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

The Future of Everything Is Lies, I Guess: Safety

aphyr.com

Post reply on HN
Print view View original

151–160 of 195 posts

Re: The Future of Everything Is Lies, I Guess: Safety

#151

Post by robot-wrangler »

zozbot234 wrote: post #128
jasonwatkinspdx wrote: post #127

Earlier quoted context omitted.

You understanding is mistaken. Graeber's "everyday communism" is not a market, and his whole larger point is that contorting everything to the lens of markets is simply ahistorical and unempirical. I'd strongly suggest reading his books. They profoundly changed my understanding of how human institutions and society form.

Unless it's some sort of complete post-scarcity, it has to be understandable in market terms. What happens if people try to free-ride on the whole "communist" system? If they get excluded from its benefits, that's equivalent to enforcing some bundle of property rights.

> it has to be understandable in market terms

I like economics and math too, but the whole discussion of markets is a terrible starting place for deriving results in ethics/psychology. If you insist though, notice that unions will happen unless some other organization is working to prevent them. What do you suppose this means? People are aligned with each other exactly because they've noticed their coworkers are not corporations or governments.

Although the two are entangled, politics is a more relevant framing than economics here. If people weren't broadly aligned on basic stuff, then autocrats, theocrats, kleptocrats and so on would simply not be interested in dismantling democracies. They make that effort because they must.

Re: The Future of Everything Is Lies, I Guess: Safety

#152

Post by apsurd »

seizethecheese wrote: post #124
jagged-chisel wrote: post #46

"Alignment" In what world would I ever expect a commercial (or governmental) entity to have precise alignment with me personally, or even with my own business? I argue those relationships are necessarily adversarial, and trusting anyone else to align their "AI" tool to my goals, needs, and/or desires is a recipe for having my livelihood completely reassigned into someone else's wallet.

Why would relationships with a commercial entity be "necessarily adversarial"? A commercial relationship depends on the product providing more utility than the cost (for the consumer) and providing more revenue than cost (for the commercial entity). This means that while some components of the relationship may be adversarial in some areas, it cannot really be entirely adversarial.

I think we're living in times where the one place that this doesn't hold is now somehow all legal: addiction.

Re: The Future of Everything Is Lies, I Guess: Safety

#153

Post by BloondAndDoom »

I don’t even see the pint of alignment or anything about security in LLMs. I feel like this is how “some people” reacted to the internet when I was young (lots of censorship), how hackers don’t let it happen, then how we are back to that world in the hand of corporations and governments who “think of the children). LLMs are out of the bottle and not going back there, only option is building for the new world on the defender side, everything else is politics.

LLMs can hack, but also nmap made hacking easier do we make nmap illegal? We already have drones who kills people, now there is less human involvement, results are same. LLM can also make defending easier (at least for cyber security) but I guess real world security is not that different. Now evil things can be done faster, easier and at more scale. Also good things have the properties.

It’s another tool in the toolbox, the idea that some entity will able to censor or align it as naive as thinking internet can be controlled. Some will do and manage anyway, but it’s not any different china’s firewall.

Alignment is sold to us by companies like OpenAI and Anthropic , not because they care, because that gives them power and more control. When was the last time a big corporation actually cared about soft topics like this? Yes, never.

Re: The Future of Everything Is Lies, I Guess: Safety

#154

Post by JumpCrisscross »

uoaei wrote: post #150
JumpCrisscross wrote: post #147

Earlier quoted context omitted.

Fair enough. We are a social species. But those alignments occur in small groups. You don’t need effort by “corporations and governments” for nations of millions of people to schism. If anything, those large institutions drive broad-based alignment.

Methinks you've been sitting in your armchair too long. Broad-based alignment doesn't come from nothing, but it is surprisingly easy to achieve when a population recognizes a shared stake. A synthesis between selfishness and altruism emerges when you consider who you can call a "neighbor".

> it is surprisingly easy to achieve when a population recognizes a shared stake

Sure. But it takes work for anything larger than a small, close-knit community. I’m pushing back on the notion that this comes naturally and is a default state. It’s not, at least not relative to people naturally forming in and out groups.

The armchair commenters are probably folks who have never organized a group of people before outside a commercial context.

Re: The Future of Everything Is Lies, I Guess: Safety

#155

Post by zozbot234 »

robot-wrangler wrote: post #151
zozbot234 wrote: post #128

Earlier quoted context omitted.

Unless it's some sort of complete post-scarcity, it has to be understandable in market terms. What happens if people try to free-ride on the whole "communist" system? If they get excluded from its benefits, that's equivalent to enforcing some bundle of property rights.

> it has to be understandable in market terms I like economics and math too, but the whole discussion of markets is a terrible starting place for deriving results in ethics/psychology. If you insist though, notice that unions will happen unless some other organization is working to prevent them. What do you suppose this means? People are aligned with each other exactly because they've noticed their coworkers are not…

> the whole discussion of markets is a terrible starting place for deriving results in ethics/psychology.

Historically, we did essentially the opposite. We figured out many aspects of human ethics and psychology first, and deduced from them how and why markets work as they do.

> ... If people weren't broadly aligned on basic stuff, then autocrats, theocrats, kleptocrats and so on would simply not be interested in dismantling democracies. They make that effort because they must.

This implies that people are only weakly aligned in the first place, otherwise no such attempt at dismantling could ever succeed. That's not a very interesting claim; it does not refute the usefulness of some external mechanism to more directly foster aligned action. Markets do this with a maximum of decentralized power and a minimum of institutional mechanism.

Re: The Future of Everything Is Lies, I Guess: Safety

#156

Post by jasonwatkinspdx »

zozbot234 wrote: post #155
robot-wrangler wrote: post #151

Earlier quoted context omitted.

> it has to be understandable in market terms I like economics and math too, but the whole discussion of markets is a terrible starting place for deriving results in ethics/psychology. If you insist though, notice that unions will happen unless some other organization is working to prevent them. What do you suppose this means? People are aligned with each other exactly because they've noticed their coworkers are not…

> the whole discussion of markets is a terrible starting place for deriving results in ethics/psychology. Historically, we did essentially the opposite. We figured out many aspects of human ethics and psychology first, and deduced from them how and why markets work as they do. > ... If people weren't broadly aligned on basic stuff, then autocrats, theocrats, kleptocrats and so on would simply not be interested in dis…

> Historically

This is not the history, it is a mythology in opposition to the empirical evidence.

Which is why you should read Graeber.

Re: The Future of Everything Is Lies, I Guess: Safety

#157

Post by cwmoore »

JumpCrisscross wrote: post #148
dns_snek wrote: post #117

Earlier quoted context omitted.

Critical bit: > i.e. without brain-washing and deliberately working to create out-groups

And if my grandmother had wheels she’d be a bicycle. The process of creating an in group naturally creates out groups. The “brainwashing” OP describes is just as natural as social alignment through an innate drive for conformity.

Conformity I think follows the innate drive to coerce the nonconformant into compliance

Re: The Future of Everything Is Lies, I Guess: Safety

#158

Post by JumpCrisscross »

cwmoore wrote: post #157
JumpCrisscross wrote: post #148

Earlier quoted context omitted.

And if my grandmother had wheels she’d be a bicycle. The process of creating an in group naturally creates out groups. The “brainwashing” OP describes is just as natural as social alignment through an innate drive for conformity.

Conformity I think follows the innate drive to coerce the nonconformant into compliance

Sure. Push and pull. The point is that needs effort to work at larger scales. We don’t “naturally” organize into nations of three hundred million or a billion. To the extent we do, we also “naturally” go to war.

Re: The Future of Everything Is Lies, I Guess: Safety

#159

Post by fmbb »

Imnimo wrote: post #135
fmbb wrote: post #63

Earlier quoted context omitted.

Well, through natural selection in nature. Large language models are not evolving in nature under natural selection. They are evolving under unnatural selection and not optimizing for human survival. They are also not human. Tigers, hippos and SARS-CoV-2 also developed ”through evolution”. That does not make them safe to work around.

>Tigers, hippos and SARS-CoV-2 also developed ”through evolution”. That does not make them safe to work around. Right, but the article seems to argue that there is some important distinction between natural brains and trained LLMs with respect to "niceness": >OpenAI has enormous teams of people who spend time talking to LLMs, evaluating what they say, and adjusting weights to make them nice. They also build secondary…

We already have humans, we were lucky and evolved into what we are. It does not matter that nature did not guarantee this, we are here now.

Large language models are not under evolutionary pressure and not evolving like we or other animals did.

Of course there is nothing technical in the way preventing humans from creating a ”nice” computer program. Hello world is a testament to that and it’s everywhere, implemented in all the world’s programming languages.

> If the author believes that evolution was able to produce something robustly "nice", there's good reason to believe the same can be achieved by gradient descent.

I don’t see how one means there is any reason, good or not, to believe it is likely to be achieved by gradient descent. But note that the quote you copied says it is likely some entity will train misaligned LLMs, not that it is impossible one aligned model can be produced. It is trivial to show that nice and safe computer programs can be constructed.

The real question is if the optimization game that is capitalism is likely to yield anything like the human kind we just lucked out to get from nature.

Re: The Future of Everything Is Lies, I Guess: Safety

#160

Post by pineaux »

JumpCrisscross wrote: post #148
dns_snek wrote: post #117

Earlier quoted context omitted.

Critical bit: > i.e. without brain-washing and deliberately working to create out-groups

And if my grandmother had wheels she’d be a bicycle. The process of creating an in group naturally creates out groups. The “brainwashing” OP describes is just as natural as social alignment through an innate drive for conformity.

There is a pretty interesting study of a large group of chimps. I dont remember where exactly but they have been civil warring the last 15 years or so. Point is, it seems that there is some kind of innate group formation process.
Post reply on HN