Start here
- How AI and Agentic Development Are Reshaping Software Engineering: the long one. Where the judgment sits once a model writes the first draft, worked end to end against a real system design problem.
- Claude Code ships the batteries, Pi hands you the wires: two coding agents with nearly the same tools and opposite bets on who owns the harness, compared decision by decision.
- Running a coding model locally with Ollama: getting a model onto your own GPU, and the moment a vendor capability flag turns out to be a claim rather than a proof.
- Stacked PRs changed how I ship code: keeping review from becoming the bottleneck, and where the tooling for it still fights back.