Hudson Labs CEO Kris Bennatti tested the top models on real filings and found the wrong number about 30% of the time. Her answer: a no-hallucination guarantee no one else in financial AI will make.

Two daily power users still won't say the models got meaningfully better in six months. Then Fable shipped. Brett and Khe run the vibe check: evals, the Bridgewater judgment layer, and why skepticism is up.

Joe O'Donnell ran the short book at Tiger Global for nearly a decade. His claim as Canary's CEO: so many investors are misusing AI that there's more alpha available today than there has been in a long time.

Matt Stockton has spent 20 years architecting data infrastructure for firms deploying AI. His read: the model is rarely the hard part. Writing down your own process is 80% of the work, and most firms skip it.

Raj Shah co-founded Stoic Point after a partnership at Light Street. He runs a two-person fund resourced like a firm ten times its size, and says he's more worried about AI replacing him than the juniors everyone else is fretting over.

Ying Hua ran money at Citadel and Balyasny before founding Implied. Her case: the foundation labs won't eat finance, because domain knowledge is the piece they won't build. Brett and Khe get into data, dashboards, and what stays human.

Brett Caughran and Khe Hy on what actually changed in 2026: the shift from chatbots to agent workflows, where AI helps in investment research today, and the limits that are still real.