A machine-readable signal layer: how AI changes production, provenance and control in open source and adjacent domains (music, books, academia, supply chain). Each event has a date, source URLs, an A/B/C/D evidence level, and links to 10 analytical dimensions. On top sits a game-theoretic interpretation of two equilibria.
Tracks are domains where the fight takes place. The time axis is non-linear: 2013–2022 are compressed (5 events), while 2025–2026 are expanded (88). Dots are events (color = evidence level; larger outlined dots are narrative nodes with relationships). Arcs connect related events. Thick arcs are cross-domain: roughly half of the relationships cross domain boundaries. That is the point: the pattern is not local. It leaks from legal into code, from supply chain into music. Click a dot to open the card.
Arcs connect the initial event with a rollback, escalation or clarification. Click a row to open the event card.
A separate geometry: one problem branches into opposite outcomes. Law and markets start sorting the phenomenon into classes.
All signals, filterable by region, dimension, year and evidence level. Each card is one checkable fact. Machine-readable source: ai_oss_signals.jsonl.
Every event is linked to one or more of 10 dimensions. Bar length shows signal density. The core is visible (platform power, externalization, copyright) alongside a thin counter-narrative (security uplift). Click to expand.
Event density for each dimension over time. Cell color shows signal saturation. The migration of tension is visible: early years are infrastructure and copyright; later years are platform power, provenance labeling and data licensing.