As a retired lawyer, I immediately spotted the ginormous loophole with the Pentagon limiting its use of Claude to “all lawful use.” We don’t *have* laws prohibiting the federal government from using AI to conduct mass surveillance of US citizens, or requiring human involvement in a drone strike or other weapons attacks. It’s meaningless. The fact that Hagueseth refuses to agree to Anthropic’s reasonable and minimal requirements is telling.
Ta, Evan for highlighting this important point, which is something that outlets like WaPo have failed to address (quelle surprise).
The part about the legislation not having caught up randomly reminded me of a curious detail I saw in one GURPS sourcebook (specifically one detailing firearms of the "two-fisted pulp adventure" timeframe). Apparently well into the Twenties in a great manny countries the ownership of *machine guns* was actually entirely unregulated because governments hadn't gotten around to updating the firearms legislations...
Granted those things were a tad too bulky, expensive and complex for that to have been much of a practical concern - Big Joe from the wrong side of the tracks wasn't going to hold up the corner store with one - but still. (Also the military calibers they were chambered for *were* often heavily controlled which indirectly addressed the topic anyway.)
My hope is that Anthropic holds the line. I feel like this is an opportunity for them to be "the good guys" and hopefully be rewarded by consumers. That might even inspire some ethics in other AI providers?
Can't wait for the Pentagon to switch from Claude to grok and give musk the ability to remote control murder anyone he wants.
I said it yesterday and I'll say it again: calling it now that drone strikes on blue states are going to be a fact of life by 2028 and they'll have slipped into the territory of "literally unthinkable a few years ago and unconstitutional as fuck but Republicans don't care and the media is in the tank for them" like the ICE invasions of Chicago and Minneapolis or alligator Alcatraz.
Axios headline: "Sam Altman says OpenAI shares Anthropic's red lines in Pentagon fight"
Yeah, no company wants to be associated with mass domestic surveillance and autonomous killer robots. Ok, some do. Actually, lots do. . . just not these two AI companies. For now.
Oh they would totally do it if there wasn't so much public scrutiny ... basically the only thing holding them back is Hagueseth's inability to keep his goddamn mouth shut ... I mean I hope these two wouldn't but I'm too cynical to believe they wouldn't if given a chance.
They need artificial intelligence because there's sure as shit no actual intelligence in Petey's Department of WAR
I've been shitfaced plenty of times. And if there is one facial expression that says "shitfaced," it's shitface's.
As a retired lawyer, I immediately spotted the ginormous loophole with the Pentagon limiting its use of Claude to “all lawful use.” We don’t *have* laws prohibiting the federal government from using AI to conduct mass surveillance of US citizens, or requiring human involvement in a drone strike or other weapons attacks. It’s meaningless. The fact that Hagueseth refuses to agree to Anthropic’s reasonable and minimal requirements is telling.
Ta, Evan for highlighting this important point, which is something that outlets like WaPo have failed to address (quelle surprise).
I know why they change 'Defense' to 'War'.
DOD is too easily confused with 'Drunk On Duty' or 'Dept. Of Derps'.
Amodei :
“… shut down CCP-sponsored cyberattacks that attempted to abuse Claude,…”
Well, somebody had to do that, as Frag Me Daddy Hegseth shut down the DOD cyber security as soon as he took office.
Amodei must be the only AI CEO with morals and ethics. All the others are first class rich Fascists who envision themselves as Rulers of the Universe.
And the average American citizen will never know about Hegseth’s demands for a super power AI to conduct domestic surveillance on Americans.
The part about the legislation not having caught up randomly reminded me of a curious detail I saw in one GURPS sourcebook (specifically one detailing firearms of the "two-fisted pulp adventure" timeframe). Apparently well into the Twenties in a great manny countries the ownership of *machine guns* was actually entirely unregulated because governments hadn't gotten around to updating the firearms legislations...
Granted those things were a tad too bulky, expensive and complex for that to have been much of a practical concern - Big Joe from the wrong side of the tracks wasn't going to hold up the corner store with one - but still. (Also the military calibers they were chambered for *were* often heavily controlled which indirectly addressed the topic anyway.)
We're getting to a new version of what should have been a more alarming quote back from several years ago, now it is essentially:
"I could stand in the middle of 5th Avenue and shoot voters everywhere"
The only implausible bit is that he says he could stand.
My hope is that Anthropic holds the line. I feel like this is an opportunity for them to be "the good guys" and hopefully be rewarded by consumers. That might even inspire some ethics in other AI providers?
Well at least the pain of this particularly miserable news day has been mitigated by you guys reminding me of my favorite Onion article ever. https://theonion.com/nations-dog-owners-demand-to-know-whos-a-good-boy-1819565849/
I really miss him:
https://www.cnn.com/2026/02/27/politics/joe-biden-post-presidency-donald-trump-south-carolina
I will bet money that Anthropic's caution and safeguards were a major factor in them getting the contract in the first place
Can't wait for the Pentagon to switch from Claude to grok and give musk the ability to remote control murder anyone he wants.
I said it yesterday and I'll say it again: calling it now that drone strikes on blue states are going to be a fact of life by 2028 and they'll have slipped into the territory of "literally unthinkable a few years ago and unconstitutional as fuck but Republicans don't care and the media is in the tank for them" like the ICE invasions of Chicago and Minneapolis or alligator Alcatraz.
"Can't wait for the Pentagon to switch from Claude to grok and give musk the ability to remote control murder anyone he wants.
Isn't he already doing that with his self-driving cars?
"Don't you stop me Amodei." - Hegseth
Wonder how they feel about mass foreign surveillance.
What dark time is coming?
https://www.youtube.com/watch?v=PgsHZfHhI5Q
Axios headline: "Sam Altman says OpenAI shares Anthropic's red lines in Pentagon fight"
Yeah, no company wants to be associated with mass domestic surveillance and autonomous killer robots. Ok, some do. Actually, lots do. . . just not these two AI companies. For now.
Oh they would totally do it if there wasn't so much public scrutiny ... basically the only thing holding them back is Hagueseth's inability to keep his goddamn mouth shut ... I mean I hope these two wouldn't but I'm too cynical to believe they wouldn't if given a chance.