When the Agent Breaks Everything: A Postmortem
Real incident postmortem: an AI agent introduced a circular dependency that saturated production. How I detected it, stopped it, and prevented recurrence.
// construir-con-ia
You use Claude Code or Cursor daily and feel the code slipping away from you. Here: agentic discipline, tech debt, and taming the black box.
Real incident postmortem: an AI agent introduced a circular dependency that saturated production. How I detected it, stopped it, and prevented recurrence.
After letting an AI agent run free, I discovered the missing piece was architectural governance, not better prompting — and here is how I fixed it.
AI models deliver working code, but without architectural awareness they silently accumulate structural debt. Here is what that looks like in practice.
AI coding tools feel magical until 3am when your production alert fires. Here's what senior engineers won't post on Twitter about vibe coding at scale.
How to configure an autonomous development workspace that writes code, manages infrastructure, and learns at scale — with real SRE automation lessons.
A deep dive into ochoajorge.me's tech stack and how I used AI agents to design a zero-backend, scalable content platform built on Next.js and Vercel.