Epistemic Stack
Can we build a reliable supply chain for truth?
- It is harder than ever to know what’s real and what’s true. AI is making this worse — but it could be the way we make it better.
- The “epistack” is a focus area we’ve developed: at its core, a set of technologies aiming to illuminate objective epistemic relationships — every claim, human-written or AI-generated, mapped to its citations and topical context.
- AI is what makes this tractable for the first time — automating work that previously would have demanded prohibitive human labor.
When more people share a common basis for what we know — and how we know it — we can reason, create, and coordinate on firmer ground.
With the epistack in place, every claim can be assessed for what it’s worth.
- On top of the core epistemic sit assessment layers — some standard, some users can build themselves — where evidence gets weighed and discussed in the open.
- Each claim is read against its full context — rebuttals and gaps specific to the claim, plus broader context like methodology and the state of the field it sits in. Weak claims show as weak; strong ones show as strong — and in genuine disagreement, what’s actually contested becomes clear.
- When a key source is re-evaluated — a retracted study, a corrected dataset, a challenged piece of reasoning — everything that relates to it surfaces immediately, not years later.
