Sorry, but a new prompt for GPT-4 is not a paper • Hacker News Board

Re: Sorry, but a new prompt for GPT-4 is not a paper

#181

Post by Intralexical » Wed, Dec 06, 2023, 3:07 AM UTC

elif wrote: post #166

daveguy wrote: post #151

Earlier quoted context omitted.

No prompt will cause an LLM to rapidly improve itself, much less into an AGI. Prompts don't cause permanent change in the LLM, only differences in output.

You're talking about how GPT functions in 2023. I am discussing such a point where when LLM outputs become valuable LLM modifications. AI recursing on itself progressing toward an AGI.

> AI recursing on itself progressing toward an AGI.

Inbreeding LLMs will result in an "AGI"?

In this universe, if you try to get something from nothing, you just end up with noise.

Re: Sorry, but a new prompt for GPT-4 is not a paper

#182

Post by mewpmewp2 » Wed, Dec 06, 2023, 8:02 AM UTC

dartos wrote: post #50

IanCal wrote: post #43

Earlier quoted context omitted.

I'm sorry, but that's entirely ridiculous. You're mangling up a concept of burden of proof here. You can easily see this because it can be flipped around easily - you made a claim that they are being changed, even every few weeks! Should it really be on me to show that your very specific claim is false? Aside - but even if the model weights did change, that wouldn't stop research being possible. Otherwise no drug tri…

Apples and oranges comparison. You couldn’t get the same participants, but you could get the same drugs. If you could get identical participants, that wouldn’t be very helpful since humans are so varied. But for GPT based papers, what you’re actually testing could change without you knowing. There’s no way to know if a paper is reproducible at all. If you can’t reproduce results, is it really research, or just show a…

With API you can choose versions fixed to a date. Are you suggesting that OpenAI is lying about these being fixed to a date?

Why would they lie about it?

The whole point of these versions is so that when you build on top of that it would keep working as you expect.

Re: Sorry, but a new prompt for GPT-4 is not a paper

#183

Post by d4rkp4ttern » Wed, Dec 06, 2023, 1:20 PM UTC

Yes these papers are optimizing for social media hype, I.e what is the quickest and easiest path to make noise on social media ?

Re: Sorry, but a new prompt for GPT-4 is not a paper

#184

Post by hnfong » Wed, Dec 06, 2023, 2:27 PM UTC

mewpmewp2 wrote: post #182

dartos wrote: post #50

Earlier quoted context omitted.

Apples and oranges comparison. You couldn’t get the same participants, but you could get the same drugs. If you could get identical participants, that wouldn’t be very helpful since humans are so varied. But for GPT based papers, what you’re actually testing could change without you knowing. There’s no way to know if a paper is reproducible at all. If you can’t reproduce results, is it really research, or just show a…

With API you can choose versions fixed to a date. Are you suggesting that OpenAI is lying about these being fixed to a date? Why would they lie about it? The whole point of these versions is so that when you build on top of that it would keep working as you expect.

I'm not saying they lie about it, but as a hypothetical there could be many reasons to lie.

- realizing their model leaks confidential information against malicious prompts

- copyright claims against them forcing them to remove bits of data from the training set

- serious "alignment" bugs that need to be fixed

- vastly improved optimization techniques that slightly affect results in 0.1% of the cases

If updating the model would save the company a couple hundred million dollars, they might want to do it. And in some of the cases, I can imagine they have an incentive to keep the update low key.

Re: Sorry, but a new prompt for GPT-4 is not a paper

#185

Post by ableal » Wed, Dec 06, 2023, 2:40 PM UTC

Der_Einzige wrote: post #30

ableal wrote: post #17

Earlier quoted context omitted.

> Because sufficient amounts of quantity always beats quality Hegel (as echoed by Marx): "merely quantitative differences beyond a certain point pass into qualitative changes" ( https://www.pnas.org/doi/10.1073/pnas.240462397 ) I always found that an interesting observation, whatever you think of the rest of their works.

Hegel is a hack and doesn’t deserve to be cited here. Consider that he’s really only famous because his class motivated a bunch of other philosophers (the young Hegelian's, Marx, Stirner, Bruno Bauer et al) to meet in wine bars after class to complain about how terrible/impossible to understand his philosophy is. Anyone whose even tried to read stuff of his I.e the phenomenology of spirit will tell you that he’s a ch…

Hey, can't a blind pig find an acorn? Is it less of an acorn because it was found by a pig?

Myself, I appreciate Schopenhauer a lot more, and we know what he thought of Hegel, but I'm not fanatical about it. If a hack hits on a good line, I'll nab it.

Re: Sorry, but a new prompt for GPT-4 is not a paper

#186

Post by u32480932048 » Wed, Dec 06, 2023, 3:41 PM UTC

ChatGPT droppings have to be at least as relevant and newsworthy as these findings [1]

  [1] https://www.wbur.org/news/2022/07/27/harvard-shorenstein-research-january-6-insurrection-president

Re: Sorry, but a new prompt for GPT-4 is not a paper

#187

Post by IanCal » Wed, Dec 06, 2023, 11:21 PM UTC

starbugs wrote: post #174

IanCal wrote: post #155

Earlier quoted context omitted.

> If you can’t reproduce results, is it really research, or just show and tell? You can't start with a statement about clinical trials not being perfectly reproducible and that's fine, then say this. > what you’re actually testing could change without you knowing If people are lying about an extremely important part of their product, which they have little reason to. But then this applies to pretty much everything. S…

> You can't start with a statement about clinical trials not being perfectly reproducible and that's fine, then say this. Nobody said that to begin with. Re-read their comment. > If people are lying about an extremely important part of their product [...] Nobody is alleging that anyone is lying. It's just that we cannot be sure what the research actually refers to, because of the nature of a proprietary/closed model.…

> Nobody is alleging that anyone is lying.

Let's be clear, you have made an explicit claim that openai are lying.

> What does this have to do with the parent comment?

Because many other fields would kill for this level of reproducibility, grab an API key, spend a few quid running a script and you can get the results yourself.

Re: Sorry, but a new prompt for GPT-4 is not a paper

#188

Post by IanCal » Wed, Dec 06, 2023, 11:23 PM UTC

kjkjadksj wrote: post #179

sebzim4500 wrote: post #25

Earlier quoted context omitted.

If I book telescope time and capture a supernova then no one will ever be able to reproduce my raw results because it has already happened. I don't see why OpenAI pulling old model snapshots is any different.

But you know the parameters of your telescope at least. If openai wants to update all the time, fine, but then they should be like how every other piece of research software works, where you can list what exact version of software you used and pull that version yourself if need be.

Stability is the purpose of the versioned models.

Re: Sorry, but a new prompt for GPT-4 is not a paper

#189

Post by gowld » Thu, Dec 07, 2023, 5:05 AM UTC

j2kun wrote: post #140

blueboo wrote: post #133

Earlier quoted context omitted.

I’m not sure the scientific method itself can withstand this sort of scrutiny. After all, it’s just making guesses about what will happen and then seeing what happens!

Except there's also, you know, building coherent theories and using those theories to predict the system behavior.

That is in no way a requirement for doing science.

Re: Sorry, but a new prompt for GPT-4 is not a paper

#190

Post by nicce » Thu, Dec 07, 2023, 1:34 PM UTC

roguas wrote: post #128

nicce wrote: post #118

Earlier quoted context omitted.

> That’s like saying a biologist studying an endangered species isn’t doing science because the animal could disappear tomorrow. The permanence of a subject has no bearing on whether it is science or not. The idea that science has to happen in a lab is of course absurd as well. The main point here is that anyone can likely start studing those endangered species and try to reproduce the results while in GPT4 it is not…

Whats not possible? Do they ban people for exploring sota model that they offer?

That is not reproducible research. To reproduce the research, you need training data, source code and all the parameters.

With black-boxed generated models, there is no way to tell how they have been actually generated. It has no value for science to how to improve it further.