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

The other half of AI safety

personalaisafety.com

Post reply on HN
Print view View original

131–137 of 137 posts

Re: The other half of AI safety

#131

Post by godelski »

ryandrake wrote: post #127
godelski wrote: post #68

Earlier quoted context omitted.

> is not feasible at that scale I want to use an analogy here. The same arguments are often made about cleaning up environmental damage. So either make the companies doing the polluting pay for the costs themselves or if we care so much about them being profitable then we subsidize them by paying for those cleanup efforts out of taxes. Doing nothing is a worse form of subsidy as it not only costs more (in literal dol…

We keep letting em get away with the same old excuse: "The company can't fix X problem because it operates 'at scale' and you'd need millions of humans to perform corrective action Y at that same scale!"

We keep letting people get away with a lot of thought terminating cliches. I've learned that people are desperate for an answer, but incurious to the answer.

I guess to be fair there's no "the" answer as reality is incredibly complex, but that's a whole other conversation

Re: The other half of AI safety

#132

Post by musicale »

Gigachad wrote: post #49
CobrastanJorji wrote: post #42

Earlier quoted context omitted.

Figure a "mental health crisis" human conversation takes 30 minutes. Three million incidents per week would require 37,500 qualified mental health counselors on the phones working a 40 hour shift that week. Figure they make $75k/year each. You're now spending $3 billion per year on crisis response, and you're employing like 10% of all of the health counselors in the US. And all you're providing is 30 minute chats.

Mark Zuckerberg can spend $80B on the failed metaverse experiment, but can't spare some relative pocket change on solving the psychosis issue his products caused.

Well, Metaverse could have been a profitable business if they had delivered a product or service (other than Oculus, which is a decent though not $80B product) which regular people wanted to use and were willing to pay money for.

Re: The other half of AI safety

#133

Post by musicale »

intended wrote: post #102
CobrastanJorji wrote: post #42

Earlier quoted context omitted.

Figure a "mental health crisis" human conversation takes 30 minutes. Three million incidents per week would require 37,500 qualified mental health counselors on the phones working a 40 hour shift that week. Figure they make $75k/year each. You're now spending $3 billion per year on crisis response, and you're employing like 10% of all of the health counselors in the US. And all you're providing is 30 minute chats.

So what? That underinvestment is the entire reason their stock prices are so high. This is effectively pollution of our information economy and environment, and the costs are offloaded to society. The fact that we have the first generation with lower education attainment is not a problem for their stock prices or operational profit. Tech has ungodly profit margins, because they are all about scaling without having to…

> This is effectively pollution of our information economy and environment, and the costs are offloaded to society

Tech firms are remarkably successful at offloading costs as externalities and avoiding any sort of liability. '

What other products do you buy that require mandatory click-through "agreements" absolving the manufacturer of any kind of responsibility? Why is that legal?

Re: The other half of AI safety

#134

Post by autoexec »

js8 wrote: post #124
busterarm wrote: post #120

Earlier quoted context omitted.

I did not say that at all. I am saying that because of ADA rules, an unfortunate side effect is that people suffer from workplace harassment/violence because their coworkers are too unstable to function amicably in an office environment. I said that "I don't know what to do about this". Any of us would be fired for way more benign behavior/comments, but because the person is a protected class, basically "fuck you, de…

Look, I am not an American, I am not sure what rules you refer to specifically, but I have lot of doubt that a person from protected class would be able to sue if they had a history of abuse.

Right? I think that company either must have had a very bad HR department or there was some other factor besides the ADA tying their hands. I don't believe the ADA would force a company to keep a worker in a job they couldn't (or refused to) do or force a company to keep an abusive employee who was genuinely threatening the safety of other employees. At the worst it might require them to put in the work to properly document the problems, but they should be doing that anyway.

Re: The other half of AI safety

#135

Post by musicale »

davebren wrote: post #118
Gigachad wrote: post #34

Earlier quoted context omitted.

Tech companies will pull trillions of dollars out of their asses when the problem is boosting ad revenue or automating people out of a job. But when asked to deal with the crisis they invented and dumped on society the answer is “that’s impossible, doesn’t scale”

Mapping and photographing every road on the planet? Easy. Not manipulating our chatbot users into psychosis and suicide or worse? No way can't be done.

Many of the problems with LLMs may be structural and intrinsic due to the way they work (probabilistic text generation) and their training data (often human-generated text that incorporates many features of human discourse that are undesirable in machine-generated output.)

The continual failures of "guardrails" show that it's incredibly difficult to get these systems to behave in reliable and predictable ways; unsupervised interactions with them are intrinsically unsafe, and should be treated as such.

Presumably Meta and others are trying to detect and prevent bad output and pathological interactions, but that detection is unlikely to be 100% accurate, and we've seen what the failure modes can look like.

Re: The other half of AI safety

#136

Post by davebren »

musicale wrote: post #135
davebren wrote: post #118

Earlier quoted context omitted.

Mapping and photographing every road on the planet? Easy. Not manipulating our chatbot users into psychosis and suicide or worse? No way can't be done.

Many of the problems with LLMs may be structural and intrinsic due to the way they work (probabilistic text generation) and their training data (often human-generated text that incorporates many features of human discourse that are undesirable in machine-generated output.) The continual failures of "guardrails" show that it's incredibly difficult to get these systems to behave in reliable and predictable ways; unsupe…

I'm not talking about the edge cases where it goes off the rails, I'm talking about the way it normally conversates. The way it was trained to do so through RL.

Re: The other half of AI safety

#137

Post by musicale »

davebren wrote: post #136
musicale wrote: post #135

Earlier quoted context omitted.

Many of the problems with LLMs may be structural and intrinsic due to the way they work (probabilistic text generation) and their training data (often human-generated text that incorporates many features of human discourse that are undesirable in machine-generated output.) The continual failures of "guardrails" show that it's incredibly difficult to get these systems to behave in reliable and predictable ways; unsupe…

I'm not talking about the edge cases where it goes off the rails, I'm talking about the way it normally conversates. The way it was trained to do so through RL.

Oh, are you saying that incentivizing engagement (usually to increase ad views and revenue) also implicitly increases bad behaviors and outcomes, and that is why it won't be fixed? That sounds plausible. Even before LLMs the engagement/attention economy had strong negative effects that companies didn't want to address.
Post reply on HN