Andy Halliday

Do not, this should not, you should not be asking me to say yes or no to this question.

You should be giving me the parameters that you're asking and the consequences of each answer, right?

Like, but.

I don't know.

I think it depends on what the source intelligence is, but I almost agree with you, Brian.

I think that part of what we're seeing in research is models converging on answers, and that happens when there's some giant corpus of books and documents and that kind of stuff.

Yeah, Hitchhiker's Guide to the Galaxy and the number 44.

Right.

That is part of the source of most of the systems that were built because nobody thought it was going to work right away.

Let's just grab what was free.

The story from Anthropic is that that wasn't an official decision.

Some guy just had it on his computer and added it.

But those kinds of things are converging on answers that could be wrong.

But if you have specialized knowledge or a particular thing that it's been trained on subsequently, now you're actually having and you have multiple.

agents or multiple models that have that information i think then you're closer to yes all of them are not going to get it wrong right most of them found it on page seven and somebody found it on page 92 and probably page seven was correct by the way carl joined us hey carl how are you

Hi, Carl.

True.

Yep.

And the most recent releases, not the code leak, but the most recent releases of Clawed code equal all of those capabilities that Claw has, right?

So they basically released a scheduler.