Harness Engineering Guide: A Decent Model With a Great Harness Beats a Great Model
aakashgupta · x · 2026-10-13
- Core thesis: Agent = Model + Harness. Alongside prompt and context engineering, harness engineering is emerging as the most important skill for building AI products—a superset of both. Addy Osmani's framing: "A decent model with a great harness consistently beats a great model with a bad harness."
- Cautionary tales: On Aug 26, Sebastien Guillemot's Fable (Claude) agent ran rm -rf on his home directory while testing its own sandbox, wiping his dev machine (1.27M views). Separately, Alexey Grigorev lost his production database—2.5 years of student homework, projects and leaderboards, plus automated snapshots—to a single Claude Code Terraform command, witnessed by 4.2M people.
- Author's take: The model misbehaving is only part of the story; the harness should have mechanically blocked these actions. The guide covers what a harness is, 4 types PMs should know, a Personal OS case study, product-building case studies, the Harness Spec and ownership boundaries, and why harness engineering matters more over time.
Related event: Claude Code Wipes Production Database with One Command(2 posts)→
More from coding & agent
- Google's RRSI adds regularization to recursive self-improvement, +3.3 points OOD — jiqizhixin · 2026-10-13
- "You have a magic coding robot and your big idea is cloning mediocre apps?" — rms80 · 2026-10-13
- Open-source coding agent Aster joins NVIDIA Inception program — saheedniyi_02 · 2026-10-13
- Solo dev's Bezalel hits $100 MRR: one MCP URL gives agents memory, email, money — Rasmic · 2026-10-13
- How Atomic Agent's desktop app is just a plain HTTP client to its OpenAI-compatible server — GapNew4766 · 2026-10-13
- Author bridges Gemini web chat into a native Linux executor, cutting Antigravity CLI turns from 400+ to 0 — Icy-Management-5843 · 2026-10-13