System76 has banned AI-generated code across its Pop!_OS/COSMIC codebases as a first-party provenance-and-quality policy; comparable explicit bans (or formal rejections) by other major open-source projects — or System76's ban remaining an isolated one-off — resolve whether AI-code exclusion becomes a recognized open-source governance pattern.
state: watchingheat: lowuncertainty: mediumconvergesscott: mediumai-code-governance open-sourceSystem76
What is this?
System76's COSMIC desktop-environment project — the Rust-based DE anchoring Pop!_OS — has added an attestation to its pull-request template requiring contributors to certify they included no LLM-generated content (code, comments, or descriptions), that they fully understand the change, can respond to review comments, and have tested it; engineer Jeremy Soller says the rule responds to maintainer strain from a flood of unplanned, rarely-accepted AI-first contributions, with AI tools for non-generative uses still permitted. Supplied reporting frames this as the strictest response in a wider maintainer-overload wave: the Linux kernel and Ubuntu have accepted AI content behind quality bars rather than banning it, and adjacent episodes (a Void Linux maintainer orphaning 100+ packages in an AI-policy dispute, Go maintainers refusing AI-generated code) are circulating in the same news cycle, so the spread-or-stay-isolated question is genuinely open. The snippets are press reporting on the template change — they establish the COSMIC PR-process scope clearly but do not confirm how far the policy extends beyond COSMIC into other Pop!_OS repos, nor the enforcement mechanics beyond the attestation itself.
Why it matters to Scott
A consequential party has independently arrived at territory Scott's canon already maps: 'rarely-accepted AI-first contributions' straining maintainers is exactly his code-churn / verification-cost economics made binding policy, and an attestation attached to the PR delta at capture time is a crude instance of his provenance-coupled-work and governance-as-code claims — but System76 picks the exclusionary remedy he does not hold, which is the dated-receipts opening: his 'drive it with discipline' thesis (plus the 84% Claude merge-rate study) argues provenance-bearing, verification-gated contribution over bans. Medium rather than high because it extends his claims and hands him a distinctive position in a live OSS-governance debate (whose spread-or-stay-isolated resolution will show whether bans harden into the default) without changing any active build — though it does touch his own AI-assisted PR practice on GitHub and continues the radar's densest lineage as its strictest ban instance.
ip:source.your-ai-can-code-you-just-don-t-know-how-to-drive-it-ebookip:concept.code-churnip:concept.verification-costip:source.provenance-coupled-work-ebookip:source.governance-as-codework:project.githubradar:openjdk-ai-generated-code-banradar:rust-llm-contribution-policyradar:debian-2026-llm-governance-resolutionradar:linux-drivers-llm-code-policyradar:linux-ai-patch-review-overloadradar:linux-kernel-agents-mdradar:coding-agent-pr-merge-rates
queries asked of Scott's wikis
- AI-generated code policy
- agent contributions to open source repos
- maintainer review overload economics
- AI slop code quality
- code provenance attestation DCO
- AI-assisted vs AI-generated distinction
Measured heat
now 0 pts/hpeak 56 pts/hcomments 0/hpeers p14momentum: steady2 platformsage 290h
points/hour across evidence · reading as of 2026-10-12 02:59:37.977291+11:00 · deterministic, not a model opinion
How the heat travelled
pace: p81 vs 1188 stories at the 168h mark (now 290h old) — ahead of byteshape-qwen38-27b-quants (1.0x), behind voodoo-quant-mit-release (1.0x)
Evidence (3) — ⭐ canonical anchor
Interpretation history
2026-10-06T01:43:33Z
The attached GNOME-orbit software-quality essay confirms the AI-era quality debate is still generative, but it is discourse, not the resolution condition — no second formal ban has surfaced, so System76's act remains a settled, enforced first-party ban under pattern-watch while the case itself sits at ~0 pts/h.
2026-10-05T23:34:21Z
evidence attached: hn.story.49944392 — A prominent OSS maintainer's software-quality-in-the-AI-era essay is direct context for whether AI-code exclusion becomes a recognized open-source governance pattern.
2026-10-04T23:02:35Z
The news cycle has crested and cooled — the Oct-4 velocity spike was the HN thread's final burst and it now sits at ~zero velocity (peer percentile 10) — so the case's meaning shifts from breaking policy act to a settled, enforced first-party ban under pattern-watch. Enforcement is attested by a contributor whose Claude-assisted PR was closed, and SQLAlchemy's maintainer describes equivalent informal filtering of LLM-padded third-party PRs, but no second formal ban has surfaced, leaving the spread-or-isolated resolution condition open.
2026-10-03T19:40:54Z
origin walked (opencode/cheap-glm, conf 0.95): anchor hn.story.49946321 -> echo.github.9d08d55214 by Jeremy Soller (jackpot51, System76)
2026-10-03T19:37:41Z
grounded: converges/medium — A consequential party has independently arrived at territory Scott's canon already maps: 'rarely-accepted AI-first contributions' straining maintainers is exact
2026-10-03T19:27:26Z
case created — A concrete first-party policy act by a notable distro vendor, reported via press with HN discussion — a bounded OSS-governance episode with a resolvable spread-or-stay-isolated outcome, and no open case covers it (the Thelio Mira hardware case and the kernel AGENTS.md thread are different claims); the Neowin candidate original is press reporting, not the origin.
Decision trace
- 10-11 14:29review_screenjev screen: no material development (noul=0.17)
- 10-06 12:43repriceThe attached GNOME-orbit software-quality essay confirms the AI-era quality debate is still generative, but it is discourse, not the resolution condition — no second formal ban has surfaced, so System
- 10-06 10:34attachA prominent OSS maintainer's software-quality-in-the-AI-era essay is direct context for whether AI-code exclusion becomes a recognized open-source governance pattern.
- 10-06 10:30propose_attachA prominent OSS maintainer's software-quality-in-the-AI-era essay is direct context for whether AI-code exclusion becomes a recognized open-source governance pattern.
- 10-05 10:02repriceThe news cycle has crested and cooled — the Oct-4 velocity spike was the HN thread's final burst and it now sits at ~zero velocity (peer percentile 10) — so the case's meaning shifts from br
- 10-04 08:21sensor_dirtyvelocity_spike
- 10-04 07:22sensor_dirtycomment_update
- 10-04 06:40promote_anchororigin walk conf 0.95
- 10-04 06:37groundA consequential party has independently arrived at territory Scott's canon already maps: 'rarely-accepted AI-first contributions' straining maintainers is exactly his code-churn / verif
- 10-04 06:27createA concrete first-party policy act by a notable distro vendor, reported via press with HN discussion — a bounded OSS-governance episode with a resolvable spread-or-stay-isolated outcome, and no open ca