Decisions

Use this file for durable project decisions. Keep entries short and dated.

2026-07-05: Create A Separate Learning Lab Repo

Decision: RS AI Learning Lab will live in its own standalone repo at projects/rs-ai-learning-lab.

Reason: The Learning Lab supports AI Readiness but has its own curriculum, contributor, community, and pilot operations. Keeping it separate avoids absorbing the AI Readiness Kit and lets contributors collaborate without access to Obsidian or internal servers.

2026-07-05: Use GitHub Issues For Contributor Work

Decision: GitHub Issues and a GitHub Project board are the v0 operational queue for work items, assignments, kanban, and reviews.

Reason: Multiple contributors may participate, and they will not all use Obsidian or have access to RS servers.

2026-07-05: Keep Markdown As Durable Source

Decision: Curriculum, templates, decisions, roadmap, policy, and pilot notes remain Markdown files in the repo.

Reason: Markdown keeps the learning artifacts portable, reviewable, and renderable into multiple future formats without committing v0 to an LMS, app, or publishing pipeline.

2026-07-05: Keep V0 Narrow

Decision: V0 builds the project spine, templates, initial backlog, first module sketches, contributor workflow, and pilot prep materials.

Reason: The Learning Lab needs enough structure to start review and pilot work, but not a full curriculum, LMS, database, app, generated schemas, or automation layer.

2026-07-10: Capture Broadly During Module Outlining

Decision: During the outlining and brainstorming stage, module sketches should favor capturing useful content over fitting a final session length. Pruning, sequencing, and final lesson boundaries will happen before instructional buildout.

Reason: It is easier to shape a strong learning journey from a well-developed source pool than to recover important ideas that were removed too early. Introductory modules are expected to lead into follow-up modules with greater depth and tool-specific instruction.

Implementation note: Keep likely deferrals in the source and mark them as Successor Module Candidate when they appear better suited to an advanced, tool-specific, or deeper follow-up module. A future pruning pass should allocate those ideas before deleting them.

2026-07-13: Use Internal-First Article Pairs

Decision: Substantial Learning Lab articles begin as internal-facing source documents. A public-facing counterpart is a separate, deliberate adaptation that remains visible on the internal intranet while it is drafted and reviewed.

Reason: Internal writing can state strategy, maturity, evidence, and open questions plainly. Public writing needs its own audience, claims, accessibility, privacy, and approval review. Keeping both visible together preserves traceability without treating a placeholder as an external release.

2026-07-13: Separate Production Lifecycle From Learner Availability

Decision: Keep the existing production lifecycle (Seed through Field Notes) separate from learner availability (Available, Drafting, Pilot, Planned, Deferred).

Reason: A module’s editorial state does not tell a learner whether they can use it now. Separate labels let the Lab be candid about both content maturity and participation options.

2026-07-13: Defer Status-System Changes Until Module Evidence Exists

Decision: Retain the existing separation between production lifecycle and learner availability during the first complete module-development cycle. Do not introduce a replacement maturity model, disposition field, catalog-wide migration, or new learner-facing status badges yet.

Reason: The current system already distinguishes editorial production from learner access. A new model would overlap with existing labels before the Lab has defined field ownership, transition criteria, format-level status, migration rules, or learner-facing presentation. The AI Without the Hype build and one contrasting early-stage module will provide evidence for a focused governance pass.

Checkpoint: Complete the governance pass after the first module package has reached cold review or pilot preparation and after one earlier-stage module has been sampled. Complete it before standardizing a third module or automating status displays across the catalog.

2026-07-13: Use Learner-Directed Modular Participation

Decision: The RS AI Learning Lab is learner-directed and modular. Tracks organize related modules, references, activities, and suggested paths; learners may follow a recommended path or choose individual modules based on their goals, prior knowledge, time, and interests. A prerequisite is required only when it is genuinely necessary for safety or meaningful participation.

Reason: The Lab can begin with a useful set of modules, grow in response to learner needs, and support people with different goals without pretending that everyone needs the same route.

2026-07-13: Give Every Track a Track Guide

Decision: Every track will have a Track Guide explaining its audience, scope, concepts, suggested starting points, related modules and tracks, access requirements, available formats, optional capstone, support, stewardship, and current limitations. A Track Guide is not Module 0 and is not a required lesson.

Reason: Learners need a map around the modules even when they choose their own route.

2026-07-13: Make Track Capstones Optional

Decision: Each track should eventually have its own optional capstone module. Capstones can support cumulative practice, portfolio evidence, completion recognition, or deeper application, but are not required for selected-module participation.

Reason: Different learners need different finish lines.

2026-07-13: Establish Global Safety and Responsibility Guidance

Decision: Shared safety and responsibility guidance will live at the RS AI Learning Lab level. Tracks and modules should include concise, content-specific reminders until an approved global guidance page exists; they must not link to a destination that is not yet available.

Reason: Privacy, consequential decisions, credentials, payment information, third-party data, verification, and human responsibility apply across the Lab and should not drift through copied wording.

2026-07-13: Rename the AI Literacy Track and Introduction

Decision: Rename Zero-to-Hero AI Literacy to AI Literacy for the Rest of Us, and Welcome to AI Without the Hype to AI Without the Hype.

Reason: The new names better reflect an invitational foundation for ordinary people. The track does not promise mastery; it supports safe curiosity, practical skills, and informed judgment.

2026-07-14: Integrate Work-Systems and Bounded-Delegation Ideas Into Existing Modules

Decision: Place the Work Is a System, Not a Job material inside Friction Mapping and the From Shadowing to Bounded Delegation material inside Decision Ownership Before Delegation. Do not retain either as a standalone module route.

Reason: The ideas strengthen existing learner jobs and artifacts—discovery before recommendations, then accountable authority before delegation—without creating duplicate curriculum or implying a larger active scope.

2026-07-16: Centralize Hands-On Activities In The Lab Library

Decision: All hands-on activity packages live in the top-level labs/ library. Module pages link to their related labs, and each lab declares its primary module, related modules, and tracks in frontmatter.

Reason: A single canonical location lets contributors discover, delegate, review, and enrich activities without maintaining competing module-local and central patterns. Existing activities were structurally migrated without claiming that missing v1 support material already exists.

2026-07-16: Permit One Static Learning-Asset Exception

Decision: The Module to Interactive Single-Page Application package may include its existing offline static browser prototype as an internal learning asset.

Reason: The prototype reinforces a defined context lesson without accounts, APIs, analytics, cloud storage, persistence, or learner-data collection. This exception does not authorize an app platform, LMS, or broader product-development lane.

2026-07-31: Require Explicit Glossary Derivative Releases

Decision: Canonical glossary status makes a term eligible for derivative use; it does not publish or update the intranet, a visual, or a Short automatically. A separate human-reviewed release records independent website, visual, alt-text, and Short decisions and is bound to the canonical term versions and hashes.

Reason: Meaning, presentation, and platform publication mature independently. Explicit releases preserve human control, expose stale derivatives, and prevent a canonical edit or publicly observable video from silently changing another surface.

Implementation boundary: Glossary derivative production happens as the first working collection in the isolated rs-ai-learning-media-production project. The Astro intranet consumes only an approved release bundle, and the workspace-owned promotion command is the only workflow that copies approved artifacts into Astro.

2026-08-04: Prove Artifact-Scoped Reviews Before Migrating Every Lab

Decision: Use Forty-Five Seconds to Make It Clear as the master proof for artifact-scoped reviews and independent final QC. During the proof, its older review URLs may route into the shared desk so existing bookmarks remain usable. After the proof is accepted, migrate every lab to the generated manifest and shared receipt contract, audit that no artifact identity, receipt, comment, or route intent was lost, and then remove the legacy review implementations. Permanent backward compatibility is not a requirement.

Reason: A temporary compatibility bridge makes the proof reversible. Keeping two review architectures after the full migration would add cognitive and maintenance cost without adding provenance value.

2026-08-04: Make The Canonical Activity Learner-Readable

Decision: Use lab.md as the complete learner-readable Activity. Keep package placement, lifecycle, review, and source-location metadata in the hub and generated provenance records. A presentation renderer may reorganize approved Activity sections, but it may not silently rewrite their meaning, instructions, safety guidance, or sequence. Wording changes require a new Activity review or a separately fingerprinted presentation-copy review.

Reason: Reviewing an internal content specification does not establish that a learner can complete the rendered Activity. One learner-readable canonical source reduces drift while still allowing the future public page to use a flexible layout. Deliberate internal/public adaptations remain available for articles and other artifacts serving genuinely different audiences.

2026-08-17: Add Invitation To Be Curious Again As An Active-Development Track

Decision: Add Invitation To Be Curious Again as a Draft / Drafting, mixed-age, learner-directed track. Use the suggested path Becoming Curious Again → shared Safety baseline → learner-chosen practice. Keep Creative Projects, Simulations, and Learning Anything visible as future directions without creating empty modules.

Reason: Curiosity needs a clear learner path and complete practice packages without pretending that every learner needs the same sequence or that AI is the destination.

Implementation boundary: Curriculum remains canonical in this repo. The Astro project renders and verifies it. New lab video records remain Deferred; no media belongs in this repo or the website. Availability cannot advance without learning-design, mixed-age, accessibility, privacy/security, applicable-rights, and pilot review. Automated or synthetic QA is not pilot evidence.

2026-09-07: Make It Starts With a Question The Curiosity Signature Lab

Decision: Add It Starts With a Question as the canonical signature lab inside the Becoming Curious Again learner journey. Keep the lab in the central Lab Library and link it from the module and track. The activity follows one learner-chosen question through one bounded exploration, distinguishes observation and source material from generated material and inference, keeps uncertainty visible, and leaves the next step or stopping decision with the learner.

Reason: The Curiosity track needs a repeatable practice that embodies its own purpose. Housing it as a canonical lab makes the activity portable without separating it conceptually from the signature module.

Implementation boundary: AI remains optional and is never the destination of this lab. The Question Path may be spoken, drawn, arranged, or recorded and is never a required submission. Beta means the core activity is runnable; it does not replace learning-design, mixed-age, accessibility, privacy/security, applicable-rights, or learner-pilot review.

---

Source: decisions.md