Architects vs gardeners: developer argues AI's plan-maxxing obsession hurts good work
nptacek · x · 2026-10-16
A developer pushes back on the current obsession with plan-maxxing before building with AI. She splits people into "architects" who envision everything upfront and "gardeners" who discover the work as they write/build/paint, arguing AI is no different. All her consequential work came from excitedly tugging on interesting threads — never from planning projects to appease PMs — and she worries the pre-planning trend will be to our detriment.
More from coding & agent
- CoPE adds one forecasting loss to GRPO, doubling long-horizon agent sample efficiency — Lianhuiq · 2026-10-17
- laya-guard 0.5.0 adds AI attack simulation and file-tool hooks to guard Claude Code — Extreme_Ad5709 · 2026-10-17
- Laude's open-source RevEngine review bot takes #1 on GitHub ReviewBench in a week — lateinteraction · 2026-10-17
- Laude open-sources Headlong: a sub-10K-line Bash microharness for agents that never stop thinking — lateinteraction · 2026-10-17
- Hacktoberfest 2026 pivots to open-weight AI: 500+ events, 3,000 projects in week one — jonmarkgo · 2026-10-17
- MCP server runs Claude Code, Codex and Gemini as subagents with account rotation — Kynori20 · 2026-10-17