Jyunmi

It's an official release.

On Anthropic.

OK, so the subtext of the story in this is Gwen and I are having a conversation in the chat, too.

Anthropic is the good guys, but it does not mean that they're good in the way that like they're not the white knight hat people.

They just have some standards that they're not willing to go beyond.

And when you hear them talk, it sounds more like because we don't trust the AI right now to do the thing correctly.

So we don't want to be involved in that.

It's not that the system can't use an autonomous weapon to kill somebody.

This is my paraphrase, so if I'm totally wrong about the thing, come let me know.

But it's not that AI can't control an autonomous weapon to kill somebody.

It's that we don't believe that it is sophisticated enough to do that in a reliable way

And from the cheap seats, it's like, don't y'all just drop bombs on a house to kill one guy?

But yes, if you don't have to do that and you're making sophisticated weapons.

I have one quick also.

Okay.

Yesterday, Anthropic announced that they had rolled out a bug that had dealt with prompt caching that caused usage limits to be consumed faster than normal.

So if you are using cloud code, make sure that your version is 2.1.62 and they reset everybody's usage limits so everybody started from scratch yesterday.

Have a good weekend.

What you're seeing is people riding bikes in a gym and doing rowing machines.

Very humanoid.