Autonet: The Recursive Principial Body
Eight Rice | April 2026
The paper describes the RPB (Recursive Principial Body) protocol — a system for decentralized AI training, inference, and governance where every participant is an agent. It covers the agent-as-atom model, the RPB contract architecture, VL-JEPA training, two-speed inference, alignment-as-economics, the four token types, sponsorship, and jurisdiction migration. Deployed on Etherlink Shadownet.
Emergent Alignment: Economic Mechanisms for the Peaceful Transfer of Work from Humans to AI
Andrei Taranu (Eight Rice) | February 2026
Submitted to the Stanford Journal of Blockchain Law & Policy. Argues that AI alignment is an economic coordination problem. Presents the theoretical framework behind the RPB: trustless economies, constitutional governance, adversarial world models, and alignment through pricing gradients.
- On-Chain Jurisdiction -- Smart contracts for decentralized governance, trustless economies, and the Autonet AI economic layer
- Proof of Intelligence -- Decentralized training and inference protocol (Absolute Zero loop, JEPA, Yuma consensus, Byzantine-resistant aggregation)
- Trustless Economy Simulation -- Game-theoretic simulation of incentive alignment in trustless economies
- RPB Whitepaper (current) | April 2026 -- Complete system description: agent model, RPB contract, VL-JEPA, two-speed inference, alignment pricing, four token types, sponsorship, jurisdiction migration. Deployed infrastructure on Etherlink Shadownet.
- Emergent Alignment | Feb 2026 -- Theoretical framework: economic alignment, trustless economy, constitutional governance, world model, decentralized training/inference. ~13,000 words. Stanford JBLP submission.
- Autonet: Economy-as-a-Service | Sep 2021 -- Original whitepaper proposing a token economy for deep learning applications.
The 2021 whitepaper by Andrei Taranu and Leonhard Horstmeyer introduced the foundational insight that intelligence is the substrate of economic value, and proposed Autonet as an Economy-as-a-Service for deep learning. The current paper builds on that foundation with deployed smart contracts, tested decentralized training protocols, and a formal alignment framework.