Bittensor creator Const on Affine, dTAO, "mining reasoning," and more | E2326
Summary
Jason Calacanis interviews Jacob Steves, founder of Bittensor, on how the platform applies Bitcoin’s permissionless mining model to artificial intelligence. They discuss Bittensor’s subnet architecture, where independent networks compete to provide AI services like inference and training, validated by cryptographic proofs. The episode highlights how this decentralized approach creates hyper-competitive markets that drive down costs and foster innovation against centralized tech giants.
Topics discussed
Intro: Bitcoin speculation and BitTensor introduction
BitTensor origins: Connecting Bitcoin and AI
Abstract consensus and subnet architecture
How subnets work: Incentives and validation
Permissionless mechanism selection and competition
Mining reasoning models and latent space
The Templar subnet controversy and rug pulls
Foundation role, market dynamics, and J-Trade strategy
Listen ad-free on Castria