Laplace Research

Foundation models for calibrated forecasting

A forecast is only useful if you know how much to trust it. Models now produce predictions at scale, and almost none of them arrive with confidence you can take at face value. Laplace is our attempt at closing that gap.

Two problems stand in the way, and we work on both:

• Aggregating resolved forecasting data across markets, history, and expert judgment. Resolution is the only signal that says whether a stated probability was honest.
• Understanding the internal mechanisms that produce calibrated belief, so that calibration becomes something we can build toward rather than something we notice afterward.

The two tighten each other. Better data sharpens calibration, and understanding why a model is calibrated tells us which data to collect next.

If this works, the result is not simply better forecasts. It is that judgment can be spent on choosing which questions matter, rather than on second-guessing the answers.

We’re a small team working out of the Bay Area. If you’re working on forecasting, calibration, or interpretability, reach out.

• Laplace Research on X
• Bruce Nshuti Hirwa
• Arya Somu