ML we've come up with, LLMs are imperfect. They get things wrong. This is more of an indictment of yeeting a pure AI chat interface in front of a consumer than it is an indictment of the underlying technology itself. LLMs are incredibly good at doing some things. They are less good at other things.
There are ways to use them effectively, and there are bad ways to use them. Just like every other tool.
The problem is they are being sold as everything solutions. Never write code / google search / talk to a lawyer / talk to a human / be lonely again, all here, under one roof. If LLM marketing was staying in its lane as a creator of convincing text we'd be fine.
This happens with every hype cycle. Some people fully buy into the most extreme of the hype, and other people reverse polarize against that. The first group ends up offsides because nothing is ever as good as the hype, but the second group often misses the forest for the trees.
There's no shortcut to figuring out what the truth of what a new technology is actually useful for. It's very rarely the case that either "everything" or "nothing" is the truth.
I blame the experts. It's their responsibility to explain things to the public and engage in forums that the public is paying attention to (e.g. podcasts). They don't have to bloviate about everything under the sub, but they do have to be able to break down and communicate their ideas to the non-expert public. Failure to do so creates a vacuum that is filled by the Marc Andreesens and Peter Thiels of the world.
If you go on Marc’s Twitter he spends most of his time subtweeting with emojis and one word responses. And he has millions of followers (for what reason?).
A scientist, aside from their day job, is now also supposed to spend time debunking whatever half baked topic of the day is?
The only world where that works is one in which MA’s reputation is built on not saying dumb stuff all the time, like a scientist’s reputation is. If his follower count dropped for example. But it’s not, and that’s not how it works. People like him will move on to the next thing tomorrow.
Absolutely not. That turns the experts into politicians and pundits. Experts should stay in their lane and provide accurate and trustworthy information.
Yes, it should be accessible and digestible, but should not be pushed.
Yeah I definitely struggle with this. You need downtime to relax but it's easy to "over relax" just like it's easy to oversleep or overeat or overdo any other number of things that are healthy and necessary but only at the right amplitude and frequency. I think that's why it can feel so good to be in a rhythm. You get a nice oscillation going that rides the wave of momentum instead of some monotonic rise or fall that is going to lead to burnout or stagnation.
This is the first time I've heard anyone argue that a food product must be good for you because Americans are consuming a large amount of it. How on earth could you come to that conclusion given how unhealthy our population is?
Have you ever tried the LeetCode live competitions? I found those to be really fun with a great community. Just grinding problems in isolation can definitely be depressing.
I agree generally, but I feel like we're slowly coming to realize that maximal leisure and safety might not necessarily be the recipe for a happy and fulfilling life.
This is more or less the point of Ted Kaczynski’s manifesto. Technology is essentially evil as it replaces the things that truly connect man to his nature. That technology has had a destabilizing effect on society, has made life unfulfilling, and has caused widespread psychological suffering. Most everything we do now is a “surrogate” activity” to try and get control/power over ourselves and find meaning, which has been vaporized by technological progress.
In particular the Industrial Revolution but he himself sees an ideal world basically void of technology. That the revolution should strive to destabilize industrial society enough to force its collapse.
I think he has many interesting observations that are really worth exploring. He also predicted a lot of things more or less. But end of the day life is what you make of it. You don’t really have to participate in modern society. And people have always had existential problems and a desire to find meaning, even if they had less time to ponder it (and possibly be antagonized by it) because they were tending a field, etc. They probably had a long time to think about it as an infection slowly ate its way through their body causing immeasurable pain over days or weeks.
By certain angles it is even less important than spiritual, community and philosophical purpose. It's not something that you can measure directly but by many possible proxies (Amish levels of depression, the Roseto effect, various studies on happiness of underdeveloped countries), people need very few material things to be happy. Perhaps it is a kind of mass delusion or confirmation bias that happiness must correlate linearly, or even logarithmically, with disposable income.
Arguably we write instructions: instead of writing out the problem and what the solution looks like, we describe a set of steps we go through—and if those steps are incorrect, there's nothing to compare against, because that was what we called the "specification".
Whether there's a difference there is in the eye of the beholder, but it does look like that specification languages such as TLA+/PlusCal/Squint or Alloy, or theorem proving languages like Coq (to be renamed Rocq) or Lean look a lot different from the likes of C, JavaScript or even Haskell.
Did the site owner remove the malicious ads since this was posted? I wanted my report to count so I didn't use the link in the blog post and instead googled for "notepad++ download" and clicked on the offending website which was ranked third for me. I didn't see any ads on any pages. I don't doubt the complaint but some screenshots and timestamps would be helpful since it's so easy for the site owner to cover their tracks.
Yup, I remember when I got a loaner car that had those new momentary switches and I couldn't believe how terrible they were. What made it all the more baffling/disappointing was that earlier BMWs had some of the nicest turn signal switches available that were so satisfying to use.
reply