I’m surprised it took this long for investors to devalue those stocks. There have been AI tools for finding security issues for a while now and they’re better than what those companies offer. Anthropic’s new feature may not even be the best one. Typical security auditing tools have always been bad. All they do is find simple patterns in code that could lead to security issues. They give you a list that’s full of false positives and you have to look over them yourself and see if any are legitimate. AI tools are non deterministic, but they find much more complex patterns that are more likely to be real issues.