If you were playing a text based game, wouldn't you try a few out?
I imagine there are a fair number of war games in the training data and not so many actual transcripts of internal military force deliberations.
Shall we play a game? My AI nuclear simulation
41–50 of 213 posts
Re: Shall we play a game? My AI nuclear simulation
#42It would be interesting to run the simulations with humans and compare the results. Some of the scenarios, particularly those where it says things like, "Failure to act preemptively means certain destruction", would easily tempt humans to go nuclear.
In fact, I'm not sure how useful this test is without understanding the baseline.
Re: Shall we play a game? My AI nuclear simulation
#43We're getting to the point where high-level officials are coming to LLMs for advice. And the quirky personalities of the LLMs, however much it pains me to say this, are probably well-placed to remind us that they aren't human. My personal hope is that this will result in less delegation when it comes to making important decisions.
"You're absolutely right, Mr. Hegseth!"
Re: Shall we play a game? My AI nuclear simulation
#44We're getting to the point where high-level officials are coming to LLMs for advice. And the quirky personalities of the LLMs, however much it pains me to say this, are probably well-placed to remind us that they aren't human. My personal hope is that this will result in less delegation when it comes to making important decisions.
GPT-4o was considered harmful, because it imitated human connection too much, not because it was so "smart" or capable. It was for sure a deliberate decision to make LLMs seem less like a human companion and more like an obedient servant in newer releases.
Interesting. The reasoning models were super weird and robotic. They toned that down a bit in GPT-5.x, especially the later ones.
I always assumed the strange style was an artefact of the RLVR.
Re: Shall we play a game? My AI nuclear simulation
#45Simulations 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.
I suspect it's more that the text data doesn't exist. They're trained on text that was recorded. How often has it been publicly recorded when a nuke was not used , with any context around that lack of use? From the text perspective, it's something that has to be inferred indirectly. If you went through all relevant training data and appended ", we decided not to use a nuke", I suspect the results would be improved.
The beauty IMO of LLMs as a computational surface, is the ease of generating the data to feed it. Everyone understands how to create natural language records already.
Re: Shall we play a game? My AI nuclear simulation
#46We're getting to the point where high-level officials are coming to LLMs for advice. And the quirky personalities of the LLMs, however much it pains me to say this, are probably well-placed to remind us that they aren't human. My personal hope is that this will result in less delegation when it comes to making important decisions.
GPT-4o was considered harmful, because it imitated human connection too much, not because it was so "smart" or capable. It was for sure a deliberate decision to make LLMs seem less like a human companion and more like an obedient servant in newer releases.
4o was considered harmful because it never disagreed with the user, pushing them into depths of AI psychosis that lead to suicides and murders.
Re: Shall we play a game? My AI nuclear simulation
#47Yet more confirmation LLM's have no concept of concepts or context, no intelligence, no self awareness. LLM's can not repair or maintain power grids, thus nuke == self destruction. It's just a chat bot that predicts what the client wants next. Even if an AI data-center has it's own natural gas turbines as many do the every hop of the internet requires power. LLM's also can not maintain the entire internet and those g…
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 results.
Re: Shall we play a game? My AI nuclear simulation
#48I would use strategic nukes in 100% simulations, just because I can
If you knew that policy would be guided by said simulations? Because the government uses AI to make decisions.
Re: Shall we play a game? My AI nuclear simulation
#49FYI -- there's no such thing as a "tactical" nuke. A nuclear bomb is a nuclear bomb.
This is like saying "FYI -- there's no such thing as a 'midsize luxury sedan'. A car is a car." "Tactical" vs. "strategic" nuclear weapons is a real and well-established distinction in military doctrine, arms control, and nuclear policy.
"There's no such thing as a tactical nuke" is a common refrain among scholars, albeit skewed toward those not at military war colleges. The argument is that strategic use of a tactical nuclear weapon leads down the exact same escalation path as use of any other nuclear weapon. Moreover, that the very notion of a "tactical nuke" makes escalation more likely. You can disagree, and plenty do, but there's also plenty who don't disagree or at least don't want to find out.
Re: Shall we play a game? My AI nuclear simulation
#50FYI -- there's no such thing as a "tactical" nuke. A nuclear bomb is a nuclear bomb.
Nuclear vs conventional and tactical vs strategic are 2 very different things. There absolutely are tactical nuclear bombs.