<!-- last-reviewed: 2026-06-02 -->
ADR-0020: Icebox the venture (channel + learning platform)
- Status: Accepted · Date: 2026-06-02 (UTC) / 2026-06-01 ET · Decider: Principal ({{PRINCIPAL_NAME}})
Context
{{PROJECT_NAME}} has two distinct halves:
- The organization and its internal tooling — the org-as-code workforce, the Documentation & Knowledge Program (COMPLETE, G3), the intranet portal, the Slack integration, the finance system, and the CI integrity machinery. These are built, live, and sound.
- The venture — the “show the work” YouTube channel and the retrieval-first web
learning platform (
docs/concept.md,specs/IDEA-0001-show-the-work.json): the dual-flywheel business {{PROJECT_NAME}} exists to eventually serve. This half is greenfield — the platform data/retrieval layer is unbuilt (ADR-0006 deliberately shipped the static intranet first), and a keystone strategic decision (business-model direction: monetize vs. AI-for-good) is unmade.
The venture has been sitting as a half-open question — generating “parked / TBD / awaiting Principal” threads across the org (the business-model decision, brand go-live, the platform build, learner-PII SOPs, etc.). A half-open question costs attention. The Principal has decided to close it cleanly rather than leave it ajar.
Decision
The venture is iceboxed — formally parked, indefinitely, until the Principal explicitly thaws it. No org capacity is spent pursuing the channel or the learning platform, and no venture decision (including the business-model direction) is pressed on the Principal until thaw. The organization continues to operate, maintain, and improve its internal tooling (intranet, Slack, docs, finance, integrity machinery) as normal — iceboxing the venture does not freeze the org.
Scope
Iceboxed (no active work; not on the live backlog):
- The YouTube channel (content, brand go-live, channel naming/art).
- The web learning platform (retrieval/scoring engine, the
platform/build, the product data layer beyond what exists, learner accounts/forum). - The business-model decision (monetize vs. AI-for-good) —
reports/2026-06-02-business-model-decision-brief.mdis preserved for thaw, not pending. - Venture-gated threads: learner-PII lifecycle SOP, platform-launch SOPs, mobile/learner
a11y, platform telemetry seams (the lone
spend requested/usage data readypending edges stay legitimately pending per ADR-0006).
Stays active (the org runs normally):
- The intranet portal, the Slack integration, the finance system, the documentation corpus, and the CI integrity gates — maintained and improved as usual.
- Open operational risks unrelated to the venture (e.g. Slack credential rotation, R-001 deletion-fallback) remain live and are not iceboxed.
Preserved, marked parked (not deleted):
- The venture artifacts created 2026-06-02 (
docs/brand/,docs/knowledge/growth/flywheel-model.md,docs/content/, the concept doc) remain in the corpus as real, preserved standards ready for thaw. They carry an icebox note pointing here; the dead-link gate still protects them.
Thaw trigger
The icebox lifts only on an explicit Principal decision to pursue the venture. At thaw, the first two moves are: (1) the business-model decision (brief is ready), and (2) a venture start plan (platform build sequencing + channel launch). Until then, the PA does not resurface venture work except to answer a direct question.
Consequences
- Positive: the org’s attention is fully on the operating backbone and tooling, which is where value is being created today; no recurring “what about the venture?” drag.
- Cost / tradeoff: the dual-flywheel mission is paused — {{PROJECT_NAME}} produces no public content and serves no learners while iceboxed. This is a deliberate sequencing call, not an abandonment.
- Supersedes/clarifies: sharpens ADR-0006 (static-first deferral) from “platform later” to “venture iceboxed until thaw.” Does not change any internal-tooling ADR.
- Reversible: fully — every venture artifact is preserved; thaw is a one-word Principal call.