The Rollup Rollup TV: AI Supercycle, Rumor Mill, Privy, Stabled Up
Oct 10, 2025
In this conversation, Sam Battinelli, a builder and executive at Rise, dives into the intricacies of Ryze's low-latency design and the evolving order-flow market. He discusses the potential of on-chain decentralized exchanges and the strategic significance of their revenue models. Meanwhile, Hang Yin, an AI and privacy researcher, explores the challenges of confidential computing and the implications of recent TEEs vulnerabilities. He emphasizes the need for decentralized, auditable AI systems to empower communities and maintain privacy in the digital age.
AI Snips
Chapters
Transcript
Episode notes
Perps Are Eating Spot Market Share
- Perps adoption on DEXes is outpacing spot and may flip spot market share as infrastructure matures.
- Sam highlights open interest as a better metric than raw volume to track derivatives traction.
Encryption In Use Is The Hard Problem
- Encryption today mainly covers transit and rest; encryption-in-use remains the hardest problem for AI systems.
- Hang emphasizes confidential computing to protect data during model inference and training.
Decentralize Keys And Governance For Trust
- Design confidential AI stacks with decentralized key management and governance to avoid single points of failure.
- Use open standards and auditable evidence so users can verify software, hardware, and operational integrity.