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

Meta disbanded its Responsible AI team

theverge.com

Post reply on HN
Print view View original

401–409 of 409 posts

Re: Meta disbanded its Responsible AI team

#401

Post by smegger001 »

RcouF1uZ4gsC wrote: post #338
0xDEAFBEAD wrote: post #311

Earlier quoted context omitted.

Let's be honest here, a lot of HN users have a conflict of interest on this topic. AI entrepreneurs trying to get rich from LLMs benefit from LLMs being open source. But the downside risk from e.g. bioweapons is spread across all of society. It's the same sort of asymmetrical cost/benefit that tobacco companies and greenhouse gas emitters face. Of course if you went to an online forum for oil companies, they'd be hop…

I heard that reading is very dangerous. Reading allows people to for example learn how to build bio weapons. In addition, reading can spread ideas that are dangerous. Many people have died because they were influenced by what they read. It would be much safer if reading were strictly controlled. The companies would have “reading as a service” where regular people could bring their books to have them read. The reader…

Stop that your giving Florida ideas.

Re: Meta disbanded its Responsible AI team

#402

Post by _djo_ »

smegger001 wrote: post #400
_djo_ wrote: post #361

Earlier quoted context omitted.

Nuclear weapons have absolutely not been democratised, and still stand as a technology that has largely been restricted and not proliferated. Only 9 countries in the 190 or so out there currently nuclear weapons, and 3 countries (South Africa, Ukraine, and Kazakhstan) decided that maintaining stockpiles was more trouble than it was worth. Huge effort has been made to keep nuclear weapons out of the hands of non-state…

To be fair many countries dont want or need them or rely on other allied countries that do have them so they dont have to feild the expense. Other have them but dont officially admit to it. Some cant afford them. But a basic gun type nuclear bomb design like little boy is not technically difficult and could be made by a highschool ap physics class given access to the fissionable material.

That countries don’t want to handle the expense is one of the signs that they haven’t been commoditised.

The number I provided also includes both declared and undeclared nuclear powers.

It’s not just about the ability to make a bomb, it’s about being able to maintain a credible arsenal along with delivery platforms.

Re: Meta disbanded its Responsible AI team

#403

Post by Retric »

Retric wrote: post #299
peyton wrote: post #231

Earlier quoted context omitted.

I just don’t see what value this brings to the table. Sounds like Meta might not either. We’ll just have to leave it at that.

Meta doesn’t think it’s worth spending money on, that doesn’t mean they don’t see value in it.

it’s worth spending [more] money on.

ie: It only suggests they don’t find the need to endless consider such things. Which seems fair enough as AI isn’t actual progressing that quickly.

Re: Meta disbanded its Responsible AI team

#404

Post by kazinator »

otabdeveloper4 wrote: post #54
kazinator wrote: post #19

Earlier quoted context omitted.

GNU/Linux is open source. Is it being used responsibly? What is the "it" that no single entity has control over? You have absolutely no control of what your next door neighbor is doing with open source. Hey, if we want alcohol to be made responsibly, everyone should have their own still, made from freely redistributed blueprints. That way no single entity has control.

Alcohol is probably the most open-source food product of all time.

And, so, that doesn't help, does it.

It is very common for it not to be used responsibly. Alcohol causes health problems, and is implicated in accidents and violence.

Many countries have laws restricting either the sale of liquor in various ways, or its consumption in public places, or both, not to mention production and distribution.

Re: Meta disbanded its Responsible AI team

#405

Post by ben_w »

tticvs wrote: post #396
ben_w wrote: post #372

Earlier quoted context omitted.

That you’re phrasing it as “training an LLM to keep secrets” suggests a misapprehension: A downloadable LLM that knows secrets fundamentally cannot be expected to keep them. The fixable condition is much weaker: capabilities, if an LLM has the capability to do dangerous things. The problem for secrets in general (which is a separate issue to LLMs, I interpreted you asking me about the EU Chat debate as an attempted g…

But LLMs only have the capability to statistically associate strings of words. That's all they are. There is no other capability possible there. And you admit that they cannot be expected to keep secrets. So what is the point of trying to have a "security" team hammer secret keeping into them? It doesn't make sense. I bring up chat control since I've noticed most "AI Safety" advocates are also vehemently opposed to g…

> But LLMs only have the capability to statistically associate strings of words. That's all they are. There is no other capability possible there.

The first sentence is as reductive, and by extension the third as false, as saying that a computer can only do logical comparisons on 1s and 0s.

> So what is the point of trying to have a "security" team hammer secret keeping into them? It doesn't make sense.

Keep secret != Remove capability

If you take out all the knowledge of chemistry, it can't help you design chemicals.

If you let it keep the knowledge of chemistry but train it not to reveal it, the information can still be found and extracted by analysing the weights, finding the bit that functions as a "keep secret" switch, and turning it off.

This is a thing I know about because… AI safety researchers told me about it.

> Which is fundamentally incoherent.

𐀀𐀩𐀏𐀭𐀅𐀨

Re: Meta disbanded its Responsible AI team

#406

Post by baby »

hibernator149 wrote: post #63
baby wrote: post #58

I really really hate what we did to LLMs. We throttled it so much that it's not as useful as it used to be. I think everybody understands that the LLMs lie some % of the time, it's just dumb to censor them. Good move on Meta.

What useful thing could the AI do that it can't do any longer?

basically you can't ask political, medical, or really anything that's deemed "controversial" or "potentially dangerous" by their prompt

Re: Meta disbanded its Responsible AI team

#407

Post by tticvs »

ben_w wrote: post #405
tticvs wrote: post #396

Earlier quoted context omitted.

But LLMs only have the capability to statistically associate strings of words. That's all they are. There is no other capability possible there. And you admit that they cannot be expected to keep secrets. So what is the point of trying to have a "security" team hammer secret keeping into them? It doesn't make sense. I bring up chat control since I've noticed most "AI Safety" advocates are also vehemently opposed to g…

> But LLMs only have the capability to statistically associate strings of words. That's all they are. There is no other capability possible there. The first sentence is as reductive, and by extension the third as false, as saying that a computer can only do logical comparisons on 1s and 0s. > So what is the point of trying to have a "security" team hammer secret keeping into them? It doesn't make sense. Keep secret !…

> The first sentence is as reductive, and by extension the third as false, as saying that a computer can only do logical comparisons on 1s and 0s.

Please, provide any example to prove me wrong.

Re: Meta disbanded its Responsible AI team

#408

Post by nradov »

0xDEAFBEAD wrote: post #393
nradov wrote: post #392

Earlier quoted context omitted.

No, that's not how this work. You made the claim, so the burden of proof is on you. You're just making speculative statements based on zero scientific evidence and a childish misunderstanding of the basic technology.

Did you read this post? https://www.lesswrong.com/posts/uMQ3cqWDPHhjtiesc/agi-ruin-a... It might require context you don't have. Perhaps try one of these intros: https://www.lesswrong.com/posts/T98kdFL5bxBWSiE3N/best-intro...

You've got to be kidding. Those are garbage articles containing unsupported, unscientific articles written by worthless grifters and alarmists.

The entire LessWrong site is mostly just random people making shit up while trying to seem smart. There may be some useful scraps there but overall it's not a site for serious people. You can do better.

Re: Meta disbanded its Responsible AI team

#409

Post by int_19h »

bakuninsbart wrote: post #187
vasco wrote: post #160

Earlier quoted context omitted.

Let's say someone figures out alignment. We develop models that when plugged into the original ones either in the training as extra stages or as a filter that runs on top. What prevents anyone from just building the same architecture and leaving any alignment parts out Practically invalidating whatever time was spent on it.

Hopefully the law.

And how would you enforce it?
Post reply on HN