Logan Kilpatrick

Hello.

Happy Friday.

Apologies for being a few minutes late.

I was just jamming with the team about a bunch of cool new stuff coming in February.

So the targets are already set for next year, which is exciting.

I'm relieved because for a second I thought you were going to say next week.

Not next week, next week.

Folks are rightfully taking time off.

So it'll be nice for people to relax and have a moment of calm before the storm reemerges.

Thank you.

Yeah, yeah, the model's incredible.

I feel like it's a, I think it's a testament to the work that's been happening on research.

It's a testament to like what happens when you have great frontier capable models with Pro as well, sort of like being the model that then distills down into Flash.

I think a lot of the,

I've seen a bunch of comments online of folks being like, this model sort of defies the paradigm of what a Flash model actually has historically meant, which I think is actually a good, reasonable comment, just because Flash used to mean small and fast, and now it's like actually the Flash model is...

in some cases, like more capable than the pro model, which is really, really cool.

So that's right.

Yeah.

And I mean, part of this, there's like a very, very simple answer to this, which is just given like one of the ingredients of making great models is time.

And the pro model was done, you know, probably like, I don't know,