Jyunmi Hatcher

Or they'll be making, you know, drastic changes to who they are, right?

Which I find interesting in this positioning because of this story from the Wall Street Journal, which is...

They've changed their responsible scaling policy or their RSP, which is a volunteer frontier model safety framework and works in practice.

So from Time and the Wall Street Journal says the company removed a central pledge that would have blocked it from training new systems unless it could guarantee adequate safeguards.

Anthropic said that it's updating the RSP to reflect uncertainty in evaluation of

and evaluation science and the reality that unilateral commitments can break if competitors keep building.

So instead of a blanket no-go promise, the updated approach emphasizes transparency and conditional actions, including publishing the frontier safety roadmaps and the risk reports.

tied to capability thresholds.

Now, this may be a non-starter type story where here we're just making it more flexible to be able to tell you more, but not necessarily the coincidence, but the timing of having to run into this military conflict, plus you making any form of change to your safety policy or safety framework,

For the most part, all of these foundational model groups, they are engineer focused, right?

That's who they started.

We even get mixed messages from OpenAI, even though it's, you know, led by a big time CEO and been in that space.

But we've talked about how...

startup messaging, being a startup and messaging versus being a monolithic organization messaging and how they've already run into problems with that.

You're totally right about Anthropic.

They're sort of in the space of like...

They're known for cloud code, right?

They're engineer mindset.

It's like, yeah, yeah, none of this matters.

Let me just do it.