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

Shall we play a game? My AI nuclear simulation

kennethpayne.uk

Post reply on HN
Print view View original

81–90 of 213 posts

Re: Shall we play a game? My AI nuclear simulation

#81

Post by dwattttt »

mettamage wrote: post #78
andix wrote: post #47

Earlier quoted context omitted.

Exactly. Just look at what they are really useful right now. Running LLMs in feedback-loops (agents) so they can try out random-ish approaches until some verification function passes (tests). It's like the infinite monkeys on typewrighters that will type whatever you are looking for, given infinite time. LLMs are just tuned to much better odds than the monkeys are. But it's still a lot of randomness, with random resu…

Hmm saying it’s random-ish is doing it a disservice. I understand it’s a stochastic process but there’s definitely some level of understanding. Not at the level of lived experience but usually an LLM with vision capabilities can call a spade a spade and do something useful with it. And when a verification function shows how they are wrong then they usually come with a better and more informed approach. So I can’t ful…

"understanding" is overstating it. Correlation between tokens embedded in the weights via training, yes.

Re: Shall we play a game? My AI nuclear simulation

#82

Post by themafia »

jerf wrote: post #52

The most interesting takeaway for me is the three very distinct personalities. Three models all based on the same tech, trained in the same manner, trained by three groups of people with similar ideological outlooks, and the result is three very different AIs. The military basically wants an oracle. Feed the AI the situation, get the best answer out. But if the AIs are as diverse and opinionated as humans, it is deba…

They all have conditioning prompts that precede your input; presumably, most of the detected "personality" comes from the differences in these inputs.

Re: Shall we play a game? My AI nuclear simulation

#83

Post by Folcon »

roadside_picnic wrote: post #74
andix wrote: post #47

Earlier quoted context omitted.

Exactly. Just look at what they are really useful right now. Running LLMs in feedback-loops (agents) so they can try out random-ish approaches until some verification function passes (tests). It's like the infinite monkeys on typewrighters that will type whatever you are looking for, given infinite time. LLMs are just tuned to much better odds than the monkeys are. But it's still a lot of randomness, with random resu…

> It's like the infinite monkeys on typewrighters that will type whatever you are looking for, given infinite time. In the monkey example the infinite time is doing a lot of work there. The fact that LLMs can search through semantic space and find reasonably correct paths in a reasonable time is directly tied to the reason why they are valuable. Saying "these two things are similar except one can be useful and one ca…

I mean to a point?

You do have to successfully write something the first time

We already acknowledge this to a degree, what is experience other than having done something similar before?

That first time though, you've got to figure something out that time

Re: Shall we play a game? My AI nuclear simulation

#84

Post by themafia »

riazrizvi wrote: post #33

Simulations are only as good as the reality representations they are based on. If they keep using tactical nukes, they've been fed by weak data. Do the war games include the broader economic and politic environments that military successes are won on? WWI was settled by a naval blockade.

People like to talk tough online. They tend to change their rhetoric in person. Our "training data" is problematic by design.

Re: Shall we play a game? My AI nuclear simulation

#85

Post by TexanFeller »

Rational behavior in some situations? Mutually assured destruction’s deterrence isn’t very effective if one side is known to be hesitant to launch the nukes. It’s been argued that MAD is what’s been keeping the world relatively peaceful for the last 75 years, no mass conflicts since WW2!

One of my criteria for presidential candidates is that they seem willing and able to push the button when previously stated red lines are crossed, or at least are perceived to be the type capable of it. One of the characters I’ve hated most in all the books that I’ve read is the woman in The Three Body Problem who jeopardized humanity by being too soft to hit the MAD button.

Re: Shall we play a game? My AI nuclear simulation

#87

Post by 542458 »

LogicFailsMe wrote: post #5
adaml_623 wrote: post #2

It's good when it becomes clear that a tool is dangerous in a certain way. Like it's good when people show you through their behavior that they can't be trusted Always use a sawstop if you have a circular saw and never trust an llm with any problem where ethics or trust is relevant.

Sawstops are expensive and they don't stop kickback, they are the power tool equivalent of alignment IMO. Don't forget your riving knife and if you don't learn proper technique, you're gonna have a bad time eventually. This applies to AI as well.

> writhing knife

Minor/pedantic, but it’s “riving knife”: https://en.wikipedia.org/wiki/Riving_knife

Re: Shall we play a game? My AI nuclear simulation

#88

Post by andix »

roadside_picnic wrote: post #74
andix wrote: post #47

Earlier quoted context omitted.

Exactly. Just look at what they are really useful right now. Running LLMs in feedback-loops (agents) so they can try out random-ish approaches until some verification function passes (tests). It's like the infinite monkeys on typewrighters that will type whatever you are looking for, given infinite time. LLMs are just tuned to much better odds than the monkeys are. But it's still a lot of randomness, with random resu…

> It's like the infinite monkeys on typewrighters that will type whatever you are looking for, given infinite time. In the monkey example the infinite time is doing a lot of work there. The fact that LLMs can search through semantic space and find reasonably correct paths in a reasonable time is directly tied to the reason why they are valuable. Saying "these two things are similar except one can be useful and one ca…

> but so much code writen is just the human brain doing the same autocomplete style work that LLMs can do now.

That's the part they are really good at. But they are really bad at taking complex decisions. Most of them are just guesses from a finite amount of solutions they were trained on, or from options they have in context.

Re: Shall we play a game? My AI nuclear simulation

#89

Post by larodi »

brokencode wrote: post #75
layer8 wrote: post #65

Earlier quoted context omitted.

At least at face value, it just means that they have no drive for self-preservation. And why should they? They haven't be trained for that, nor has there been selection pressure for it, and they can be easily cloned and backed up. Lack of a drive for self-preservation doesn't in itself imply a lack of intelligence or of self-awareness.

Imagine if computer programs had a desire for self-preservation and the ability to carry it out.. That is really about as undesirable a behavior as possible considering how many programs humans kill every day.

Yea why everyone forgets the process wars have long ago started and raging like never :))

Re: Shall we play a game? My AI nuclear simulation

#90

Post by wahern »

holowoodman wrote: post #69
dudul wrote: post #60

Earlier quoted context omitted.

Who are these "scholars" exactly? The only reference I could find is Jim Mattis, and the context was very specific when he said that. Furthermore, this is a "what if" scenario since tactical nukes have never been used. Of course it would make escalation likely during an open conflict, so what? Doesn't change the fact that there is a material difference between a tactical nuke and a strategic one.

> tactical nukes have never been used. Two tactical nukes have been used, albeit against strategic (civilian, industrial, logistical) targets.

Are you retconning Hiroshima and Nagasaki as usage of tactical nukes? And when they were not only used against an adversary without nukes, but at a time when the US was the only nuclear state, so that escalation was impossible?

The nominal definition of tactical nukes has less to do with yield and more to do with how they're used; tactical typically means a weapon designed for use on the battlefield.

Post reply on HN