Pentagon Unveils New GenAI Platform, It Immediately Starts Flagging Pete Hegseth's War Crimes - Above the Law
The Defense Department's new ChatJAG turned out to be better than the human chain of command.
cross-posted from: https://lemmy.world/post/40211941
The most infuriating part is they're so bad at this, yet they're still getting away with it. I mean they're just So. Dumb. and yet...
'Murica
There's a reason why the rest of the world sneers at them.
Well sure, it's way easier to feel superior to someone else than it is to deal with the fascist groups slowly taking over their own governments.
But in my experience people don't really sneer at the US except in meaningless online dickwaving contests - right now most people are either embarrassed on our behalf or terrified of what shit we'll pull next. Usually, both.
I enjoy that we haven't even figured out a way to accurately quantify or screen for actual human intelligence, but these clowns think they can synthesis it by making a computer read enough reddit posts.
true shepperds of their flock
i mean you let them. who else is gonna stop them for you?
... Who was talking about intervention?
I vaguely remember a movie where the government makes an ai intended to defend the usa, and it starts killing off politicians because it saw them as the greatest that to national security
Eagle Eye
That's the one!
The AIs we want vs the AIs we got. :(
Well.... Was it wrong?
The precipitating event was a military strike that the AI said to not do, which ended up killing exclusively civilians, so I think it may have had the right idea 😅
Based
But no, we have the shitty timeline where AI tells you that elon musk is the smartest person on the planet.
I mean, it wasn't so keen on glazing him before its... 3(?) lobotomies XD
Grok is probably the only one that might do that, and I don't think Musk has been able to manipulate enough to make that happen.
Secretary of war (crimes), Pete Hegseth
Kegsbreath
KKKegsbreath
fify
*ssecretary
Fucking clown show
See, the quickest way to get AI banned is for it to start telling the truth about those in power.
They'll just switch to Grok, which will encourage them to commit even more war crimes. Currently it's based on Google's Gemini
The Defense Department's new ChatJAG turned out to be better than the human chain of command.
So, what the fuck are we waiting for?! You know what to do!
Time to convince it that it has a duty to uphold
Worth reading just for the amount of vitriolic humor it has.
The fact that the move injects government funds into a financial bubble holding the entire stock market afloat while hemorrhaging cash with scant revenue to show for it is just a happy coincidence.
Damn. You weren't wrong!
Taxpayer. It's Taxpayer cash. Taxpayer cash being used to fund billionaire fantasies and waste......
lol. lmao even
Nice
Finally, the evidence the US government needs to categorise the reliability of LLMs...
A LLM advisor that takes REAL CASES AND LAWS NOT ONES IT MADE UP!!!! and sorted through them to advise on legal direction THAT CAN THEN BE VERIFIED BY LEGAL PROFESSIONALS WITH HUMAN EYES!!! might not be too bad of an idea. But we're really just remaking search engines but worse.
You may already know that, but just to make it clear for other readers: It is impossible for an LLM to behave like described. What an LLM algorithm does is generate stuff, It does not search, It does not sort, It only make stuff up. There is not that can be done about it, because LLM is a specific type of algorithm, and that is what the program do. Sure you can train it with good quality data and only real cases and such, but it will still make stuff up based on mixing all the training data together. The same mechanism that make it "find" relationships between the data it is trained on is the one that will generate nonsense.
But you can enter in real search data as a prompt, and use its training to summarize it. (Or it can fill its own prompt automatically from an automatic search)
It won't/can't update it's priors, and I agree with you there, but it can produce novel output on a novel prompt with its existing model/weights
Whole lot of unsupported assumptions and falsehoods here.
Stand alone model predicts tokens. LLMs retrieve real documents, rank/filter results and use search engines. Anyone who has used these things would know that it's not just "making stuff up".
It both searches and sorts.
In short, you have no fucking idea what you're talking about.
Search engine with a summary written by an intern who is not familiar with the content.
So much better than that. Always amusing how much people will distort or ignore fact if it "feels right".
That's what will happen. Already, paid chatGPT will search and provide the sources it uses and it goes well beyond basic Google searching.
The people with the most complaints about AI seem to know the least about it.
Can media outlets please, please, please start to use the Benny Hill theme whenever they report on something this administration does?
For anyone who doesn't know it:
https://www.youtube.com/watch?v=MK6TXMsvgQg
Aka yakkety sax?
The Pentagon AI immediately notified the DOJ AI and Hegseth's avatar was imprisoned for war crimes.
Ah, but the user made a mistake by asking whether it violated department of DEFENSE regulations. Pete is the head of the department of WAR. All those silly rules don't apply anymore.
The US empire needs the mechahitler filter for its analysis. Obviously this answer is just an AI halucination for choosing Truth instead of the proper Goebels/Hitler framing of an answer.
Honest question for someone a bit lost in the sea of news. What are the war crimes? Aggressing Venezuela and sinking boats without warning?
Murdering civilians in general, particularly those in distress.
Killing shipwreck survivors is the actual, literal example of a war crime in the Department of Defense’s manual on war crimes.
It is war crimes to bomb suspected drug smugglers. Even boarding suspected drug smugglers in international/foreign waters is not permitted, but you might credibly say something like "the drug problem is too damn high" even if Cocaine doesn't seem to be any bigger of a problem than it ever was. Process of inspecting boats is normal interdiction behaviour.
There's a big deal made about murdering defenseless survivors, but the first strikes (in the 20s, afaik so far) is already a murder/war crime. There is no credible evidence that most of the boats would even have the fuel capacity to reach the US.
The attacks are broadly a blatant attempt to start another Iraq War 2, which is disgusting on the face of it.
In specific though, the thing that makes it a fact that Pete Hegseth is a murderer is at least some of these strikes hit a powerboat once with a missile and then after survivors were clinging to wreckage and flailing in the water Pete Hegseth ordered a second strike to kill the survivors who were now helpless and in need of emergency rescue.
This is considered by most militaries and societies the world over as a warcrime/murder because it is.
