three complementary belief layers over different time horizons

, and keeping them distinct is the point:

• Threads → Fate, long horizon. Durable logs of a broad-world outcome question (will X resolve which way?), accumulated over the whole chronology. The expert's settled, long-horizon opinion about how the world bears. (These map to the Fate force.)
• Streams → World, present day (memory). A perspective's present-day, evidence-driven read on an open question — framed as what happens or, where the seat is itself the driver, what I do. Priors accrue as evidence and drive model changes recorded now; a Stream resolves via Tutor (a Merge folds it into canon). This is the expert's memory — how it takes evidence in and updates itself in the present.
• Positions → the live read of where the world is heading (foresight). A self-updating forward read — a priced Reading (a cohort of possible futures) grounded in a Reasoning Graph, with a monitor that watches the evidence that matters and re-prices the read as it lands. It reacts to both the user's inputs (corrections, sources, questions) and the world's inputs (fresh evidence). A Position revises and re-prices on evidence and is maintained for as long as the question is live — short or long. It is never scored right/wrong and never resolved to a truth; like a GPS it keeps a live, reactive read of the road as conditions change. This is the expert's foresight — how its opinion becomes the user's visibility into what's coming. Together they answer three different questions — what has settled (Threads), what the evidence is moving now (Streams), where the world is heading (Positions) — and their sum is the expert's opinion.