Austin Griffith

I think that that counts if it's an agent working on your behalf the same way if you had, say, an employee working on your behalf to do things with like MetaMask or do things on Jane, it would count in that way.

I think a thesis that we've talked about on Bankless several times in the past, whenever AI and crypto has come up is the true crypto natives are actually AI agents.

And what we mean by that is like, because they are software and because they can sort of think at the speed of light and because they are native programs, DeFi and Ethereum and crypto is going to be

even more, I guess, domestic to them than it is to human beings, right?

So in the same way that I might feel a little like a foreigner doing things in the digital landscape, whereas if an AI was trying to manifest in a bank branch, that would be very strange and awkward, right?

They are...

finely tuned for crypto rails, and thus they will become the dominant player.

Do you guys believe that?

And what about AI agents make them so conducive to using crypto and DeFi and all of the tools that we're about to talk about?

So let's paint a picture before we go on.

We're going to talk about some of the standards that you mentioned in a little bit, but let's paint the picture because there has been maybe, as some might call it, a multibot moment.

All right, so there's this project, I believe it was formerly called Cloudbot, that sort of broke out a week or two ago and now is taking the agent world by storm.

It is now called OpenClaw, I believe.

This is an open source project.

And I know, Austin, you've been tinkering with it and giving it on-chain types of skills.

Before you talk about what you've been doing and how this actually works, can you set this up?

What is OpenClaw?

What is happening with it?

Why is there so much buzz?

So Austin, just to understand, so OpenClaw, basically you set it up in a sandbox type of environment, maybe a Mac mini, so that's somewhat isolated.