Harness debate: it should be model-provider infrastructure, not your engineering
robleclerc · x · 2026-10-31
Replying to mattshumer's hot take that "harness engineering is usually stupid," robleclerc argues the confusion should end: a harness is just the loop + general-purpose tools that let an LLM do work, it should be fungible and ideally provided by model providers — no OpenClaw or Hermes needed. What organizations must actually build is the "cockpit and vehicle": the net of workflows, rules, tools and data around agents, which is not the harness.
Related event: Shumer Sparks Debate: Is Harness Engineering a Mistake?(3 posts)→
More from coding & agent
- Dev tool UIs are converging: Vercel, Supabase, Cloudflare and ChatGPT sidebars look alike — ColleenMBrady · 2026-11-01
- 9.1B tokens in 105 days: a power user's AI usage stats — Angaisb_ · 2026-11-01
- Anthropic publishes Skill authoring best practices: keep SKILL.md under 500 lines — bibryam · 2026-11-01
- Dev builds entire Three.js city with no 3D models using Codex and GPT-6 Astra — OpenAIDevs · 2026-11-01
- Compiler-backed coding agent Benzi hits 82.6% on SWE-bench Verified for $9.26 — DonkeyTheKing · 2026-11-01
- Ronie: free open-source AI agent that makes Ray-Ban Meta glasses actually smart — Odd-Sun6296 · 2026-11-01