Subvolts take Build custom AI by turning company knowledge into training data. Feb 10, 2026 - Developers
- Learn about a new approach to building and training AI agents using specialized environments and proprietary data.
- Business
- Discover how to transform your company's institutional knowledge into powerful, custom AI models.
This episode provides a solid interview with founders discussing a novel approach to AI training and development tools.
Building the GitHub for RL Environments: Prime Intellect's Will Brown & Johannes Hagemann
Audio streams from Sequoia Capital. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Sequoia Capital's Training Data podcast features Will Brown and Johannes Hagemann of Prime Intellect discussing their Environments Hub. They explore the shift to environment-based AI development, focusing on Recursive Language Models and agentic Reinforcement Learning. Their vision is for companies to leverage institutional knowledge as training data, creating specialized AI models that outperform generic systems.
Imagine teaching AI like training a dog, but with virtual worlds instead of a park. Prime Intellect's platform lets companies create these training worlds using their own data, making AI smarter for specific jobs.
This episode offers insight into a new paradigm for AI development, moving beyond static prompts to dynamic, environment-driven training. It's for AI researchers, developers, and business leaders interested in building more capable and specialized AI systems. The discussion on agentic RL and leveraging proprietary data aligns with the frontier labs' ongoing work in creating more autonomous and context-aware AI.
- HostSequoia Capital
- GuestWill BrownPrime Intellect
- GuestJohannes HagemannPrime Intellect
- Episode Title: Building the GitHub for RL Environments: Prime Intellect's Will Brown & Johannes Hagemann
- Show: Training Data (an official AI show)
- Hosts: Sequoia Capital
- Length: 45 minutes
- Key Concept: Shift from static prompting to environment-based AI development
- Platform Mentioned: Environments Hub
Will Brown and Johannes Hagemann of Prime Intellect discuss the shift from static prompting to "environment-based" AI development, and their Environments Hub, a platform designed to democratize frontier-level training.
The conversation highlights a major shift: AI progress is moving toward Recursive Language Models that manage their own context and agentic RL that scales through trial and error. Will and Johannes describe their vision for the future in which every company will become an AI research lab. By leveraging institutional knowledge as training data, businesses can build models with decades of experience that far outperform generic, off-the-shelf systems.
Hosted by Sonya Huang, Sequoia Capital
- RL Environments
- Will Brown & Johannes Hagemann
- Virtual worlds or simulations where AI agents learn by trying things and getting feedback.
- Recursive Language Models
- Will Brown & Johannes Hagemann
- AI models that can manage and understand their own context, potentially improving reasoning.
- agentic RL
- Will Brown & Johannes Hagemann
- AI that learns through trial and error in an environment, acting more like an independent agent.
1 h 08David Silver and Lasse Espeholt of Ineffable Intelligence: A Superlearner for the Age of Experience
WithDavid Silver, Lasse Espeholt
1 h 041 h 0552 minMaking Cities Awesome: Peregrine’s Nick Noone & Ben Rudolph
WithNick Noone, Ben Rudolph
Questions people ask
- What is the shift in AI development discussed?
- The episode discusses a shift from static prompting to environment-based AI development, utilizing agentic Reinforcement Learning and Recursive Language Models.
- How can companies build better AI models?
- Companies can leverage their institutional knowledge as training data to build specialized AI models that outperform generic systems.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Sequoia Capital.