ANTHROPIC WILL LET OUTSIDE EVALUATORS OBSERVE ITS MODELS BEING TRAINED AND RUN, AND DARIO AMODEI WANTS THE WHOLE INDUSTRY TO DO THE SAME Asked whether slowing down means Anthropic will stop releasing more advanced models: "It doesn't mean that." He says the evaluators may end up setting the "speed... limit." "The first step is that we put in place embedded evaluators, embedded external evaluators from a third-party nonprofit or perhaps someday government organization, that can observe the process of us training and running our models... I want this to be done across the industry so that whenever anyone builds an AI model, there's a third-party evaluator who understands the best safety practices, who can verify whether that company is following the safety practices that they're committed to. It's like a food inspector."show more

@epictrades1 So people that left your company to work at these outside companies … gotcha not suspicious or anything

@epictrades1 Who are these “Evaluators” ?

Kinda like the nuclear weapons inspectors in Iran?

Outside observers is classify: access, not assurance. Owner for the useful bit is the publish/exit clause — and whoever can walk findings out without a lab veto. Amodei's invite isn't the rule until that's in a contract the Office can lean on. Stop equating "may observe training" with enforceable third-party eval.

Transparency on training runs is a meaningful shift, but the real test is whether evaluators get access early enough to influence release decisions.

@StockMKTNewz So the tail is wagging the dog, why can’t you set your own fucking speed limit ffs

deepseeek puts their hand up first

third-party evaluators setting the speed limit makes a lot of sense. if this becomes standard across labs, safety checks could finally have some real teeth.

How can you be so smart but then talk like you need guidance over yourself control of what you created. This makes no sense at all, and you got all the approval from the commander in chief to keep pushing forward. Our president has spoken so there’s no need for regulation.

In other words, open source will destroy our margins and mess up all future investments. Long live the Ai bubble trade

David Sacks challenge is what comes next. Why should voluntary restraint by the companies actually at the frontier require a new regulatory architecture governing everyone else?

I wonder how many employees at Anthropic are watching him and thinking all that money they were going to make on that IPO is going down the drain.

Nice one