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

OpenAI's comment to the NTIA on open model weights

openai.com

Post reply on HN
Print view View original

61–70 of 74 posts

Re: OpenAI's comment to the NTIA on open model weights

#61

Post by benreesman »

simonw wrote: post #36
benreesman wrote: post #15

Earlier quoted context omitted.

The latter, in the context of the justification. It was pulled because while it caught on the order of 25 of pure AI output, it missed the rest. But I’m cool with that number, in a world where so much AI in human writing? That’s a total win on low-effort propaganda. Anyone should be cool with that number. Unfortunately they had to pull it because something like 9% of human authored text got hit with the AI flag. Agai…

A great deal of harm is being caused by "AI detectors" today. There is an endless stream of stories about students who have their work flagged as "AI generated" by a detector when it wasn't - and then get failing grades, from teachers who don't understand that AI detectors are inherently inaccurate. Here's a paper about how unreliable they are from last year: https://arxiv.org/abs/2303.11156 OpenAI shutting down thei…

Nah, while the list of good things they’ve done is getting shorter, they have pushed the pace in a few areas utterly consistent with their ostensible mandate and they deserve credit for it:

- whisper is really useful, it has good applications in strictly socially positive settings (accessibility as one example), and its scope for abuse is very consistent with how they’ve opened up the weights. the whisper people either still are, or until recently were, doing the right thing. - the TTS stuff is a little less cut and dried, but TTS is among the more potentially dangerous capability increases, and I can see an argument for going a little slower there, for a number of reasons. I still think they’re behind on opening that up, but the voice group has a case there, even if it is thin and I personally disagree. - the detection stuff, they were pushing the pace on tools researchers and institutions need. they deserve the same credit for doing that, which I’m giving them, as blame for pulling it, which I’m giving them. that was consistent with their stated mandate.

If you’re going to criticize and institution stridently to the point of calling it a menace, as I am, you are held to a higher standard on being fair, and I acknowledge the good and positive and non-remunerative work, or at lead the headline stuff.

Turning off the good stuff is one more really, really red flag.

Re: OpenAI's comment to the NTIA on open model weights

#62

Post by segmondy »

It's a shame, I think they were shocked with how far everyone caught up to them. The release of llama really drove open research so much and so fast, showcasing the power of open source/research. Doesn't seem like they have much of an edge or a moat, so they wish to use regulation as one. Quite a shame.

Re: OpenAI's comment to the NTIA on open model weights

#63

Post by benreesman »

visarga wrote: post #45
benreesman wrote: post #15

Earlier quoted context omitted.

The latter, in the context of the justification. It was pulled because while it caught on the order of 25 of pure AI output, it missed the rest. But I’m cool with that number, in a world where so much AI in human writing? That’s a total win on low-effort propaganda. Anyone should be cool with that number. Unfortunately they had to pull it because something like 9% of human authored text got hit with the AI flag. Agai…

> Again, some people are starting to write like it. It’s gonna happen. Interesting point you touched here. Let's do the math for OpenAI: 100M users, with 10k tokens/user/month, that means 1 trillion tokens/month are read by people. That has got to influence speech and circulate information faster between all fields.

It’s what makes the precision/recall tradeoff a no-brainer if you’ve worked in spam. I worked in Abuse Detection at FB in 2016: there is a consensus on how to use binary classifiers responsibly in the presence of lame attempts at fraud.

Sophisticated fraud, that’s complicated.

But most fraud is dumb.

Re: OpenAI's comment to the NTIA on open model weights

#64

Post by bamboozled »

vouaobrasil wrote: post #55
SirensOfTitan wrote: post #20

Closed weight LLMs are unethical forms of theft that will privatize profits on a work that includes virtually all of humanity’s digital written work and serve to ultimately heighten wealth inequality as they get more sophisticated and start eliminating jobs. The only path forward is open model weights, Sam Altman is on the wrong side of history here, and I hope he fails to convince regulators.

> The only path forward is open model weights, Sam Altman is on the wrong side of history here, and I hope he fails to convince regulators. The third path would be an uprising against AI that would ultimately lead to an outright ban of it, a dissolution of all AI companies, and a moratorium on research on AI.

Which is impossible at this stage ?

Re: OpenAI's comment to the NTIA on open model weights

#65

Post by bamboozled »

jrflowers wrote: post #18
internetter wrote: post #16

Earlier quoted context omitted.

For the copyrighted content they stole

They are working hard to invent SkyNet and we must fund them in perpetuity so they can protect us from SkyNet

If I have to be murdered by Skynet, I want it to be an open source version.

Re: OpenAI's comment to the NTIA on open model weights

#66

Post by bamboozled »

CuriouslyC wrote: post #34
chefandy wrote: post #30

Earlier quoted context omitted.

The fact is though, every corporate actor in this entire landscape is just playing their hand. Anybody's stance on anything at any given moment doesn't mean they're more or less ethical-- the moment they perceive a strategic benefit to walling everything off which would surpass the PR cost, they will. They've probably already got PR folks workshopping angles for the press release.

This is true to a degree though there are high profile actors such as Yann LeCunn who have ethical boundaries. Yann wants AI to be open source and available to all, and he's straight up said that he won't work for a company that doesn't follow this principle. Zuck might not have a hand to play in terms of AI products, but even if he did he'd have to tread carefully because the guy that sets his whole AI direction and…

Same could be said for Ilya once upon a time ?

Re: OpenAI's comment to the NTIA on open model weights

#67

Post by benreesman »

benreesman wrote: post #57
maeil wrote: post #49

Earlier quoted context omitted.

Unfortunately that kind of generalization being inaccurate is exactly what makes this problem so difficult. Every case needs to be judged independently, but that takes a lot of time and effort that no one person individually has. A corporation is purely a legal structure. There exist people who do actually spend their time "doing good", and they too use such legal structures when that's helpful. It's unfortunate that…

My comment history has been calling Altman a menace for a minimum of a year or two, probably longer. For a long time it got me downvotes and dirty looks. Then it was few votes and the occasional “no shit”, then it was just kind of quiet, and it’s about to be consensus. There was a big PR campaign around him personally, and it worked as intended, and if he hadn’t gone this far this fast, it might have remained effecti…

Edit: The Valley has turned, he’s done.

Re: OpenAI's comment to the NTIA on open model weights

#68

Post by vouaobrasil »

bamboozled wrote: post #64
vouaobrasil wrote: post #55

Earlier quoted context omitted.

> The only path forward is open model weights, Sam Altman is on the wrong side of history here, and I hope he fails to convince regulators. The third path would be an uprising against AI that would ultimately lead to an outright ban of it, a dissolution of all AI companies, and a moratorium on research on AI.

Which is impossible at this stage ?

Why? I am talking about a revolution against AI.

Re: OpenAI's comment to the NTIA on open model weights

#69

Post by int_19h »

Havoc wrote: post #3

> a number of nation-state cyber threat actors who were abusing our GPT-3.5-Turbo and GPT-4 models to assist in cyberoffensive operations. Not sure I buy this. Sure there was that half hearted case they blogged about. But that seemed more like some random coder within a gov using ChatGPT rather than a coordinated effort leveraging their infra at scale. Besides a nation state easily has the capability to spin up a loc…

From what I've heard, Yandex's Alice is already largely on par with GPT-3.5 (censorship aside), and that's just what the public gets access to.

Re: OpenAI's comment to the NTIA on open model weights

#70

Post by int_19h »

lolinder wrote: post #40
sadeshmukh wrote: post #38

I'm a little confused why everybody seems to want to mandate open weights. Maybe a system similar to copyright, but by mandating open weights on a system they developed, it somewhat stifles creativity.

Sorry, who's the "everybody" who wants to mandate open weights? The only discussion I'm familiar with is that some people and governments want to ban open weights. Where are you seeing people argue for mandating them?

It is not something you hear from the people in power - so it has close to zero chance of becoming policy - but you can see it mulled on places like HN.

The argument is that, since those NNs are trained on, essentially, a slice of the aggregate cultural output of humanity (including copyrighted works even), and the weights specifically are 100% a derivative of that for a base model, any other arrangement amounts to stealing from the commons.

Post reply on HN