AI Composer - Safe Vibe Coding

LLMs generate code that appears correct yet omits critical security logic. In adversarial environments like DeFi, this is a structural failure, not a bug. This talk examines why GenAI cannot reason about adversaries, and why formal constraints are re