Reinforcement learning

Jul 17, 2026

Research to Reality with Google DeepMind — Benoit Schillings, Google DeepMind, VP of Technology

Benoit Schillings, VP of Technology at Google DeepMind, explores the evolution of AI's role in software development, highlighting the transition from human-limited coding to an AI frontier where syntax generation is solved. He delves into the power of self-play for model training, the shifting economics of software engineering, and the imperative for active guardrails. Schillings also discusses the need for inductive architecture, advanced model planning, multimodal reasoning (as seen in Gemini), and the potential for AI to drive scientific breakthroughs in fields like chemistry and biology by uncovering patterns imperceptible to human bias.

Jul 17, 2026

AI Can't Learn The Way Humans Do - This Could Fix That

This discussion explores world models as a promising path to solving sample efficiency in AI and achieving AGI. It contrasts deterministic control (Newtonian physics) with stochastic environments (RL), explaining the challenges posed by large action spaces in Go, self-driving, and robotics. The episode delves into how synthetic data, video diffusion models, and latent space architectures like JEPA are making world models practical, while also highlighting remaining open problems in physics modeling, real-time adaptation, and rich sensory integration.

Jul 15, 2026

Computer-Use 2.0: Agents Just Got Multi-Cursor — Francesco Bonacci, Cua

This talk introduces cua driver, an open-source tool enabling AI agents to interact with computer GUIs in the background across macOS, Windows, and Linux by leveraging accessibility APIs. It details CUABench, a robust evaluation framework with over 130 verifiable tasks designed to benchmark and ensure the trustworthiness of computer-using agents, revealing current limitations in complex tasks like circuit design. Finally, it presents cua fleet, an infrastructure solution that optimizes GPU utilization and reduces costs for RL training of these agents by dynamically managing sandbox environments.

Jul 15, 2026

Recursive Model Improvement — Lee Robinson, Cursor, SpaceXAI

Lee Robinson of Cursor outlines a comprehensive strategy for recursive AI model improvement, centered on a two-loop training framework. He details how Cursor enhances both user-feedback-driven outer loops and high-quality evaluation inner loops, introducing novel methods like textual feedback and addressing reward hacking. The discussion extends to scaling compute infrastructure through partnerships with SpaceX, Colossus, and Terafab, and leveraging agent-based automation to streamline research and foster a future where models continuously train and improve themselves.

Jul 13, 2026

The Prime Intellect Stack — Will Brown, Prime Intellect

Deep dive into Prime Intellect's open-source ecosystem for post-training LLMs, covering the modular Verifiers V1 environment design, the asynchronous and scalable Primer RL training framework, and the Lab platform for hosted training and fine-tuning. Learn about advanced reward systems, the interception server pattern, and tokenization control with the Renderers library, all designed to enable frontier agentic model development.

Jul 07, 2026

How we taught agents to use good retrieval - Hanna Lichtenberg, Mixedbread AI

Mixedbread AI addresses the "Oracle Gap" – the disparity between LLM reasoning and retrieval capabilities – by developing agents trained to use advanced search tools effectively. They demonstrate how current LLMs generate poor queries due to training biases and introduce a sophisticated agent harness with diverse search tools and a unique training regimen, including supervised fine-tuning and reinforcement learning with custom rewards, to teach agents to form precise semantic queries. This approach significantly improves performance on benchmarks like Oblique Congress and Snowflake's Match QA, closing the gap between theoretical perfect retrieval and real-world agent performance.