conductor(video_analysis): Initialize 12 child + 1 synthesis spec scaffolds

Each child spec is lightweight (~100 lines): references the umbrella, gives video details, specifies the 7 deliverables (transcript.json, frames/, ocr.md, report.md 1000-10000 LOC, summary.md), and the 5-phase pipeline. Children in execution order: 1. cs229_building_llms (Stanford CS229, Cluster E) 2. probability_logic (Cluster A) 3. entropy_epiplexity (Cluster A) 4. score_dynamics_giorgini (Cluster A) 5. platonic_intelligence_kumar (Cluster B) 6. free_lunches_levin (Cluster B) 7. generic_systems_fields (Cluster C) 8. brain_counterintuitive (Cluster C) 9. neural_dynamics_miller (Cluster C) 10. multiscale_hoffman (Cluster C) 11. cs336_architectures (Stanford CS336, Cluster E) 12. creikey_dl_cv (Cluster D) Plus 1 synthesis track (video_analysis_synthesis_20260621) blocked_by all 12 children.
2026-06-21 15:03:10 -04:00
parent b184250b78
commit 1b40fa5345
13 changed files with 1238 additions and 0 deletions
@@ -0,0 +1,96 @@
+# Track: Video Analysis — Stanford CS336 Lecture 3: Architectures
+
+**Status:** Not started (umbrella published 2026-06-21)
+**Type:** Research-only child track (Pass 1 of 3)
+**Owner:** Tier 2 Tech Lead (execution)
+**Cluster:** E (Stanford course VODs >1hr)
+
+> **Parent:** Child #11 of the [video_analysis_campaign_20260621](../../video_analysis_campaign_20260621/) umbrella. See [umbrella spec.md](../../video_analysis_campaign_20260621/spec.md) for full design.
+
+---
+
+## 1. Video
+
+| Field | Value |
+|---|---|
+| **Title** | Stanford CS336 — Language Modeling from Scratch, Spring 2026, Lecture 3: Architectures |
+| **Author** | Stanford CS336 Spring 2026 |
+| **URL** | https://youtu.be/lVynu4bo1rY |
+| **Cluster** | E (Stanford course VODs >1hr) |
+| **Estimated duration** | >1hr (Stanford course lecture) |
+| **Slug** | `cs336_architectures` |
+| **Execution order** | #11 of 12 (deep dive on transformer architectures; pairs back to CS229 with full context from prior videos) |
+
+**Pre-execution note (2026-06-21):** This video's oEmbed API fetch returned 401. This may indicate a private/age-restricted video; `yt-dlp` may still work. **Phase 1 of this track must verify yt-dlp access before downloading the mp4.** If `yt-dlp` also fails, fall back to manual transcript sourcing (if available) or escalate.
+
+**Position rationale:** Placed late (after Clusters A/B/C) rather than early so the Tier 3 worker has full context from CS229 + math + Platonic + biological videos when analyzing the architecture details. CS229 (#1) sets the "why" for transformer architectures; CS336 (#11) is the "how" deep dive.
+
+---
+
+## 2. Deliverables
+
+| Artifact | Path | Target |
+|---|---|---|
+| Transcript | `artifacts/transcript.json` | All segments |
+| Download log | `artifacts/download.log` | yt-dlp output |
+| Frames | `artifacts/frames/*.jpg` | 50-500 |
+| Extraction meta | `artifacts/extraction_meta.json` | Frame paths + hashes |
+| OCR | `artifacts/ocr.md` | Full OCR per frame |
+| Deep-dive report | `report.md` | **1000-10000 LOC markdown** |
+| Summary | `summary.md` | 200-400 words |
+
+---
+
+## 3. Pipeline (5 phases)
+
+- [ ] **Phase 1: Acquire.** Run `extract_transcript.py` + `download_video.py`. Verify `yt-dlp` access first.
+- [ ] **Phase 2: Keyframes.** `extract_keyframes.py` with threshold 0.4.
+- [ ] **Phase 3: OCR.** `ocr_frames.py`.
+- [ ] **Phase 4: Synthesis.** Tier 3 worker: transcript + OCR + frames → `report.md` (1000-10000 LOC).
+- [ ] **Phase 5: Verification.** Idempotency + audit + end-of-track report.
+
+---
+
+## 4. Report structure
+
+8 sections per umbrella spec §FR6.
+
+```
+# Stanford CS336 Lecture 3: Architectures
+**Source:** https://youtu.be/lVynu4bo1rY
+**Author:** Stanford CS336 Spring 2026
+**Cluster:** E
+**Slug:** cs336_architectures
+
+## 1. TL;DR
+## 2. Key Concepts  ← expect transformer architectures, attention, FFN, MoE, etc.
+## 3. Frame Analysis  ← expect dense formulas/diagrams; OCR critical
+## 4. Transcript Highlights
+## 5. Mathematical / Theoretical Content  ← dominant
+## 6. Connections  ← heavy cross-refs to CS229 + prior cluster videos
+## 7. Open Questions
+## 8. References
+```
+
+---
+
+## 5. Connections
+
+- **Forward to:** `creikey_dl_cv` (applied capstone — uses these architectures in games).
+- **Backward from:** `cs229_building_llms` (sets "why transformer architectures"), `score_dynamics_giorgini` (training dynamics), `platonic_intelligence_kumar` (representations inside the architecture).
+- **Likely rich cross-references:** `cs229_building_llms` (most direct — same LLM topic, different depth), `platonic_intelligence_kumar` (representations inside architectures).
+
+---
+
+## 6. Verification
+
+- [ ] All 7 deliverables present
+- [ ] `report.md` 1000-10000 LOC
+- [ ] Tests pass
+
+---
+
+## 7. See also
+
+- [Umbrella spec.md](../../video_analysis_campaign_20260621/spec.md)
+- [Umbrella README.md](../../video_analysis_campaign_20260621/README.md)