Timestamped crypto video analysis

Bitcoin Policy Institute Discusses AI Frontier Models, Code Vulnerabilities, and the Clarity Act

In this episode of the Bitcoin Policy Hour, Zack Cohen, Ken Egan, and Zack Shapiro from the Bitcoin Policy Institute (BPI) analyze the intersection of frontier artificial intelligence capabilities, open-source Bitcoin infrastructure security, and US legislative timelines. The panel details BPI's open letter—backed by more than 35 digital asset companies and development entities—calling on frontier AI laboratories (such as OpenAI and Anthropic) to grant vetted security researchers access to advanced automated red-teaming capabilities. The discussion centers on recent security incidents affecting Bitcoin infrastructure, notably Boltz pulling its swap infrastructure offline following targeted hot-wallet assaults, alongside security reviews in BTCPay Server. BPI researchers argue that attackers leverage AI tools to rapidly audit pre-AI codebases for exploitable vulnerabilities. Because US frontier labs restrict defensive red-teaming under standard safety guardrails, defenders have turned to foreign open-weight models (such as Kimi K3 and GLM) to find zero-day vulnerabilities before malicious actors can exploit bearer assets. Zack Shapiro highlights the convergence risk where misaligned autonomous agents seek bearer capital like Bitcoin, arguing that pre-AI code vulnerability represents a more immediate operational danger than quantum decryption ('Q Day'). Additionally, Ken Egan outlines the legislative outlook for the Clarity Act, noting that Senate floor action was postponed through the August recess and scheduled for a procedural cloture vote on September 15. The team also assesses Meta's open-weight AI distribution model and Google's internal compute allocation dynamics, while concluding with a call to support open-source Bitcoin security red-teaming via OpenSats.