Episode Summary
Executive Summary: Brian Armstrong argues Coinbase is evolving from a crypto exchange into an "everything exchange" for tokenized assets, stablecoin payments, and AI-driven finance. He sees agentic commerce and AI wallets as a major new market, while internally Coinbase is using AI to speed software development through recursive self-improvement. He also describes New Limit’s longevity work using AI to discover epigenetic reprogramming therapies, plus broader views on cognitive enhancement and special economic zones.
Main Topics: Coinbase as an "everything exchange" (Priority: 5/5): Armstrong describes Coinbase's expansion beyond Bitcoin into a broad on-chain financial platform where users can trade stocks, commodities, crypto, derivatives, perpetual futures, and prediction markets in one place. Stablecoin payments and agentic finance (Priority: 5/5): He argues stablecoins are becoming a global payments layer and that AI agents should have financial accounts, self-custodial wallets, and spend tools so they can transact autonomously. AI transforming financial services (Priority: 5/5): The conversation covers Coinbase’s 'AIFY' strategy: AI advisors for users, agent accounts, and new protocols like X402 enabling cheap machine-to-machine commerce. AI inside Coinbase and recursive self-improvement (Priority: 4/5): Armstrong explains how Coinbase is using internal AI 'brains' for codebases and teams, so agents can ingest company context, make changes, and improve future performance through feedback loops. Tokenization and global access to assets (Priority: 4/5): He discusses tokenized stocks, future tokenization of credit, treasuries, and deposits, and the idea that billions of people could gain access to U.S.-style investing via wallets. New Limit and longevity biotech (Priority: 5/5): Armstrong outlines New Limit’s work on epigenetic reprogramming, large pooled screens, and early therapeutic targets in liver, vascular, and immune cells, with the aim of rejuvenating aged tissues. Broader civilizational ideas: cognitive enhancement and special economic zones (Priority: 3/5): He advocates for improving human cognition and health, discusses embryo editing and bioenhancement, and supports special economic zones/freedom cities as innovation sandboxes.
Key Arguments: Agentic commerce needs crypto rails because many AI transactions are tiny; card fees make sub-dollar payments uneconomical. AI agents will need their own accounts and spend authority, just like human assistants or employees. Stablecoins and tokenization will modernize financial services by making money move at the speed of information. Coinbase can use AI not just for coding assistance but for recursive self-improvement by feeding review corrections back into a shared team 'brain.' Prediction markets may extend beyond sports into policy analysis and even 'big life questions,' making them broader social information markets. New Limit’s approach is to target the root aging process through epigenetic reprogramming rather than treat individual age-related diseases one by one. Improving human cognition and health, including through embryo editing and other bioenhancement tools, may become normalized over time. Special economic zones could reduce regulatory friction and serve as safe sandboxes for high-velocity experimentation.
Data Points: Minimum card transaction fee: About 30 cents flat plus a percentage - Used to explain why card rails are impractical for very small agentic payments. Share of agentic commerce transactions under 30 cents: About 76% - Armstrong cites this to show why crypto-native payments are better suited to AI commerce. Coinbase revenue from non-Bitcoin trading: 88% - From the company’s recent earnings deck, showing Coinbase is already broader than Bitcoin trading. Prediction markets revenue run rate: $100 million annualized within months of launch - Armstrong says prediction markets scaled quickly inside the Coinbase app. Prediction markets growth: Over 100% quarter over quarter - He describes rapid growth momentum for the product. Global access gap to brokerage/U.S. investing accounts: About 4 billion people - Used to motivate tokenized stocks and broader global investment access. Global crypto users/holders: About 700 million - He estimates the size of the broader crypto-holding population. Monthly active crypto users: About 50 to 100 million - Armstrong distinguishes active users from total holders. New Limit team size: About 50 to 60 people - Describes the current lab and research operation in South San Francisco. First clinical trial timing: Next year - He says New Limit’s first phase one trial is planned for the following year. Initial New Limit indication: Alcoholic liver disease (ALD) - The first liver program targets patients with high unmet medical need. Potential market size for ALD program: Around $20 billion - Armstrong estimates the therapy could be highly valuable if successful. Public support for embryo editing for disease prevention: About 80% of Americans - He cites Pew Research to argue the topic may become less controversial.
Pivotal Quotes: "We don't want the AIs to be unbanked. You know, we want to bank the AIs." — Brian Armstrong: On why AI agents need financial accounts and autonomous payment rails. "It's kind of like recursive self-improvement" — Brian Armstrong: Describing Coinbase's internal AI system that learns from human reviews and feeds corrections back into the team 'brain.' "I think crypto was really, really good for humans and it's going to be essential for AI." — Brian Armstrong: On the future role of crypto rails in both human and agentic economies.
Implications: Coinbase is positioning itself at the intersection of crypto, AI, and tokenization, while biotech and civilizational tools like longevity, enhancement, and regulatory sandboxes could reshape access, productivity, and markets over the next decade.