Fundamentals
Clean data in did not mean clean answers out. The checks that still catch 4 to 9% of AI output, the other ways to verify it, five things AI did in production that I did not see coming, and a cheat sheet for where AI belongs.
Thirteen hours a day, eight of them with AI. Why AI is not like a bug you can fix, what building with it really looks like, and how much to trust it depending on what is at stake.
Layered security from hardware to runtime: TPM/HSM, WAF, CI/CD scanning, secrets rotation, and encryption. Practical implementation beyond HTTPS.
How to build professional grade security layers on a lean, single node stack using Rails and Kamal.
Mitigations for secrets handling, OAuth bypasses, dependency typosquatting, prompt injection, and more.
Security is not easy and that is why so many products get it wrong part 1.
