Mythos, China, and a New Era of AI Regulation

So we've had quite a wild roller coaster ride.

Basically, going back to Mythos and Anthropix's decision to not release this to the public, we've had an executive order that was canceled at the last minute.

And then a different version was signed about voluntary pre-release testing, among other things.

And then most recently, we have these sudden export controls on both Mythos and Fable, which is the sort of safety-restricted version of Mythos.

And the reaction from Anthropic was basically, you know, if this is going to be export controlled for foreign countries and foreign nationals within the US, that means they effectively had to block both models entirely.

So here we are.

What are your reactions? What are your views on all this?

We will get into some of the details, but I think the high level is a bit of anthropic champion this view and convinced a lot of people that these things are of great risk and are reaping some of that.

I think it's...

It's not great because their virtues in some ways, I think, are valiant for what is coming.

But they are ahead of the game, which is causing whiplash sooner than it is needed, which is disruptive.

And then the government being... much worse of an actor and how they respond to this in a mix of punitive measures we've seen i mean like pete had seth heg seth had this tweet which was insane which they just do this stuff it's like okay this is obviously at least partially punitive which is way over the line and then even the implementation is a bit clunky and what's happening because i think realistically i think there would not be an immediate crisis.

Even without the safeguards, if they were without the classifier safeguards, I think there would probably be most likely things would be okay.

There's a chance that something really bad happens.

I think I've priced in for a while that leading AI models would cause minor bumps along the road and things will keep getting worse.

But it seems unlikely that Fable is actually that based on everything we've seen.

And a lot of people have had access to Mythos even before.

Even people who aren't supposed to have access have had access to Mythos.

There's like the one fabled Discord server that had access for a long time.

So there's a lot of signs of Anthropic being a bit high on its supply and sanctimonious about safety and implementing features which are really mostly good for their business.

So it's really hard to unpack because there is truth that AI safety is an important thing.

And this company has the leading models, but also they are pretty antagonistic towards China, open weights and open science.

And why I was mad was because that shifted from rhetoric to action on the open science front.

I think it was pretty obvious that they thought they're the only people that should do AI research.

But if they... stay out of the way it's not going to be a big deal because i think there's a lot of people that talk a lot in these communities of ai safety and don't really do anything so i was kind of giving them a pass until they actually did anything and then i was like well now they're now they really mean it and they can't get a pass on this so maybe we should break down more of these and you should Prod me to clarify with the different issues at play.