Shreya Murthy

They can decline to not permit the company to move forward.

And I would expect that even a small result, like we send them their model, they reviewed it for 30 days and they didn't say it was bad.

will be a good signal for AI labs that submit through this process.

But Dean Ball continues, he says, this is a fairly major win for the safety contingent within the admin and a significant loss for the SACS accelerationist wing and is surprising to me.

I continue to think this EO is a mistake.

This is clearly teeing up the infrastructure for a model licensing regime.

And the fact that the administration is classifying the details of how this voluntary system will work is egregious.

So everyone wants to know,

what are they going to be evaluating?

Are they going to be doing, you know, RKGI tests or some other benchmark or humanities last exam or just red teaming it or just talking to it?

Who knows?

It's classified.

Everyone is interested to know.

Dean Ball's certainly interested to know, but we don't know yet, but maybe in the future.

The public, he continues, the public and the employees of the labs have a right to know how this works.

It is very frustrating when you're interacting with the federal government and you don't know their

rubric for grading you.

That is one of the most frustrating pieces of going back and forth with the regulatory body.

It's best when there's a very clear test for the DMV.

You show up.