Use the briefing in three passes
- Scan: start with the top new and re-sighted signals.
- Interpret: check trend momentum and horizon placement.
- Act: review provisional high-priority items or export into briefs.
What changed, why it matters for intergenerational fairness, and what deserves attention next. The complete evidence base remains available in Signals.
This dashboard works best as a decision tool. Scan the newest movement, check which trends are accelerating, then jump into the review queue or ask for a quick synthesis.
0 are currently high priority and 0 have analyst confirmation.
Try: "What is new this week?", "Which trends are accelerating?", or "What needs analyst review?"
Confirmed trends are locked claims. Provisional trends are detector hits that passed the event gate. Watch domains are topic buckets.
Dated decision moments on the way to the preferred future, grouped by the principle they advance. Expand a moment to see what success and failure look like.
Use this to trace how each major trend radiates outward — on-theme hubs, first-order signal links, then second-order implications as outer rings.
Each wheel reads from the centre outward: the trend, then on-theme hub evidence, then first-order signals, then inferred second-order implications. Farther out means farther from the trend in the analysis chain, not more important. Click any node to pin its details in the panel (outer nodes show IF → THEN). Hover to preview. Green = supports, red = contradicts.
Use this when you need momentum — which storylines are gathering signals, and when attention shifted.
The stacked flow shows weekly sightings across every trend on one time axis — total height is the field's overall pulse, each coloured band is one trend's share of it. Click a band or a legend chip to jump to that trend. Below, every trend gets the same aligned weekly chart, so the shapes are directly comparable, with a growth badge for the last 30 days against the 30 before.
Strain-direction signals that belong to no tracked trend — candidate emerging issues.
Confirm, override or archive these AI-provisional high-priority signals. Easiest: run node scripts/dev-serve.mjs and open the click-through review console (sliders, one-key decisions, saved instantly). Or copy a command below for the terminal.
Use this to judge whether the platform earns trust — public, dated predictions and the human-review record, kept honest over time.
Falsifiable calls with a due date and a stated confidence, made before the outcome is known. They resolve to hit or miss on their date — no retroactive editing. This is the accountability record; calibration (are we right as often as we claim?) sharpens as they resolve.
Where a human has checked or corrected the AI's provisional scoring. Every override is recorded in the append-only review journal.
Browse by theme: pick a domain below, then choose a topic for a thematic synthesis page with composition, tensions, and watchlist. Combine with search (press /) and click a column header to sort the signal list.
Use this to compare what still holds (H1), what is being fought over (H2), and which seeds are live (H3) — three columns on one screen, or theater for a wall brief.
Sharpe’s method is a tension, not a pile of dots. Read the three columns together. Map plots every signal (height = score; dense lanes use smaller dots, high-priority on top); Read lists every title; By domain lines climate against fiscal against digital. Hover a dot to see the same domain in the other two lanes. Theater fills the screen (Esc to leave).
Use this when triaging by impact vs uncertainty — act, prepare, monitor, or leave in the background. This is a radar, not a scenario 2×2.
Scores cluster; treat quadrant as the unit of action, not the pixel.
Every signal on one impact × uncertainty map — impact rises upward, uncertainty grows to the right (lower credibility plus higher novelty), and the median splits carve four action quadrants. Bubble colour marks the dominant domain; numbered bubbles are signals sharing an identical score cell. Click a quadrant to expand it into a zoomed map with a domain breakdown and the ranked signal list; single bubbles open the signal card directly.
Use this when geography and regional precedent matter — drag to rotate, click markers for country-level signal lists.
Geographic footprint of the knowledge base. Drag to rotate, click a marker or a country in the list to see its signals. Marker size = signal count; red markers contain high-priority signals; hollow teal rings mark countries mentioned in signal text (sized by mention count). Click one to see what cites it. Signals tagged “International” are listed separately because they belong to no single country.
Use this as the advocacy calendar — clocks, then four named COP31–2028 worlds, then backcast pathways.
Four worlds to COP31 and the 2028 Pact review. Expand a cell for the GYC ask, the critic pass, and the signals that would make that world true. Dated moments sit below. Live calls sit on Finance and .
Use this for live calls, fellowships, and windows with a deadline — joined to the Finance graph for funder context.
Public scan of funding and participation windows. Amounts and funder graphs live on Finance. Decision clocks for UNGA 81 and COP31 sit on .
Use this to brief like a futurist — IE lens by horizon, typed as strengthening, weakening, wild card, or weak signal. Not a triage scatter.
Sectors are CEB/DFG lenses. Rings are H1 (inner) to H3 (outer). Colour is the Futures Platform type. Click a dot for the signal; the compact strip on Overview is the same map.
Grounded retrieval over signals, threads, moments, worlds, Pact, and funders. Citations jump. A critic pass flags what the answer does not prove.