Skip to content

Strengthen cooperative learning through evidence and quality experiments - #35

Merged
NellInc merged 3 commits into
mainfrom
codex/cooperative-quality-rehearsal-20261018
Oct 18, 2026
Merged

NellInc merged 3 commits into
mainfrom
codex/cooperative-quality-rehearsal-20261018

Conversation

@NellInc

@NellInc NellInc commented Oct 18, 2026

Copy link
Copy Markdown
Owner

Summary

  • Bring completed cooperative handoff review into the decision loop, with early-trial limits and obligations kept explicit.
  • Connect four cooperative roles to captured plant facts and existing inspection views.
  • Add an optional shared quality hypothesis and real isolated-flow rehearsal: packing output never grants dispatch permission.
  • Preserve live stock, quality dispositions, pay, ballots, role authority, interface and world assets.

Checks

  • 34 focused tests pass, up from 29 at the final pass baseline.
  • Typecheck, lint, format, depth/shader policy, design detector, production build, bundle, reachability and inhabited-world checks pass.
  • Final build: 431 files, 169.88 MiB, under the unchanged 170 MiB budget.
  • 23 final native Metal desktop/mobile views across day, dusk and night; known/missing inputs, frozen prediction, reset and state purity checked. Earlier passes retain their separate native evidence.
  • Secret and 100 MB outgoing-tree/history checks pass.

Boundaries

  • Existing pacing/adaptive-quality issue remains open separately, with its failed receipt retained and no waiver.
  • Physical player/device/manual first-person checks remain deferred.
  • Muted throughout. No new dependencies, remote compute, saved schema or runtime actor/inventory injection.

@NellInc
NellInc merged commit bc09c3c into main Oct 18, 2026
5 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant