Now is probably one of the best times to build, but it's also one of the noisiest and craziest time to build.

For every idea, there's 10 competitors you know of and up 20 you never even heard about.

You don't even know existed.

Can we talk about the deep seek moment?

High Flyer was very quant heavy.

We bought all these GPUs to do heavy trading.

The government basically banned high frequency trading.

It was like, we have all these

GPUs, we're very smart people. We can't really trade in the same way we could. What do we do?

Right? Let's go do some AI. Imagine what happens if you said, hey, all those smartest people

from Wall Street, like instead of doing high frequency trading, go do AI. How much further ahead

could we even a AI? I either look at the technologies being something that's very closely matching our brain

or that our brain is quite not what we thought it was, you know?

As we understand LLMs and these models better, perhaps that our brain was not what we thought

it was.

We're not always rational first principle logical thinkers, right?

As a cost of building software that goes to zero,

building a tech company ends up being kind of like drop shipping.

You take the AI-generated code from Antarctic Open AI,

you arbitrage some like go to market,

and you repackage it and do a pretty scary thought.

The takeaway there for people, I think, is like,

...