Peter Elias

Peter Elias is the founder of Probably, which raised $9 million in seed funding from Andreessen Horowitz, TechCrunch reported in June 2026. His goal is to keep hallucinations and simple factual errors from ever reaching the user, and to reach the kind of 99.99% accuracy common in deterministic systems.

Model Guesswork at Probably

Elias describes Probably's harness as a "data science mech suit." The model's first-pass answers are checked against a deterministic validator, which bounces back any results that don't match the dataset.

"What we learned building this was that the better your harness engineering is, the weaker the model can be," Elias told TechCrunch. He said the current version runs on a model "four classes weaker than the frontier models," small enough for a desktop computer.

"I think it's really interesting that the big AI labs have not even attempted to do this," Elias said. "They're incentivized not to, because they make money the more times you have to correct the model."

Starting with Data Analysts

Probably's first product is a data science tool that produces quick answers from complex datasets, with a citation and an audit trail for each result. Elias says the same engine can extend to accounting or medical services, or "any precision-sensitive use case."