I'd love to hear from you the nuances there and how have you identified the trends early and the conviction you're getting on the moonshot idea here?

Yeah, I still think we're in the era of reinforcement learning.

Reinforcement learning has gotten a little bit of a pushback because all the VCs just decided to pile into reinforcement learning.

I think like 90% of them don't even know what reinforcement learning is.

But the truth is reinforcement learning has generalized far better than any other domain.

If you do reinforcement learning on one thing, the model just gets better at everything else because it learns how to solve for that reward function and it learns new strategies.

very much like the pathway to recursive self-improvement, even maybe to continual learning, may heavily depend on reinforcement learning.

Tell me more about when you see founders pitching, what gets you excited? What makes for a great pitch that helps you build conviction?