Mythos, Anthropic’s most advanced AI model to date, has sparked fears about the threat to traditional software security after the AI ‌startup said the preview had uncovered “thousands” of major vulnerabilities in “every major operating system and web browser.”