It's Friday, February 27th, 2026.

From Peachfish Productions, it's the GISTI, Mike Pesca.

So Anthropic, the company that runs Claude,

if you want to say it in German,

they're making big news today.

They're getting attacked a bit by Donald Trump.

He doesn't want to allow them in government offices again.

Okay, we'll get to that in a second. But let's go back a couple days to set the scene.

Earlier this week, Anthropic rewrote a key promise in their responsible scaling policy, backing away from a 2020 commitment to slow down potentially dangerous AI development.

There is no point, really, if a competitor is going to get more dangerous anyway, and they trail

the competitor.

Ben Anthropics says they have two major concerns here, and it's because they just don't think

the AI systems are either reliable enough for this. These are things like autonomous weapons. They say it's just not reliable enough yet to be able to aim and shoot a weapon. They're also concerned about mass surveillance of US citizens, saying just our laws and regulations have not caught up to what AI can do when it comes to mass surveillance. And they say that the Pentagon's attempted changes did not satisfy that, and they are not budging on this.

And they say the military is welcome to go elsewhere.

They will help the military off board Anthropics products. They'll help them on board somebody new. And Dario Amadeh, the CEO of Anthropics, saying in a statement, regardless, these threats do not change our position. We cannot in good conscience accede to their requests. So basically what cloud, what Anthropic is saying, is that we won't be the most dangerous

kid on the block.

And since they are a leader, that means something, but not as much as it used to mean.

Here's the quote.

The policy environment has shifted towards prioritizing AI competitiveness and economic growth,

while safety-oriented discussions have yet to gain

meaningful traction at the federal level.

The premise of we will go slower for safety subject to a collective action problem.

...