Anthropic has rejected the Pentagon’s “best and final” offer, less than 24 hours before theUnited States Secretary of War Pete Hegseth 5:01 p.m. deadline.
https://x.com/shanaka86/status/2027158948445646942
BREAKING: Anthropic has rejected the Pentagon’s “best and final” offer, less than 24 hours before theUnited States Secretary of War Pete Hegseth 5:01 p.m. deadline.
The company reviewed the offer overnight and found insufficient progress on its two red lines: no mass surveillance of Americans and no autonomous weapons that fire without a human in the loop. Anthropic’s position has not changed. The Pentagon’s chief technology officer responded on CBS News today: “At some level, you have to trust your military to do the right thing.”
Everyone expects a Pentagon response today. The options on the table: invoke the Defense Production Act, cancel the $200 million contract, or designate Anthropic a “supply chain risk” alongside Huawei.
But here is what nobody has put together yet.
On the same Tuesday that Anthropic rejected Hegseth to his face, the company gutted its own founding safety policy. The Responsible Scaling Policy was Anthropic’s core promise since 2023: if a model becomes too dangerous and the safety measures cannot keep up, stop training. That was the tripwire. The whole reason Anthropic existed instead of just being OpenAI.
Anthropic removed that commitment. Its chief science officer told TIME: “We didn’t really feel, with the rapid advance of AI, that it made sense for us to make unilateral commitments if competitors are blazing ahead.”
Think about what that means. A company that just surrendered its founding safety promise to competitive pressure looked the Secretary of Defense in the face and said no to a $200 million contract, no to classified network access, no to the threat of blacklisting, no to the Defense Production Act, and no to being labeled a supply chain risk alongside Huawei.
They gave up the principle that made them Anthropic. But they would not give up the principle that protects you.
Two weeks earlier, the head of Anthropic’s Safeguards Research Team resigned. Mrinank Sharma wrote: “The world is in peril. I’ve repeatedly seen how hard it is to truly let our values govern our actions, where we constantly face pressures to set aside what matters most.”
The company proved him right by abandoning its broadest safety commitment. And then, in the same breath, proved him wrong by holding the two lines that actually matter for 330 million Americans.
Today at 5:01 p.m. we find out what happens when a company that has already given up almost everything refuses to give up the last two things the government wants most.
Full analysis on Substack.
open.substack.com/pub/shanakaans
https://x.com/shanaka86/status/2027158948445646942
posted by Satish Sharma at
05:21
0 Comments:
Post a Comment
Subscribe to Post Comments [Atom]
<< Home