Computer Says Maybe Nodestar: Turning Networks into Knowledge w/ Andrew Trask
Sep 12, 2025
Andrew Trask, Executive Director of the OpenMind Foundation, specializes in pioneering decentralized AI solutions. He envisions a world of 'broad listening,' where understanding emerges from abundant information without corporate gatekeeping. Addressing challenges like information overload and privacy, Andrew discusses creating open-source protocols that empower data owners and enhance transparency. He delves into the evolution of communication, emphasizing the need for secure, collaborative frameworks to navigate the complexities of truth in today's digital landscape.
AI Snips
Chapters
Books
Transcript
Episode notes
Path From NLP To OpenMined
- Andrew traced his path from early neural-net curiosity to work on private-data AI and homomorphic encryption.
- He founded OpenMined to bring cryptography, deep learning, and distributed systems communities together.
Attribution-Based Control
- Attribution-based control lets data holders decide who benefits from their information without necessarily sharing raw data.
- That decentralizes power by connecting data owners directly to those who need predictions.
Broadcasting Vs Broad Listening
- Broadcasting scales one-to-many communication while broad listening is synthesizing many perspectives into a better model.
- Broad listening remains far behind broadcasting and is central to current power concentration problems.
