4 Days of Markdown Docs Saved 4 Months: The Underrated AI Coding Artifact
alex_verem · x · 2026-09-08
Engineer mardehaym argues the most underrated artifact in AI engineering is a folder of markdown files. Before agents touch legacy codebases, teams should build a knowledge graph documenting every module, dependency, data flow, and domain term — and feed it to the agent before any code is written.
Results in practice:
- At a logistics client, building the graph took 4 days and saved roughly 4 months: the agent stopped violating naming conventions, importing deprecated modules, and wiring services through abandoned patterns.
- At a healthcare client, the same knowledge graph doubled as onboarding documentation for human engineers.
Most teams skip this step because pre-code documentation feels slow — the author argues it's actually high-leverage.
More from coding & agent
- Google GTIG: threat actors now run agentic AI attacks, harvesting credentials in under 6 hours — ChuckDBrooks · 2026-09-08
- Someone got Claude Opus 5 to play Ultima Online — mariofilhoml · 2026-09-08
- Dev hacks a Codex usage chart to watch rate limits plummet in real time — gpt2chatbot · 2026-09-08
- Agentic software factories: AI speeds coding but bottlenecks shift to review and ops — Pavan_Belagatti · 2026-09-08
- Turn PRs into RL environments at scale — startups already raised seed rounds on this repo — lvwerra · 2026-09-08
- Self-Improving AI's First Payoff Is Cheaper, Not Smarter: Ben Lorica on the Harness Advantage — bigdata · 2026-09-08