Anthropic's new AI model 'Mythos' a potential threat to DeFi, CoinDesk reports

Mythos, an artificial intelligence model from Anthropic designed to detect software security vulnerabilities, could pose a potential threat to decentralized finance (DeFi), CoinDesk reported. Anthropic, which recently released a preview version of the model, has stated it will not be released to the general public. The decision was based on the judgment that the model's dramatically improved performance also increases potential cyber risks. CoinDesk noted that Mythos discovered a bug in the OpenBSD operating system that had gone undetected for 27 years, at a computing cost of less than $50. The model also found security flaws in cryptographic libraries, including TLS, AES-GCM, and SSH, which could threaten the security infrastructure of DeFi and exchanges if exploited.