Trump orders US agencies to stop using Anthropic technology in clash over AI safetyhttps://apnews.com/article/anthropic-pentagon-ai-hegseth-dario-amodei-b72d1894bc842d9acf026df3867bee8a

[rysiek@mstdn.social]

Basically, Secretary of Defense Pete Hegseth wants to be able to use Anthropic's AI models for things Anthropic is not comfortable with: mass domestic surveillance of Americans and fully autonomous weapons.

Anthropic said "no". Whiskey Pete is throwing a tantrum, and Trump is supporting him in it.

This is highly entertaining on many levels.

One level is that a company peddling dangerous technology to fascists is suddenly learning the hard way how instrumentally the fascists treat them.

[rysiek@mstdn.social]

Another level is that even in their rebuttals of Department of Defense (as it is still officially called!), Anthropic appeases the fascists by using the unofficial "Department of War" moniker:
https://www.anthropic.com/news/statement-comments-secretary-war

That's one of the reasons why I am not taking Anthropic's pushback seriously. It's not about any actual moral stance, it's about being able to do business with the fascists while not looking too bad publicly.

[rysiek@mstdn.social]

In that context, ironically, Hegseth is not wrong about how sanctimonious the rhetoric of effective altruism is:
https://xcancel.com/SecWar/status/2027507717469049070

But here's the kicker – all the AI bubble barons are in the effective altruism and the wider TESCREAL bundle:
https://techwontsave.us/episode/198_how_effective_accelerationism_divides_silicon_valley_w_emile_torres

It's only become a problem for Whiskey Pete now because Anthropic is being difficult.

[hans@social.woefdram.nl]

@Michał "rysiek" Woźniak · I'm surprised nobody seems to understand that in a democracy such fundamental questions about how far a nation can go in autonomous weapons and mass surveillance, should be a matter for parliament.

This shouldn't be decided between a trigger-happy minister on one side and the director of a commercial enterprise on the other.

[edelruth@mastodon.online]

@rysiek

I'm waiting to see if Anthropic caves to this naked, coersive, bullying.

[rysiek@mstdn.social]

But the best part is that, as noted by Dean W. Ball – a former Senior Policy Advisor for Artificial Intelligence and Emerging Technology at the White House Office of Science and Technology Policy, under Trump! – "DeepSeek is now treated much more kindly by the United States government than anthropic":
https://xcancel.com/deanwball/status/2027540776402747853#m

Hegseth designating Anthropic as a "supply chain risk" means that no entity providing services to the US government can do business with Anthropic.

[rysiek@mstdn.social]

Consider what this means in practice. Again Dean W. Ball:

> Nvidia, Amazon, Google will have to divest from Anthropic if Hegseth gets his way. This is simply attempted corporate murder. I could not possibly recommend investing in American AI to any investor; I could not possibly recommend starting an AI company in the United States.

X Cancelled | Verifying your request

(xcancel.com)

So…

Would be super funny if that's what causes the bubble to start popping.

[rysiek@mstdn.social]

@Edelruth it will.

Shareholder primacy is a well-established doctrine in US law, at least since Dodge v. Ford Motor co. (1919):
https://en.wikipedia.org/wiki/Dodge_v._Ford_Motor_Co.

So, Hegseth threatens Anthropic with (effectively) corporate murder, this is bad for shareholders, shareholders can sue Anthropic to force it to cave.

I'm sure somebody will go into more detail and correct my simplistic hot take, but in broad strokes I believe it's correct.

[drwho@masto.hackers.town]

@rysiek

[drwho@masto.hackers.town]

@rysiek It wouldn't surprise me if ICE "mistook" a few of Anthropic's key engineers for "illegals" and kidnapped them.

[rysiek@mstdn.social]

Oh, and I forgot about another level of comedy here: US government throwing a tantrum because *checks notes* it does not have enough technological sovereignty from US tech behemoths.

While at the same time actively pushing against EU's policies related to technological sovereignty:
https://www.reuters.com/sustainability/boards-policy-regulation/us-orders-diplomats-fight-data-sovereignty-initiatives-2026-02-25/

This is truly comedy gold.

[viss@mastodon.social]

@rysiek whats the bet that all these military laser mishaps are because of ai, and swapping out anthropic models for deepseek ones will let china steer the lasers

[rysiek@mstdn.social]

@Viss Hanlon's Razor applies. Incompetence and pettiness (on part of DoD) in communicating with the FAA is a sufficient explanation. Or are you talking about some other mishaps?

[rysiek@mstdn.social]

Also, this would be a *great* time to ask other AI hype peddlers – Microsoft, OpenAI, Meta, Google – if they are comfortable with US DoD using their models for mass domestic surveillance of Americans and fully autonomous weapons.

Put them on the spot. Make them squirm trying to answer this question in a way that is not a major PR disaster, and not find themselves on Hegseth's "naughty" list.

[viss@mastodon.social]

@rysiek no, thats the one. yesterdays and then 2ish weeks agos version of the same. its just amusing to think that they tried to get claude to drive the laser around and it just started shooting at everything

[lordbowlich@masto.hackers.town]

@rysiek

They do their little song and dance now to buy some time to set up the necessary subsidiaries to launder their works through and so in a decade they can go "whoopsie didn't know they were doing that"

[woozle@toot.cat]

@rysiek I keep hearing that they've refused to drop their safety pledge, but then TIME reports that they did -- so I am now confused.