It's Protest Time
Added 2026-02-26 04:14:12 +0000 UTC
Comments
It's a big topic that would warrant a full video to explain in detail. My understanding is that Anthropic refused to allow the Pentagon to use Claude for mass surveillance of Americans or autonomous killing machines, and the Pentagon responded with "change your mind or else". Anthropic chose "or else" so the Pentagon is trying to hurt Anthropic by cancelling their contracts with the pentagon and designating them a supply chain risk. The supply chain risk is interesting because it's never been publicly applied to a US company before. The Pentagon also threatened to use the Defense production act to force Anthropic to remove their safeguards from Claude. How can the Pentagon say that Anthropic's products are a supply chain risk, and also simultaneously that anthropic's products are so important we're forcing them to work with us? Unclear. In the same week, apparently unrelated to the Pentagon stuff, Anthropic rolled back one of their major AI safety promises, so it's not like they're perfect angels either, but it is good to see an AI company walking away from big money contracts on safety principles. A lot of the AI Safety community dislikes Anthropic because 1. they're still an AI company 2. they don't mind military applications of AI 3. they're trying to build superintelligence Personally I think Anthropic are the least bad AI company, but that's kind of like being the least bad parasitic worm species in your breakfast In summary, in my opinion, Anthropic is in the right and the Pentagon is being insane
Siliconversations
2026-03-05 16:23:00 +0000 UTCHey Silizium Gespräche, what's up with the Anthropic mess? I mean what's going on? Is the Pentagon's ban justified? I mean on first glance it seems to be and it seems to be about being cautious of AI. But then again it seems the reason the Pentagon is going after them is precisely because they DIDN'T want to help the pentagon build Terminators and help them spy on people. So could you shed some light on this?
Emmanuel Goldstein
2026-02-28 19:29:57 +0000 UTCpresumably many people who sincerely care about AI safety!
Siliconversations
2026-02-26 15:12:51 +0000 UTCI'm still in rural northern canada so I can't be there unfortunately, but I'm sure it'll be great!
Siliconversations
2026-02-26 15:12:38 +0000 UTCI do wonder who I'll meet there
C_Bookie
2026-02-26 11:02:13 +0000 UTCI wish I could go but I've got work and a birthday party to go to but have fun in London!
raided
2026-02-26 08:45:57 +0000 UTC
