Ex-Meta engineer: AI usage generates a fragile codebase — 90% AI code didn't 2X output
bibryam · x · 2026-11-22
Evan X. Merz, laid off from Meta in May (partly, he suspects, for resisting AI coding adoption), argues AI usage produces fragile codebases ill-suited for scalable SaaS.
- He generated at least 90% of his code with AI in his final months (the minimum for a decent rating), yet AI didn't even 2X his output.
- He saw brittle code from AI-heavy peers — including a top engineer merging JS-event-based mouseover effects instead of the :hover selector.
- Now at another large company, he's 100% AI-generated code because it's expected, but argues AI solves what fits its context window and is oblivious to long-term architectural issues.
- The post lists four specific ways AI usage results in a fragile codebase.
More from coding & agent
- App Store Connect CLI 7.1.0 Adds Product Page A/B Test Results in the Terminal — rudrank · 2026-11-22
- Auto Ontology MCP: open-source semantic layer lets Claude and ChatGPT query your DB safely — solyarisoftware · 2026-11-22
- 10 skills to fix the cookie-cutter look of AI-generated websites, drop into Claude Code or Cursor — sven_ai · 2026-11-22
- A catalog of 2,256 real agent use cases with copyable prompts, from bill-negotiating bots to 11-agent trading teams — sven_ai · 2026-11-22
- Dev ports Inkling-2-Mini to a single AMD R9700: 196K context at ~110 tok/s — abhinand05 · 2026-11-22
- Disney Parks MCP Server offers structured queries for attractions, dining and Lightning Lane status — modelcontextprotocol · 2026-11-22