Illia Polosukhin of NEAR on AI Operating Systems & Blockchain Infrastructure

Illia Polosukhin, co-founder of NEAR Protocol, joins The Smart Economy Podcast to discuss AI operating systems, agent harnesses, chain abstraction, and the growing role of AI agents in blockchain ecosystems. Polosukhin explains how NEAR’s original vision emerged from AI infrastructure challenges, why AI harnesses are evolving into operating systems, and how NEAR Intents enables users and agents to interact across blockchains without dealing with backend complexity. The conversation also explores AI security, formal verification, confidential inference, and why blockchain infrastructure may increasingly operate behind the scenes as AI agents become primary users.

In this episode of The Smart Economy Podcast, host Dylan Grabowski is joined by Illia Polosukhin, co-founder of NEAR Protocol, a modular blockchain platform focused on AI infrastructure, chain abstraction, and agent-based systems. Together, they explore the evolution of AI harnesses into operating systems, how AI agents may reshape blockchain usage, and why blockchain infrastructure could increasingly disappear into the background for end users. Polosukhin explains how his background in machine learning and distributed systems led to the creation of NEAR, originally inspired by the need for a scalable global payments infrastructure to support AI data-labeling networks. The conversation traces how that early vision evolved into NEAR Intents, confidential AI infrastructure, and agent operating systems like IronClaw.

What you’ll learn:

Why Illia Polosukhin believes AI harnesses are evolving into operating systems
How IronClaw approaches memory, permissions, and agent workflows
Why most mainstream AI usage still resembles “glorified Google” search
How NEAR Intents aims to abstract away blockchain complexity for users and agents
Why blockchain infrastructure may increasingly operate behind the scenes
How AI agents are beginning to transact across crypto ecosystems
Why bridge security remains one of crypto’s largest attack surfaces
How formal verification could improve smart contract security
Why businesses are beginning to adopt AI agents operationally
How confidential AI inference helps preserve user privacy
And much more!

Illia Polosukhin is the co-founder of NEAR Protocol and a longtime researcher in AI and distributed systems. Before launching NEAR, Polosukhin worked in machine learning and natural language processing, including contributions to the Transformer architecture that helped shape modern AI systems. His work at NEAR focuses on AI infrastructure, chain abstraction, confidential inference, and agent-based systems designed to support large-scale autonomous coordination. His perspective is especially relevant as AI agents increasingly intersect with blockchain networks, financial systems, and decentralized infrastructure.

Watch on YouTube: https://youtu.be/HT5NzWIpQM4

Did you enjoy this episode? Check out these other great guests:

And many others!

ATTENTION all blockchain, crypto, and Web3 professionals! If you're a founder, CEO, or expert doing something innovative in this space, we're interested in speaking to you! Apply to be a guest on our platform and connect with our vibrant community of blockchain professionals: https://bit.ly/SEP-guest