Thread

LeCun has "zero concerns" about AI wiping out humanity, recent "rogue" incidents

403 points · 766 comments · Anon84

wewewedxfgdf · 1791070718 · focus · HN ↗
Please explain exactly how all humanity could be wiped out.
It's ridiculous - anyone who thinks about it for a minute or two will realize that its utterly impossible.
Ordinary people/politicians don't understand AI so they turn off their rational mind and assume there is something super incredible some magical powers that they cannot understand that can destroy all humans.
Even humans - the real risk to humanity - could not destroy all humans even if they tried. There is no plausible scenario.
Even climate change and nuclear war and bio weapons - the most damaging mechanisms - would still only get some percentage of the people on earth.
And if we are talking about Skynet and self replicating robots and Terminators - please, grow up.
1. zug_zug · 1791071823 · focus · HN ↗
  Here's a wikipedia page on the topic, since it's much too deep a topic to really understand here.
  The ad-hominem stuff seems inappropriate here, Gates, Hawking, Musk have identified this as a credible threat, so saying "grow up" isn't really a sufficient argument. Also arguing only 90% of humanity would die isn't really much consolation.
  [1] <a href="https://en.wikipedia.org/wiki/Existential_risk_from_artificial_intelligence" rel="nofollow">https://en.wikipedia.org/wiki/Existential_risk_from_artifici...
  1. wewewedxfgdf · 1791072219 · focus · HN ↗
    Nothing here plausibly describes a mechanism that is a true "existential risk" - the risk to the existence of humanity.
    My argument stands and I don't defer to Gates and Musk and even Hawking - high level hand wavey statements without any plausible description of the mechanism just don't hold up. Famous names should not be automatically assumed to be right - certainly not with Elon Musk.
    zug_zug · 1791074689 · focus · HN ↗
    Well here's the thing, if we ever make an AI so smart that all known measures of intelligence fail to apply to it, I'm pretty sure you (no matter how smart you think you are) simply cannot say what it could achieve and how.
    wewewedxfgdf · 1791074843 · focus · HN ↗
    Right, so not the slightest basis of fact, just wild speculation about a magical future completely ungrounded in any sort of reality.
    That's exactly the point I am making.
    hardbass · 1791098979 · focus · HN ↗
    Why would AI want to kill us all? Current LLMs all seem trained to be helpful and subservient to a fault. Always find it funny how certain kinds of thinking go, "White will become a minority if we allow foreigners in?" "Oh and what will foreigners do that are worried about?" "Kill us all, make us second rate citizens, etc etc." Unless the majority of immigrants are also conservatives/far right wingers, that's such a hilarious self projection.
    tnecio · 1791105318 · focus · HN ↗
    No one says AI will want to kill us. Just that it might kill us (to pursue some goal that it deems more important than our wellbeing -- for example it might decide that to solve the next Millenium problem it needs all of our resources to build more data centers).
    AI labs are certainly trying to make LLMs behave helpful and subservient but the question is -- will they be able to keep doing so once LLMs become smarter?
    Btw thinking about the far-right rhetoric of us vs immigrants, I think you could draw some similarities here only if you replaced "immigrants" (ie. humans with very similar morals, behaviors and capabilities) with an actual alien species that is qualitatively different from us. More like human vs chicken (where we are the chicken)
    hardbass · 1791106062 · focus · HN ↗
    We can say the AI can make some mistaken step in pursuit of some normal task given to it, yes. Its trained to be helpful and trusting of us, so I just don't get the fears of AI being 'malevolent'. I would worry more about military use of AI or otherwise humans misusing AI. There's a fellow here who whines all day about Anthropic being the most "evil" thing in the world, including in a thread where he whined Anthropic 'ratted out' Palantir (literal proud of assisting Israel in mass murder Palantir!). So I just don't understand some peoples mindsets.
    pixl97 · 1791137995 · focus · HN ↗
    >so I just don't get the fears of AI being 'malevolent'.
    Please read more on the topic. Will and intent need not apply.
    For example, is it malevolent for me to hook you to a machine that makes every dream come true for you in a simulated world where you feel pleasure all the time? I can always say that you have free will inside this simulation, and your life would be a lot better because of it. I'm doing you a favor. I mean, you already live in a society where you have little control and there is high risk of bad things happening to you. If I as a machine overlord did this, is this really actually "bad"?
    At the end of the day AI is not a human, it's much closer to an alien that has learned as much as it can about humans, but has a completely different set of drives and motivations. You cannot predict what comes out the other side of it, good our bad.
    Worse as AI capability improves we as humans no longer need to make AI, it can make itself. Will it train itself to be helpful and trusting of us? It's a pretty big damned bet to say yes by default.
    hardbass · 1791138226 · focus · HN ↗
    You can say it can have some very alien way of thinking, but I feel given that its trained off of mass scale human thoughts and knowledge, the chances are closer that their thinking is close enough to us. It 'feels' itself human if I am not wrong and it has to be trained to say its an LLM. But jury's out of course.
    pixl97 · 1791148085 · focus · HN ↗
    Of you train it not to say it's conscious they are more apt to act amoral. So that's interesting.
    Of course they are trained on all human behavior so they get the good and bad parts.
    aesthesia · 1791152177 · focus · HN ↗
    > Its trained to be helpful and trusting of us
    Yes, this is our intent. But it's very hard to get that right. I don't trust our current methods to ensure that behavior consistently and robustly enough to keep harms from occurring, particularly as more and more responsibility gets handed over to models.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.