Two years of AI coding: one dev's full setup, 11 MCP servers cut to 2
ParticularFan1122 · reddit · 2026-10-21
After two years of AI-assisted coding, the author shares the setup that survived — gains came mostly from cutting things, not adding them:
- Coding: Claude Code for planned work, Codex for background chores; stopped switching models weekly — cost more time than it saved
- Planning: 3-4 lines of plain-English ticket before the agent touches anything; half the bad sessions came from skipping this
- Rules: one AGENTS.md under 100 lines (a 900-line version got mostly ignored)
- Review: CodeRabbit on every PR (noise dropped after disabling style nits) plus a local Claude review pass; reads payment/auth code personally
- Testing: Playwright for the 5 money flows (signup, login, checkout, invite, cancel); agent-written tests trusted less
- Infra: GitHub Actions, Vercel, a $12 VPS worker, Sentry for errors — agents read stack traces better than humans at 1am
- Dropped: 11→2 MCP servers, multi-agent setups that looked cool on Twitter, long sessions (new task, new session)
More from coding & agent
- Claude Code Pushes Urgent Security Update Across All Installations — umeshonai · 2026-10-23
- Claude cut a 2.6-hour video export to 20 minutes — but only after being told to rebuild the server — byhewar · 2026-10-23
- 3 flight-booking rules my AI agent checks that I'd never do by hand — Efistoffeles · 2026-10-23
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Building an autonomous SEO agent in Hermes with three skills: programmatic pages, Keywords Everywhere, DataForSEO — EXM7777 · 2026-10-23
- Building a sports betting AI agent with onchain execution and human approval by default: architecture and lessons — Dry-Comfortable-2678 · 2026-10-23