What Comes After Prediction?

Pattern recognition scales. Causal reasoning doesn't... yet. Most AI is extraordinarily capable until the moment it isn't: when the situation is novel, when uncertainty is real, when the question stops being what usually happens and starts being what happens if we do this. That's not an edge case. That's where consequential decisions live.
‍
Correlation is abundant. Causation is hard. The gap between them is where Anthos works. We're a frontier AI lab building systems grounded in structural causal models, counterfactual inference, and decision theory — fields built specifically for reasoning through change, under conditions no training set anticipated. Not predicting the most likely future. Modeling what shifts, what cascades, and what breaks.
‍
Founded by applied mathematician DJ Passey PhD and serial entrepreneur Devin Young, our team brings together folks from Google X, Citi Bank and numerous VC-backed startups — people who've spent years inside institutions where decisions are hard and the cost of getting them wrong is real. We're building the tools that don't yet exist, with a small number of partners close enough to the work to help us sharpen them.

Want to know more? Reach out to hello@anthos.systems.