According to CoinMeta, a Bitcoin security “red team” used multiple AI models to audit approximately 150 Bitcoin-related codebases and has already disclosed over a dozen vulnerabilities, involving projects such as wallets, cryptography libraries, and infrastructure. Team members stated that they can currently find about one serious vulnerability per person per hour on average. In the past 12 hours, they have submitted security reports to multiple projects, but the specific affected projects have not yet been made public. The team is also developing an open-source AI security auditing platform, with the current scanning cost being around $10,000 per day. Recent security incidents involving coldcard and boltz also indicate that AI is being used by both security researchers and attackers to more quickly identify software vulnerabilities.