Hello,
The previous episode with Eric Saraniecki, co-founder of Canton Network, was about what institutions need to start using blockchain rails. For the latest, we talk about how blockchains need the ability to reason and act autonomously to be embedded in an agentic economy.
I recently sat down with Mike Hanono, co-founder and CEO of Talus Network, to talk about how crypto can help AI move forward.
The $342 billion invested in AI by five major tech companies would rank as the 45th-largest GDP in the world. However, almost none of that investment has been spent on making AI accountable or verifiable. Most companies have internal governance frameworks for verifiability, but there’s nothing visible to the consumer. You are flying blind when you use most models. I’ve written about AI markets and verifiability here in detail.
Half the traffic on the web is bots, and computers have already started driving us around. It won’t be too long before the agentic economy is real, too. By that, I mean agents that hold assets, make decisions, and coordinate with other agents. These agents will need to live inside rule-based systems, interacting programmatically and economically, where every action leaves an audit trail. And this is where blockchains enter the picture.
Talus is building the infrastructure to make this happen. Blockchains of today are strict. Input A gives output B. They lack a brain, the ability to reason and adapt. And they lack hands, the ability to act autonomously. Talus’s bet is that incorporating both reasoning and autonomy into blockchains is essential to advancing AI commerce.
As a market participant, I’m curious about where the value accrues in our experiments with AI. Remember the fat protocol thesis? The idea that in crypto, protocols capture most of the value while applications remain thin. But we’ve seen that change.
Applications that users actually want to use are starting to capture value, too. Platforms like Aave, Base, Hyperliquid, and Jupiter are living proofs that it’s not just about the protocols. I believe something similar happens with AI. Value concentrates at both ends. Chip manufacturers like Nvidia, on the one hand, and applications that build relationships with users, on the other.
See you in the next one,
Suarabh Deshpande
