Machine
Foresight*
State T₀ Futures T₁ → Tₙ
A continuously trained world model that turns the present into calibrated possible futures — so institutions can test a consequential move before reality tests it for them.
* The future, as an input.
Models learned to answer
before they learned
to anticipate.
Language models predict what comes next in text. Consequential decisions depend on what comes next in the world.
Fluency is not a stable belief about the future. A decision needs more than a good answer: possible outcomes, calibrated odds, connected consequences, and a view that changes when the evidence changes.
This is not a claim that models cannot forecast. They can produce predictions. The deficiency is that nothing in their default form maintains a persistent, calibrated, continuously tested view of what happens next.
The next token is not the next state.
For a consequential decision,
the answer is not the output.
The calibrated future is.
The next frontier is not a more articulate model. It is a system that holds an explicit belief about what happens next, keeps it current as evidence arrives, and is scored when reality decides — so the belief can be trusted before it is acted on.
Know the state. Run it forward.
Learn from reality.
Latent inside. Explicit outside.
Built for decisions that become
expensive after you commit.
Made before
the answer was known.
A foresight report is usually written by the party it benefits. The scenarios are picked after the conclusion, and nobody goes back to check. It cannot be audited, so it cannot be wrong.
Kaldun commits instead. A forecast is fixed with its date, what the Engine could see, and the rule that will decide it. Selected forecasts are cryptographically committed before the outcome is known, so nothing can be rewritten afterwards.
Integrity by commitment. Accuracy by resolution.
Cryptography proves the record is intact. Only the score proves it was right.
Build the systems
that learn from time.
A fellowship for people obsessed with prediction, uncertainty, and what becomes possible when reality supplies the training signal.
Do not write about the future.
Build systems that survive it.