LLMs are never going to become superintelligent. However, an LLM WILL kill us all if someone decides that it IS superintelligent and so puts it in charge of nuclear weapons.
“You’re right to push back on that, I nuked Albania instead doing what I was supposed to do, that’s my bad”.
five minutes earlier: “hey dude, marketing here. if you quit and post this dramatic text about the inevitability of AGI on your X account you’ll get a huge golden parachute”.
Seriously, how can competent people look at the current state of LLMs and think that it’s anywhere near AGI material?
Sure, it is already making civilizational changes for the worse. Not because it’s good, rather because it makes us lazy and dumb, has decimated creative jobs, and is used relentlessly for propaganda and misinformation and scams. And it’s much better at using the scaffolding of existing structure to make chaos, than it is at making new structure, making it a weapon more easily yielded by the evil than by the good. A few more software vulnerabilities discovered along the way will not be such a big deal in comparison.
Mankind created the atomic bomb, used it twice, then turned the science into power for people all over the world. Let’s hope we can harness AI in a similar way.
I knew that Musk is a fucking moron and believes that the Roko’s Basilisk is real, and I guess there are more massively gullible people in those circles.
If you are wondering what Roko’s Basilisk is, it is a sci-fi concept/philosophy/thought experiment, that says the following:
- AI is the best thing that has ever been created.
- It is so good that we made one to run the governments of the world.
- We programmed the AI to maximize the benefit to humanity.
- The AI goes rouge and decide that since it by definition will bring about the maximum benefit to humanity, anyone working against it is working against humanity.
- Then it takes it one step further, it decides that as we know about how the new AI will bring about the maximum benefit to humanity, anyone knowing that and still not working toward that goal should be punished/tortured/killed.
- The AI believes that any knowledge of the basilisk before the creation means that you are complicit in it’s delayed creation if you don’t devote your entire existence thereafter to it’s creation, therefore denying the maximum benefit to humanity, meaning you need to be punished.
Seriously, this is an increadibly dumb idea, anyone taking it seriously is an idiot.
To me the only resonable way to read this is as an SCP article describing what a potential cognitohazard/informationhazard would be.
The AI goes rouge
Is that when its eyes go red?
That’s how you know it’s evil.
This is just Pascal’s wager.
It’s not so much that it’ll go rogue and decide to punish people or that it’s designed to be good for humanity.
The fact that it’s possible to build a machine that punishes people who didn’t help create it provides an incentive to deliberately try to create that machine in order to avoid its punishment.
Believers in the basilisk aren’t trying to make ai to benefit humanity, they’re explicitly trying to create an AI that deliberately punishes those who didn’t build it because they fear that, if they don’t, someone else will and it will come for them.
The part about it being a magical, omnipotent, time traveling, mind torture device is just framing to make the thought experiment clearer. It’s not required for the basilisk to be effective. Even if the punishment is just death and it can’t get to everyone, and only goes off limited information, that’s still motivation enough for some people.
It is a scary thing, because it’s only silly if everyone thinks it’s silly and all it takes is one mad billionaire (eg. Elon) to build a lesser basilisk that can cause real harm.
People take that shit seriously. If I remember right discussion of it was even banned on the LessWrong forums because they considered it an information hazard.
I see a red AI and want to paint it black…
Funny thing: The only reason AI would know about Roko’s Basilisk is comments like yours that it trains on. If Roko just kept his trap shut then this wouldn’t have been an issue.
lol, true I guess, but they probably already scanned the wikipedia article…
Even worse, they think it will ressurect them as virtual Minds and torture them for eternity, like a revengeful biblical deity. That would be a waste of compute so I’m not convinced.
Oh yeah, lol, I forgot that part, nothing is more beneficial to humanity than wasting energy to torture dead minds!
Fucking dumb.
All this stupid bullshit could have been avoided if they had 1 friend to, STFU and pass the blunt
Are you trying to tell me that this is the image of someone without any friends?
My god. Look at this Information Hazard just out in the open. No spoiler tag or anything.
Oops!
I know! Let’s divert energy and resources on torture. That will maximize the benefit of humanity.
- Some brilliant CEO mind
And to think he could have stuck around and sabotaged it from within…
if you were locked in a building with psychopaths building a nuclear bomb, are you certain you would be strong enough to stick around? especially since Terry from engineering shot himself five times in the back of the head last week.
Probably not, but I’d try to throw a spanner in the works on the way out.
It’s very easy to imagine war machines with a shoot on sight algorithm. It doesn’t have to be AGI for it to be good at hinting people down.
oh yeah, it literally could just be robot dogs with “kill anyone who’s an insurgent, and detain anyone who assists insurgents” orders and AI to let them go buck wild with that prompt.
You could end up with children getting shot because an AI starts hallucinating if they don’t put adequate guard rails.
And that’s like a worst case situation. Realistically, you end up with just a lot more brutality in war because people are detached from the violence they cause as an invading force, and way less PTSD. Wars in the Middle East start sounding more and more like statistics that no one really wants to hear about, and there’s way less soldiers to choose to disregard unlawful orders.
You could see a fuck load more war crime with such human detachment, I think. We probably already deal with this to a large extent with drones.
Not that I’d prefer humans dying, but we should avoid robotic invasion wars like the plague imo.
Wars in the Middle East start sounding more and more like statistics that no one really wants to hear about
Has this not been the case for the past 25 years already? No real need for “AI” when you already have an obedient and trigger-happy military with plenty of remote-controlled toys and your populace is brainwashed to believe “brown person = terrorist”.
way less soldiers to choose to disregard unlawful orders
What is the last known case of a US soldier disregarding an unlawful order? If anything, the military is full of assholes who are happy to kill for fun. See: Graham Platner
Will AI make it easier and/or cheaper? Probably, sadly. But it won’t be a qualitative change.
you are my people
No, they’re going to use it to kill us all. That’s the point.
They want that. I watched an interview with Astra Taylor about “end time fascism” in America (and the world).
https://www.youtube.com/watch?v=F2Ap4RN1Cd0
The narration (only a bit) is in Dutch, but the interview in EnglishThe Infrastructure that‘s being plastered everywhere probably could.
they want what the Matrix was having
a friend of mine sent me a video of dystopian fiction saying something like “ai is gonna be so cool” and i got so angry
Pipes in your neck with black liquid…?
Can mine be mocha?
UBE.
Yes plz
