Anthropic's Methos model set off something close to panic last month. Banks, tech giants, and even governments all scrambling over fears that a new era of AI-enabled cyberattacks had arrived. But, cybersecurity experts told me that that era, it's already here.Multiple AI researchers have been able to reproduce many of Methos's most alarming results. This is finding previously unknown vulnerabilities in crucial software programs using older, widely available AI models. That technique, it's called orchestrati ...