Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

OpenAI’s policies hinder reproducible research on language models

aisnakeoil.substack.com

Post reply on HN
Print view View original

311–320 of 394 posts

Re: OpenAI’s policies hinder reproducible research on language models

#311

Post by itsaquicknote »

itsaquicknote wrote: post #275
z3c0 wrote: post #242

Earlier quoted context omitted.

I thought I must be going crazy until I saw your comment. This sounds like a bad research practice that probably shouldn't be reproduced to begin with.

Research into systemically important infrastructure cannot be damned because that infrastructure isn't public. It's a cheap moralizing argument to say "pfff, this was predictable". Maybe so, but there isn't an alternative. Much like research on Twitter. Once these companies start to drift into providing what become broadscale social utilities and public services it doesn't matter that they're private. There are(/shou…

Replying to both responses because they're all good points. My argument boils down to the fact that some private companies end up becoming social utilities and once that happens, the rules (should) change as part of the social contract which means, yeah, they can't simply "pull the rug". The research is important precisely because its into systemically significant systems.

I get that it's difficult to define the line where that gets crossed. But the idea to provide a publicly funded trust that manages legacy versions of things like this is not a bad idea.

Re: OpenAI’s policies hinder reproducible research on language models

#312

Post by ben_w »

modriano wrote: post #309
ben_w wrote: post #300

Earlier quoted context omitted.

> who elected these elites to decide what was and wasn't ethical and responsible? Nobody First: basically every American literally voted for that by repeatedly saying no to the alternative (the communist party) in every American election. Second: what exactly and specifically are you suggesting here? Because even outside of capitalism, the alternative to "people deciding they personally don't feel it's safe to releas…

Two notes: 1) less than half of Americans vote in each election (less than 63% if you restrict to the voting-age population, less than 70% if you apply the scummy rules that restrict to the voting-eligible population) And 2) it's a false dichotomy to say that US elections have ever been "whatever we have now VS communism". Maybe you could say socialism was on the ballot all those times Eugene Debs ran for the preside…

> it sounds like you would struggle to define communism if pressed

Having read the Communist Manifesto, I think that description of me is both totally fair and would also apply to Karl Marx.

Darn thing read like an unhinged run-on blog rant.

Re: OpenAI’s policies hinder reproducible research on language models

#313

Post by ghoogl »

gwd wrote: post #197
danpalmer wrote: post #186

Earlier quoted context omitted.

And those suggestions would be very in-line with the original purpose of OpenAI. A purpose they are now actively hindering in the name of profit.

I think what most of the people here are missing is how big, how paranoid, and how influential the "AI alignment" movement is. To you it looks like they're being overly careful and paranoid, perhaps as an excuse to set up a monopoly silo to extract money. But a lot of the people the OpenAI researchers work closely with -- people deep in the "AI alignment" community -- are telling them that they're being wantonly reck…

[dead]

Re: OpenAI’s policies hinder reproducible research on language models

#314

Post by typest »

whatshisface wrote: post #120
sillysaurusx wrote: post #60

Earlier quoted context omitted.

Au contraire, no one knows how large GPT-4 is, which is the single best predictor of performance (for a model trained to convergence). The GPT-4 paper spent much of its time writing about this — they did some small scale experiments with 1/1000th the compute, then picked a loss level they wanted and trained GPT-4 till it got it. Neither the exact loss level nor the number of parameters are revealed by the paper. Unfo…

I can't believe anyone considers a single number, which would work about equally well if it were 10% higher or lower, to be a trade secret.

This is not really true. The Chinchilla paper showed that a 4% difference in loss between Chinchilla and Gopher led Chinchilla to blow Gopher out of the water at most tasks, including 30x performance in physics.

Empirically, LLMs have shown to have emergent abilities appear at different loss levels. So, a 10% difference could really matter.

Re: OpenAI’s policies hinder reproducible research on language models

#317

Post by jimmydorry »

gwd wrote: post #246
bheadmaster wrote: post #228

Earlier quoted context omitted.

Applying your own reasoning, what evidence would convince you that every money-making industry is necessarily driven by profit?

First, I specifically said (emphasis added): > There are people in that community -- people not working for a for-profit company -- who would, if they could, stop all AI research of any kind until we have rock-solid techniques to prevent an AI apocalypse. Most of those individuals have absolutely nothing commercial to gain from stopping AI research. Dalewyn's response implicitly said that even these people have a fin…

They withheld GPT-4 for eight months, but continued development based on it and provided access to third parties and entered into agreements with the likes of Microsoft/Bing, etc. All they did was impair their competition that were still struggling to catch-up with their previous offering, while continuing to plow ahead in the dark.

Re: OpenAI’s policies hinder reproducible research on language models

#318

Post by pixl97 »

marcyb5st wrote: post #251
sva_ wrote: post #22

Since OpenAI didn't release the parameter count of GPT-4, I've been wondering/doubting if it is really much bigger than GPT-3. The release of GPT-3.5 has shown that they've found ways of drastically cutting down compute costs (an order of magnitude) while maintaining or even improving the quality of the model's outputs. Perhaps the reason that they didn't release the specifics of GPT-4 might be in part due to them wa…

I am not sure how much bigger, but definitely much bigger IMHO. Otherwise you wouldn't be capped at 25 requests every 3h. That number is small enough that makes me think the inference costs/hardware needed are much bigger than 3.5.

I believe I heard that running inference longer is giving the better responses we're seeing in v4. Hence why v4 is taking so much longer to output data.

Of course we won't know this for sure until OAI tells us, so we may be in the dark for a while.

Re: OpenAI’s policies hinder reproducible research on language models

#319

Post by pixl97 »

zarzavat wrote: post #181
sva_ wrote: post #22

Since OpenAI didn't release the parameter count of GPT-4, I've been wondering/doubting if it is really much bigger than GPT-3. The release of GPT-3.5 has shown that they've found ways of drastically cutting down compute costs (an order of magnitude) while maintaining or even improving the quality of the model's outputs. Perhaps the reason that they didn't release the specifics of GPT-4 might be in part due to them wa…

ChatGPT-4 is definitely slower than GPT-3.5 (and way slower than 3.5-turbo). What could be the reason for that other than much larger parameter count? I agree that the capabilities seem overhyped. In my subjective experience, 4 seems a little better than 3.5 but not by a huge amount. We just have OpenAI’s cherry-picked word that it‘s this incredible advance.

> What could be the reason for that other than much larger parameter count?

Longer inference time... I should have written it down now that people are asking about it, but a few weeks ago I was seeing people discuss the GPT-4 "paper" in what little information was released and that throwing more inference compute at the problem gives better responses.

>, 4 seems a little better than 3.5 but not by a huge amount.

Can you define that in a tangible way? I don't think most of us can since we have so little access to the product.

Re: OpenAI’s policies hinder reproducible research on language models

#320

Post by alden5 »

cfn wrote: post #124
alden5 wrote: post #84

Earlier quoted context omitted.

Just today I got Stanford's Alpaca-7b model running locally on my m1 mac, it’s just facebook’s Lamma-7b model which has been trained to complete tasks. It's getting close to the versatility of chatgpt where I could actually use it for everyday tasks. I don't think open source is that far away, especially considering how quickly Alpaca came out and how much better it is vs Lamma, which frequently would hallucinate and…

Can you link to instructions specific to the Mac. I can only find instructions for Alpaca with GPUs and PCs.

the lamma.cpp project on github has instructions for alpaca. id recommend not using the alpaca download given and finding the updated torrent in issue #324 as the download didnt work for me
Post reply on HN