Nice. High chance we’ll start seeing a ton more tooling innovation of this kind.
I view gpt-4s information constraints very similarly to blockchains; it’s a bounded context. Blockchain solved this with oracles. I think there will be a more formal, widely recognized variant for gpt-4 pretty soon. Your feature for pulling down websites is basically it.
I found the context size for gpt-3 to be extremely limiting when I was hacking at ideas. Seeing this and knowing the context size for gpt-4 is larger is giving me motivation to pick up where I left off.
I had the idea of creating a Kubernetes cluster debugger that’s gpt-4 based and runs in the cluster, consuming and troubleshooting then posting results to a Slack channel for dev teams to view. Don’t think I’ll work on it though.
I actively reject conspiracy theories on the premise that they’re conspiracy theories. However, I can’t deny that this one is interesting. What if this is actually happening?
Obviously the easiest way to provoke thought within the skeptic is to ask, “If it is within the realm of possibilities for the human race to invent a driver allowing interstellar travel, why shouldn’t an extraterrestrial race also be capable of it?”
I would assume someone would counter with the fact that extraterrestrial life hasn’t been discovered. A question to often revisited. For all we know extraterrestrial life surrounds their inhabited planets with strict ingress and egress policies.
I don't think what the OP is suggesting is necessarily a conspiracy. Institutions are wary of people who risk their credibility with claims that catch a lot of attention but may prove to be false. There's some incentive to make claims that capture a lot of attention and the field doesn't want to lose credibility.
If you think about how some scientists responded to speculation about Oumuamua you can see how they were nervous about speculation that could be sensationalized.
Airpods 100% affect / cause tinnitus. I tracked it on myself over several months once I noticed the relationship. In short, the airpods caused it and using them substantially intensifies the tinnitus I experience. I’m 24. Saddens me to know I’ll be living with this for the rest of my life.
Stumbling upon the post linked by OP after realizing the connection between my tinnitus levels and airpods is hard to put into words. The tinnitus developed during my first year using airpods. I would use them for at least an hour a day and always towards the lower 1/4 of the volume bar.
Not if you’re a for-profit VC firm looking to invest in an area with changing regulation, which could lead to trillions of dollars of economic opportunity.
Many instances, and in some cases it was more like months or even years.
One large government department struggles to hire skilled application architects, because they can only pay public service rates in a highly competitive market. The only hires they can get are 6 month contracts for relatively inexperienced architects who use the job to pad out their resume and then immediately move on to private industry. To make their resumes look good, they throw the kitchen sink at every design.
A recent project was a simple web app that's an extension to an existing web app. It implements trivial things such as adding sign-in to an otherwise static site, some email notifications, mobile push, reporting, and the like. Nothing fancy.
They were a couple of months into development, and they already had:
- Multiple web apps
- Multiple API apps
- A "mid-tier" behind the above
- Dozens of lambdas/functions
- Logic apps and workflow apps
- Multiple databases
- Multiple Kubernetes clusters despite most of the above running on a different PaaS platform already.
- ETL jobs for copying data between databases that can directly access each other.
- All of the above stretched across both Azure and AWS, just for funsies.
Etc...
It looks really impressive on a high level architecture diagram! Fantastic in a resume.
But it's a ton of work for a developer to implement, and achieves nothing. Microservices and lambdas/functions have their place, but not for a single-person web dev project!
Throwing out 90% of the "moving parts" from the design probably saved a year or two all by itself.
Then I started demonstrating how to use storage emulators for local development, how to manage dev/tst/uat/prd configurations, how to get logs and debug dumps from the PaaS environments in the cloud, etc...
That workflow stuff alone would save weeks or months, easily.
All this took was just a couple of hours of chatting.
Another meeting was less than an hour, and while the guy was scrolling through his code I yelled "Stop!" because I saw a reference to a low-level cryptography primitive. He had started trying to implement authentication and authorisation "from the ground up". I suggested using a popular JWT library instead.
Just showing a developer how to use an APM tool like Application Insights or New Relic can cut weeks off the troubleshooting time over the lifetime of a project.
I just copy and pasted 3 random Leetcode problem prompts to GPT-3. It successfully generated Python code that passed all test cases for 2 out of the 3.
Problems passed:
- Two sum
- Text Justification
edit: Newlines
I agree with your side. Here’s my take. For some people, there’s a line you don’t cross. However, if the line is going to be crossed anyway, why not cross it in the best way possible? I’m inclined to believe that if done well, even those against it being done will end up finding appreciation for the next works.
This whole situation happened with Robert Jordan and The Wheel of Time series. It was considered a mess by fans but was a critical success.
Either way at least we got an ending. Thing is- the books had been meandering for a while, and it's quite possible Jordan would have whiffed at the end, but we will never know.
With Tolkien it's just been a shallow money grab since The Hobbit trilogy made off a super short book and after the Amazon series stuff I'm over it anyway.
It feels like at this point they should have people write universe books like Forgotten Realms or something. New books by new authors in the Tolkienverse.
ASOIAF will probably never be finished, and I'm OK with that.
I used to hate not getting to see how a story ends but after so many books/shows/series/movies always ending on cliffhangers to never be completed, I'm just over it.
Seems even storytellers die heroes or live long enough to become villains ;)
> Wheel of Time ... was considered a mess by fans but was a critical success.
Can you go into why it was considered a mess? I'm a relatively recent finisher of the series so I can't speak to the contemporary reaction of the fanbase, but my impression has been that the current community is, at a minimum, satisfied with Sanderson's work. There are several things that people complain frequently about (and that I agree with) like Mat's character development and the Androl arc. Not to mention the lost possibility of follow-on series after A Memory of Light! That doesn't mean people think it's a mess, though.
It's just what I remember from release. We'd waited so long. I haven't even read it since. Checked reviews and it's sitting at 4.5. I will have to reread the series anyways, maybe it was just some of us never being satisfied, I know I've been that guy before.
I personally was glad for the ending. My main critique was with the series as a whole meandering wildly. But was still something I loved.
I believe they had lots of notes and a roadmap from Jordan as well, but I could be mis-remembering.
One weird series that had an ending that wasn't supposed to was the Gunslinger series. I think King went like 20 years between books before ending it. I really remember it because he had a note when you got to the ending saying he didn't feel the story should have an ending, and he recommended stopping there, but for those who needed an ending he provided one to give some closure.
I thought that was so cool, and read it, but prefer to think Roland is still out there somewhere chasing some new Walter, so rather than preferring an ending that time I preferred open-ended, which almost never happens.
In spite of something I wrote elsewhere--basically that sometimes it's just time to move on--I tend to agree. While it may be blasphemous to say, much as I love it, LoTR is not without its flaws--especially the whole structure of The Return of the King. Tolkien was a world building (and linguistic) genius but LoTR hasx its flaws.
This is very similar to my experience. TBH, being able to not talk through what I’m coding the _whole_ time, unlike what it shown in the official “What to expect” Google videos, was a relief. Unfortunately, that change in expectation sort of threw me off; there became uncertainty in what was expected of me. This uncertainty hindered my creative juices while problem-solving but nowhere near as bad as the interview prior, which I won’t go into since I am paranoid about the interviewer seeing this comment :)
I found the context size for gpt-3 to be extremely limiting when I was hacking at ideas. Seeing this and knowing the context size for gpt-4 is larger is giving me motivation to pick up where I left off.