Ai-Coding

Essays and field notes on how AI coding changes software delivery in practice. This tag focuses on where velocity gains are real, where verification becomes the bottleneck, and how team workflows need to evolve when generation gets cheaper.

This tag archive currently includes 2 articles. It is designed as a practical map of how this theme shows up across writing on software architecture, delivery quality, and engineering execution. Instead of treating AI coding as an isolated keyword, these entries trace how the idea appears in real system constraints, team decisions, and implementation tradeoffs.

Recent entries in this archive include Code Is Cheap. Guarantees Aren’t, The Pull Request is Dead: Surviving the AI Code Avalanche. Reading these together gives a clearer view of recurring patterns: where tooling helps, where process matters more than syntax, and where verification or coordination becomes the limiting factor.

A representative thread from this set: If models write more of the code and humans increasingly review, constrain, and verify it, then popularity in the human-coded era is not the same thing as fitness for the next one. The harder problem is no longer just producing code. It is building stacks that … Use this as a starting point, then follow the rest of the archive to see how the same problem evolves across different contexts and constraints.

Related tag themes that frequently appear alongside AI coding include programming-languages, formal-verification, code-generation, code review, SDLC, verification debt. Those cross-links are useful when you want broader context instead of a single-topic view, especially for platform-level decisions and multi-team delivery work.

  1. Code Is Cheap. Guarantees Aren’t

    If models write more of the code and humans increasingly review, constrain, and verify it, then popularity in the human-coded era is not the same thing as fitness for the next one. The harder problem is no longer just producing code. It is building stacks that can survive cheap generation without collapsing under ambiguity, review burden, and correctness debt.

  2. The Pull Request is Dead: Surviving the AI Code Avalanche

    Code production is no longer our bottleneck. The newfound velocity of AI coding agents hasn't solved our problems; it has simply moved the bottleneck further down the pipeline, creating massive SDLC backpressure. The human "Looks Good To Me" on a PR is now the single biggest liability in deployment. It’s time to stop acting like typists and start acting like architects.