The tech industry is guilty of fanning the flames of hype to get funding. A lot of people are rightly annoyed over some of the more outlandish claims that were made over the past 1-2 years.

Such as what? I think this is like a massive invention and presages massive changes in the world at least on the decades-timescale. Big claims are warranted.

Come on man I'm not going to go through countless blogposts from a year ago and then argue with you about them if they were overhyping AI or not. But if you recall the discussion a year ago was apocalyptic, both for the human race and software engineering jobs.

Yah-- what we have now is enough to be enormously disruptive as it is absorbed.

The big question is how much better it's going to get. We may be plateauing in capability-- or at least slowing in capability growth-- or we may not be.

But it doesn't need to get better for it to eat a lot of peoples' lunch.

Even if we are plateauing in GPT number (4,5, 12, whatever), there is still a lot of meat left on those bones.

For instance: Chemists using the AI backwards. Take some phrase like 'vanadium increases the Young's modulus versus silicon in 1040 steel' and then work the AI backwards from that phrase. As in, assume the AI outputted that phrase and see what inputs were most likely to generate it. There may be some real discoveries just by working an AI backwards iteratively.

Simple things like that are still open and attainable right now, it's just that AI is still so young that we really haven't explored all of what they do yet.

My opinion shifts every other day it seems, recently, Roon, the "famous" anon X account that works at OpenAI, and presumably has access to some unreleased info, said, in response to "are we plateauing soon", "not even close". I honestly don't know what to think.

OTOH Altman has said that next release will be about same advance on GTP-4 as that was over GPT-3. Dario Amodei has said something roughly equivalent - to expect solid gains this year but no reality bender.

I think the real potential gains come when they start extending the architecture, but as long as it's just scaling up and different training/inference regimes, then it seems it'll be more of the same rather than game changer.

Given how connected the whole SF/AI scene appears to be (alchohol + drugs too?), it's hard to imagine a company the size of OpenAI not leaking. If there had been any amazing discoveries there, I think there'd at least be rumors (and not just "they seem to have something called Q* going on").

I feel like for LLM's at least, we've not seen a lot of progress on the high end for a year or so. GPT4-Turbo is a little better than GPT4, but mainly only for things that can benefit from huge contexts.

We have seen progress, but mostly price progress.

Yes, and also not-always-bumping-into-capacity-limits (availability) progress.

Of course somebody working for OpenAI would say that.

I know his real identity and I doubt there is some key insider info there although he does work at OAI

>No matter how much the chattering class wants to portray this as a "fad" and not an inevitability, those of us who can type and read can see the reality for ourselves.

Right. Regardless of any "superhuman" abilities, we can talk to our computers now, and they can talk back. The decades old holy grail of HCI has been achieved. That fact alone is going to change everything.

Unfortunately human language remains ambiguous and imprecise. That's why we invented programming languages.

Exactly. When I read stuff like this, I'm always wondering "Have you actually used ChatGPT?" Both of the following can be true, and it's annoying when folks pretend it's either/or:

1. At present, AI can be massively overhyped, especially by salespeople who have an incentive to overhype it.

2. Current generative AI has made leaps and bounds improvements in the past few years, and it's present capabilities provide an enormous productivity tool for those who use it intelligently.

I mean, no, I don't think GPT 4 is AGI, nor do I really think it's that close. That said, every time I use it I'm amazed at how uncannily good it is, and it is able to save me a ton of time.

Yeah, chatbots have completely changed the way I program, and I'm not a "self-mythologizing executive and founder in Silicon Valley", also it's not only very useful but also entertaining, quite a lot of fun creating stuff with my friends using Suno or text-to-image models.

I’ve yet to find its programming help useful. ChatGPT and Copilot routinely offer bad or outright incorrect advice.

I think a big problem looming on the horizon for using llms to help with code is the same "confidently wrong" tone as when they're used as a general search engine. I've seen people blindly follow what they were being told by the llm when, if you'd just read it, it was obviously wrong code (or in one case a json config file). It reminds me of the copy/paste problem from stackoverflow except there's no voting or feedback from others that signify a more correct answer.

LLMs sound so sure of themselves and people think "well i'm dealign with the most advanced technology ever so it must be right...".

On the implementation side of things, it's hard for me to get the non-deterministic aspect of llms right in my head. I put an LLM and RAG system in prod with a team and went through rounds of the usual testing. 99 times it passed but on test 100 it would fail, so you'd adjust the system prompt. Then it'd pass 500 times and fail at 501. Adjust the system prompt, then it would pass 9 times and fail at 10. That system went to production but there's the low level worry in my mind, when is it going to fail to give the correct output? The fact that you can never guarantee the output of an LLM from a given input severely limits where they should be used IMO. I don't think it's wise to have the output of an LLM be the input to another program, there's no functional relationship between domain and range with an LLM.

That problem is usually met with "well, a human would make the same mistake.." but the reason computers exist is to do long, tedious, lists of tasks/instructions very fast that humans get wrong. Simulating a human, and all those imperfections, with digital logic seems contradictory to me.

edit: Also, just want to point out that the "testing" mentioned in my post was all manually done by humans. You can't automate testing the response of an llm unless you use another model to grade the response as correct or not but then you're right back to not being able to trust that the grader will always act consistently.

Claude 3 Sonnet is free like ChatGPT-3.5 but much more powerful, for what I need to do these chatbots + GPT-4-turbo can do most of what I need, from scripts for managing files to templates, rewriting code, actually solving difficult tasks (when it's too late for my brain to process), the more efficient way to use a library etc...

Try telling it it's a project manager and get it to write a scope of work/spec document first that you agree with. Then send it to a "software architect" to define the frameworks and architecture, and then to a "developer" to write specific files or modules.

There's a crazy theory that negative AI posts are foreign agents or even big tech trying to slow down progress of AI so they can catch up.

Crazy theory that runs afoul of HN guidelines :)