"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis User-Owned AI: On-Chain Training, Inference, and Agents, with NEAR's Illia Polosukhin
133 snips
Sep 13, 2025 
Illia Polosukhin, co-author of the influential 'Attention Is All You Need' paper and founder of NEAR, shares his bold vision for user-owned, privacy-focused AI. He discusses NEAR’s cutting-edge blockchain infrastructure that prioritizes decentralized model training and privacy through NVIDIA’s confidential computing. The conversation highlights the importance of trust mechanisms, economic security in AI, and the potential for blockchain to empower users in AI governance. Illia calls for community engagement in creating transparent practices for a future where AI truly belongs to everyone.
AI Snips
Chapters
Transcript
Episode notes
Choose Permissionless For Censorship Resistance
- Prefer permissionless networks to avoid censorship and cartel control over transaction inclusion.
- Join or run a validator yourself, and stake tokens or delegate to preserve network neutrality.
AI Makes Formal Verification Necessary
- AI increases the attack surface by finding vulnerabilities, so systems must offer mathematical/verifiable guarantees.
- Formal verification and verifiable compute become essential as automated attacks accelerate.
Confidential Computing Enables Private, Verifiable Inference
- Nvidia confidential computing enables encrypted execution where operators cannot see model weights or user data.
- This creates a permissionless, verifiable inference layer with only small performance overhead.