Post

Here's a chronic AI-security-report fail: reporting vulns against obsolete patch versions. Not that humans don't do this too, but humans using AI tools do it constantly. Part of it is AI models' inability to parse or understand patch versions; remember that LLMs are bad with numbers ...