ingestion log
2026-04-10 — initial ingestion
sources processed: 137 files across 3 origins
- ideaflow: 110 files (tagged #projectidea, #newprojectidea, #bigprojectidea)
- apple-notes: 20 files (Archived/build folder + archived root)
- obsidian: 7 files (home/ideas/ folder)
pages created: 74 idea pages, 8 cluster pages, 1 index
ideas by status:
- built: 6 (OnCue, Pause, convo-flow, connection hub, precision description engine, keystroke music)
- explored: 10 (brain rewinder, sensor capturer, smell resetter, universal habits, axon, learning suite, discord connections mapper, referral hiring, always-on AI assistant, universal agentic searcher)
- raw: 58
merges performed:
- brain rewinder: ideaflow + apple-notes potential-projects
- sensor capturer: ideaflow + apple-notes sensor-capturer + build-notes
- universal habits: 6 ideaflow notes + apple-notes potential-projects
- learning suite: 3 ideaflow notes + apple-notes potential-projects
- discord connections mapper: 4 ideaflow notes (linked notes chain)
- universal data capturer: ideaflow + apple-notes potential-projects
- oncue: apple-notes oncue + tellwell
- wifi client: apple-notes + ideaflow
- several others merged with apple-notes potential-projects-archived (which contained strikethrough copies of ideaflow ideas)
clusters identified:
- memory and context tools (9 ideas)
- social network and connections (7 ideas)
- learning and education (10 ideas)
- habits and productivity (8 ideas)
- AI tooling and research (10 ideas)
- hardware and wearables (8 ideas)
- hiring and work tools (7 ideas)
- search and discovery (6 ideas)
obsidian sources:
- song.md: merged into keystroke-music (AI songwriting interest)
- motivation.md: standalone page (motivation-education)
- disconnection_between_generations.md: standalone page (generational-disconnect)
- game_ideas.md: merged into task-optimization-game
- domain_names.md: not ingested (list of domain name ideas, no project concept)
- phone_shortcuts.md: not ingested (personal shortcuts, not a project idea)
- WRC_stuff.md: not ingested (website maintenance tasks, not an idea)
notes:
- apple-notes potential-projects-archived.md contained strikethrough copies of many ideaflow notes, confirming those ideas were moved from apple notes to ideaflow
- many ideaflow notes reference linked notes by ID (e.g., [linked note: JMXS1vOOYr]) — these were matched to other ideaflow notes by ideaflow_id and merged accordingly
2026-04-10 — google sheets ideas ingestion
source processed: sources/google-sheets-ideas.md (645 lines, 4 tabs)
- project_ideas: ~85 ideas with originality/excitement scores, MVP estimates, competitive analysis
- ranked_by_tech_depth: same ideas ranked by technical depth with tier labels (DO THIS, STRONG, WRONG FIT, ARCHIVE)
- Blue Ocean Analysis: 12 ideas evaluated for market positioning
- best: harrison's curated top picks
summary created: wiki/summaries/google-sheets-ideas.md
existing pages enriched with spreadsheet data: 38 pages
- added originality scores, excitement scores, MVP time estimates, competitive landscape analysis, tech depth rankings, tier labels, and blue ocean positioning
- data added as "spreadsheet evaluation" blocks in compiled truth sections + timeline entries
new pages created: 16 ideas not present in other sources
- acoustic-drone-detection (DO THIS, tech depth 9/10, EMERGING)
- spec-driven-dev (DO THIS, tech depth 8/10, RED OCEAN)
- ui-flow (8/10 originality, 9/10 excitement)
- hard-docs-writer (STRONG, tech depth 7/10)
- eeg-artifact-rejection (NEAR BLUE, high feasibility)
- symbolic-regression (EMERGING, high feasibility)
- context-window-optimizer (EMERGING, medium feasibility)
- cognitive-foom (STRONG, tech depth 9/10)
- predictive-maintenance-sensors (STRONG, tech depth 8/10)
- ppg-biomarker-wearable (best picks list)
- embedding-tone-interpolation (best picks list)
- robotic-arm-assistant (WRONG FIT, needs scoping)
- conversations-recorded (content/community concept)
- culture-fingerprint (visualization concept)
- math-dopamine (WRONG FIT, content/product focus)
- nueva-third-space (community design concept)
key findings from spreadsheet:
- top 5 [DO THIS] tier: acoustic drone detection, BCI/read-imprinting device, agent-based simulation, spec-driven dev kit, personalized medicine
- highest excitement (10/10): brain rewinder, articulate/precision description engine
- only true BLUE OCEAN: N-of-1 causal inference engine (personalized medicine variant)
- NEAR BLUE with high feasibility: EEG artifact rejection, content quality scorer
- best picks emphasize: audio classification on MCUs, smell necklace, data selling, embedding interpolation, computer optimizer, PPG sensors