The administration repurposed export laws to enforce model licensing.
Not at all. Hobbes would recognize the mechanism immediately: translate a domain into fear and you invite the sovereign to claim it. Once it does, it speaks its own language, decides in its own name, and ceases to be yours to manage.
The "this is what Anthropic asked for" take is stupid and wrong.
Users criticize Anthropic for pushing export controls and AI rules that would disadvantage competitors while shielding itself, and condemn regulators for enacting misguided policies they do not understand.
@shashj No, Anthropic got exactly what it wanted. The correct course of action here would have been to say “I’ll see you in court”, not immediate acquiescence to the government’s cackhanded attempt at overreach.
@kchoudhu Even if Anthropic intended to challenge this decision, it can’t ignore it. Not how it works.
@shashj The EU was very "serious" about the AI Act. It's terrible. This is the default mode of regulators regulating things they don't understand. So to that extent, it is the right response.
@shashj You are (intentionally?) missing the joke.
"But they asked for regulation that would screw competitors, not regulation that screw themselves! 😭"
Yes! We know! That's why it's fucking hilarious!
@kchoudhu @shashj I mean, we don't have to speculate what Anthropic / Dario Amodei wants: https://darioamodei.com/post/policy-on-the-ai-exponential