Would it not be possible to detect these kind of errors using machine learning now? To go through the code base and find unsafe operations? There must be enoug

Would it not be possible to detect these kind of errors using machine learning now? To go through the code base and find unsafe operations? There must be enough security bug fixes out there on GitHub that certain classes of security error become easily detectable.