Data is everywhere. Knowledge isn't.
Every one of these sectors has more data than anyone can process. What's missing is the reconciliation — turning contradictory sources into a single record that holds up under scrutiny.
Sectors
Where knowledge is broken— and where we’re going next.
Every industry drowns in data. Almost none of it has been reconciled into knowledge anyone can trust. We start where the gap between available data and usable knowledge is widest — and where getting it right changes outcomes.
Why These Nine
Where the gap is widest.
We chose domains where abundant data coexists with a surprising lack of coherent, verified knowledge — and where closing that gap has real consequences.
The stakes are high enough to justify the work.
A wrong call in markets is a loss. In medicine, a misdiagnosis. In security, a breach. These are domains where the difference between raw data and verified knowledge is measured in consequences.
AI is about to make this urgent.
The next generation of AI agents will need structured, reliable knowledge at scale — not raw feeds they have to interpret. The infrastructure for that doesn't exist yet. We're building it.
Active Sectors
Nine domains, one method.
Markets & macro · Venture funding · Forward view
Financial
Prices, fundamentals, and venture data arrive from dozens of sources and almost never agree. We reconcile them into one clean record and publish research on where things are heading.
Explore sectorEvidence grading · Claims data · Population equity
Medical
Trials, molecular data, and insurance claims that quietly contradict each other. We reconcile the evidence, grade what replicates, and check it across the populations that were actually studied.
Explore sectorPatch diffs · Attribution · Collective defense
Cybersecurity
Ninety-nine alerts in a hundred are noise. We collapse CVE feeds, telemetry, and threat intel into reachable risk — the one vulnerability that's real, exploitable, and yours.
Explore sectorConsensus · Provenance · Narrative detection
Social Knowledge
The same event yields a dozen incompatible accounts. We tell durable consensus apart from manufactured narrative — with provenance, not opinion.
Explore sectorMeasurement · Longevity · Outcomes
Mental Health
We live longer, under more stress, and the field can barely agree on definitions. We start with the hardest question — can well-being be measured? — then build the database.
Explore sectorProcess transfer · Property space · Cataloguing
New Materials
The next alloy or battery chemistry is probably already implied in what we know. We unify a century of scattered data, map which processes transfer across materials, and revive forgotten techniques.
Explore sectorCoherence · Cross-domain · Reproducibility
Scientific Discovery
A sky survey collects more data in one night than a lab processes in a year. The frontier is limited by coherence and throughput — making the known hold together across fields.
Explore sectorLocalization · Biodiversity · Resilience
Agricultural Diversity
Farming knowledge is deeply local and rarely recorded the same way twice. Monoculture wins by being legible, not by being right. We make diversity coherent enough to compete.
Explore sectorVoting records · Polarization · Governance
Politics
What a politician said, how a bill actually moved, whether we're really more polarized. The record exists but is scattered and hard to reason over. We make governance legible.
Explore sectorExpanding
Where we’re heading next.
Domains where the demand for organized knowledge is growing fast and the supply hasn’t caught up. Each one passes the same test our first nine did.
Exploring
Legal & Regulatory
Laws and regulations span jurisdictions and contradict across them. The clean, cross-referenced legal record that compliance teams need doesn't exist yet.
Exploring
Patents & IP
Patent data is public but scattered across dozens of national offices in incompatible formats. Prior-art search at scale — the problem that actually matters — remains essentially unsolved.
Exploring
Supply Chain & Logistics
Provenance and risk data fragmented across vendors, countries, and proprietary systems. Most organizations assemble it manually, project by project.
Exploring
Energy & Climate
Emissions, grid performance, and climate models sit in agency silos across formats and timescales. Increasingly demanded by regulators and investors, still poorly served.
Exploring
Insurance & Risk
Risk models rely on actuarial data that rarely reconciles across carriers, regions, or event types. The shared view that would improve pricing for everyone doesn't exist.
Exploring
Education & Pedagogy
Curriculum research and learning outcomes siloed by country and institution. What actually works is buried in what's merely published.
Know a domain that needs this?
If your industry has abundant data but no clean, coherent record — we want to hear about it.