EP 37 — Building Useful AI on Crypto Rails with Ejaaz
Episode publishedPodcast: Decentralised.coDecentralised.co
Transcript preview
Preview only
Welcome to the Deco podcast, where we bring you stories from founders and investors in Web three. Before we get started, this is not a financial or legal advice. It's just a conversation about things we find interesting in Web three. Views expressed are our own personal opinions. The co members or guests appearing on the show may have positions on assets they talk about. Alright. Welcome to another episode. Today, we have Izzaz with us. I have been following Izzaz for quite some time and his takes on AI, how deep he is into AI, his episodes with Bankless as cohost on AI weekly roll up, and he'
for having me, Saurabh. Really looking forward to this one.
Alright. So let's start big. Pretty zoomed out view. I want to understand the entire AI stack from you. Right? I mean, you've been in the so called trenches for a while. So make me understand the entire stack right from GPU marketplaces to training to inferences to agents and agent marketplaces while explaining every stack as to what unique problems that crypto solves at every layer. And, you know, does crypto need to be there in every layer? I mean, should we be present across the layers, or there are certain layers that we should be optimizing for?
Yeah. Yeah. Really good question. So I'll get into it. So the way I kind of view the stack as you, described it would be at the bottom of this layer. You kind of have what I call, AI resources. So these are like, you know, the minerals and the soil and the earth that you need to kind of build the foundations of a building, for example. Right? So what are these things? You have data, you have compute, and then you have the model themselves. Right? So think of a model as, like, a really intricate designed chip or a dam where you are about to allow water to flow through. Right? You need to design
Wonderful. Thanks a lot for that explanation. I really love it. Now while thinking of these different layers, pretty sure that different layers have different value accruals. Right? And, therefore, different business models. Right? One of the examples that comes very quickly to my mind is virtuals. Right? When things were going alright, they were making enough fees, and we did not rely too much on inflationary token supply or those things. Right? I mean, the application itself or the protocol or the framework itself was generating revenue to sustain itself. Right? What are the examples or what
Yeah. I think that that is the golden question that no one has been able to answer. And I think if anyone gives you a straight answer to that, Saurabh, they're lying. And the reason why I say this is I think the crypto business model has only been figured out for a few things, and I mean really few things. You could argue that it's a monetary asset, which is Bitcoin. Right? And, you know, you get emissions from that. And, you know, eventually, once that emissions schedule declines exponentially as it were, you're hoping that fees on the Bitcoin network will be worthy enough for miners to do th
not specific to AI. I think that that is valid. It's across the world. Yeah. Yeah.
%. It's not specific to AI at all. This is a crypto problem that needs to be resolved. Right? And then on the application layer, if I'm being honest, you kinda see a similar thing. Right? So all the AI apps that I've seen exist today in the form of agents use some kind of staking or payment model. Right? And the common example is, hey. Pay to access this terminal where this agent will now give you alpha. That's AIXBT. Right? Or stake a certain amount of this token, and now you can get different tie
Keep exploring with a free account.
A free account preserves the short preview. Upgrade to Pro for the full transcript and episode chat.