Denise Holt hosts the Spatial Web AI podcast. I asked her about a kind of AI that is modeled on how brains work, and how it relates to the sciences behind the OODA loop.
Active inference AI
Denise explains active inference AI, which draws on the same sciences as the OODA loop. She says it uses less energy than large language models, and its decisions are transparent, auditable and explainable.
The Spatial Web
Denise describes the Spatial Web as the next step in internet protocols. It gives emerging technologies a framework to work together. We look at action-perception loops for intelligent agents and how the Free Energy Principle shapes an agent’s internal and external states.
How the brain predicts
We close on perception and inference, using Adelson’s checkerboard illusion. The brain predicts the world from sensory signals, and active inference explains how it does that.
Active inference puts prediction and orientation at the center of intelligence, in people and in machines.
What would you want to audit in an AI system before you trusted its decisions?
Listen and watch:
YouTube:
Apple Podcasts:
Spotify:
No Way Out on YouTube:
https://youtube.com/@nowayoutpodcast