Can AI Learn to Think About Money? - with Bayan Bruss (Capital One)

Bayan Bruss, who leads Capital One's AI Foundations team, discusses building machine-learning systems that perceive financial ecosystems, form beliefs about the future and act on them. The conversation explains why language models are a poor fit for ingesting thousands of card transactions, why synthetic data works well for financial time series, and how end-to-end learning conflicts with regulatory demands to know exactly what a model has learned. It also compares reasoning in language with reasoning in latent space for high-stakes decisions such as long-horizon credit risk.