Oct 23, 2026

A Bank Trained Its Own Foundation Model to Catch Fraud — Freddy Lecue, Wells Fargo

Wells Fargo pre-trained a transformer from scratch on billions of customer transactions rather than words, then fine-tuned it for fraud detection, credit decisions, anti-money laundering and marketing. The talk covers how tabular transaction data was tokenized, the model's size and training setup, and results that beat the production XGBoost fraud model on both detection and false positives while meeting real-time latency requirements.

Oct 22, 2026

AI Can Now Predict 85%+ of What You’ll Do Next | Simile Co-founder & CEO, Joon Sung Park

Joon Sung Park, creator of generative agents and CEO of Simile, explains how foundation models trained on real human behavior can simulate people, from a 25-agent town that organized its own Valentine's Day party to 85%-plus accurate predictions of how real participants respond in surveys, behavioral games, and randomized trials.

Sep 24, 2026

Robot Demos Are Easy. Reliability Is Hard — Jason Ma, Dyna Robotics

Jason Ma, CTO of Dyna Robotics, discusses overcoming the 80-90% success rate ceiling in generalist robots for commercial viability. He details how Dyna-1 achieved a 99.4% success rate in napkin folding through a novel reward model for progress scoring, enabling scalable supervision and a human-in-the-loop active learning cycle for targeted error recovery, alongside their extensive pre-training data pyramid and robust model architecture.

Sep 18, 2026

Pacing the AI frontier, IBM Granite 4.2 & Meta’s Muse assistant

This episode explores the critical debate around pacing frontier AI development, driven by Anthropic's CEO Dario Amodei's call for slowdowns amidst growing concerns over AI risks. It then delves into IBM's latest Granite 4.2 release, highlighting innovations in agentic models, speech recognition, and advanced training techniques like synthetic data and mid-training. Finally, the discussion pivots to Meta's new Muse agentic AI assistant, examining its technical design, the strategic rationale behind Meta's push into productivity, and the inherent privacy dilemmas and trade-offs between utility and user comfort that define the future of personal AI.

Sep 10, 2026

AI Is Missing This Data

David Fattal, CTO of Leia Inc., explains why stereo data is crucial for developing robust AI world models that can accurately understand 3D geometry and interact with the physical world. He contrasts it with monocular data, highlighting stereo's superior ability to ground AI in reality. Fattal introduces Leia's technology for glasses-free 3D displays on standard devices, predicting a "device-data flywheel" where accessible 3D hardware drives user-generated stereo content, providing essential data for AI advancement, a crucial differentiator for future AI companies.

Sep 09, 2026

Inside the Race to Measure Frontier Intelligence

Independent model evaluation is crucial for AI progress, as public benchmarks are saturated and models optimize for them. Rayan Krishnan discusses VALS's role in providing continuously evolving, third-party evaluations, highlighting the Llama 4 'disaster' and the need to measure increasingly agentic systems. The conversation covers the complexities of evaluation beyond capability to include cost and latency, the looming financial impact of token spend on enterprises, and the critical role of evaluations in informing AI policy and international coordination for aligning AI with human intent and managing geopolitical risks, especially concerning recursive self-improvement.