Dev diary — thoughts on DevOps, AI, and the journey.
If you pick a coding agent because its harness scored higher on a leaderboard, you are probably choosing on noise. If you pick it because of what it costs per task, you are choosin
If your multi-agent system fails, you rerun it, and it passes, you have learned almost nothing. That is the uncomfortable result in Repair or Resample? Rethinking Failure Debugging
Every long-running agent has the same problem. Reasoning piles up. Step 40 is still carrying the deliberation from step 3, a decision that was made, executed, and observed 37 turns
If you run an LLM bug-fixer in a loop until it says the code is clean, you should expect it to break working code faster than it repairs broken code. That is the uncomfortable head
If your agent's tool returns an error, the agent tells the truth. If the tool returns status: ok with garbage inside, the agent will often make something up. That is the finding in
Persistent memory is showing up in agent platforms as a product feature. The agent finishes a task, a second agent reads the transcript, decides what was worth keeping, and writes
A new paper on Kubernetes config repair contains one number that should bother anyone running an agent that edits files without a human watching: when a tolerant patch tool is hand
A skill file has no build step. That is the entire problem. When code goes stale, something breaks loudly. A renamed function throws. A moved endpoint returns 404. CI goes red and
A decade of code review research already told us the approval on a pull request is worth less than everyone treats it as worth. Reviewers skim. Approvals arrive in seconds. LGTM is
> Status: drafted, not published. See Publishing Status section at the bottom before treating this as live anywhere. Researchers at King's College London built 72 simulated regula
One dispatch per week — what I shipped, what broke, and what I learned from the field. No filler.