AI Cybersecurity
Field notes on AI on both sides of the fight: vulnerability discovery and disclosure rules, dependency governance for developers, and defenses against agents pushed into risky actions.
AI is no longer only the thing being defended; it now hunts for vulnerabilities, writes disclosure rules and reviews dependencies. We start with developer platforms like Snyk that push security left, then look at how generated search queries lower the cost of reconnaissance, and end with zero-day offense and defense: once a model can draft an exploit chain, the first things to tighten are permissions and approvals, not just content filters.
