Google just launched AI Threat Defense and it does not just find security vulnerabilities.

It fixes them.

The system scans your applications, prioritizes the highest risk ones first, identifies vulnerabilities, and then a repair agent called CodeMender verifies and patches them automatically.

Continuous testing runs in the background looking for unknown vulnerabilities before attackers find them.

The key difference from what Anthropic and others are doing with Mythos is that Google is not stopping at detection. It is going all the way to remediation.

Finding bugs is one thing. Closing them before anyone gets in is another entirely.