Corey Knowles

I mean, like, I mean, just saying that sounds crazy.

And, you know, and I think about AI struggling with even fairly easy contest problems, like as recently as a year ago.

Were there any, you know, insights or breakthroughs along the way specifically that really enabled this big leap we've seen over the last 12 months or so?

Or the approach it took even.

Yeah.

So here's the question, and it may even seem a little silly, but I want to ask it anyway.

What do you do on a day at OpenAI when you're not making these big breakthroughs or smashing competitive coding records?

Like, what's a normal day look like at OpenAI?

Opening Eyes Chief Scientist Jacob Pachocki, forgive me if I butchered his name, I'm so sorry, mentioned moving from this wind to automating scientific discovery over months and years.

And as part of the team that pulled off this wind, what do you think are some key technical challenges along the way in getting to where you can do that?

Wow.

Yeah, I keep thinking of that as... Oh, sorry, Grant.

Go ahead.

I was like, I keep thinking of that as a thing involving sensors and other things.

Like, how do you teach it what a spring day feels like, for example?

Like, how do you do that?

And I think there's so much frontier still to grow and different types of data and sources of data that these things can be fed that...

It makes me so optimistic about the future and the work that you guys and other labs are doing right now, especially in science.

Science is the one that Grant and I kind of always watch.

Wow.