Million Minds

What is Million Minds?

Million Minds is building a new kind of AI inspired by how brains work. Instead of huge neural nets trained on massive datasets, it copies brain tricks: sparse encodings, repeating cortical columns, sensorimotor grounding, and continuous learning. The system learns all the time, keeps only what’s needed, and runs using far less data, memory, and energy. It can work with text, images, audio, video and sensor inputs, and aims to form direct and abstract models of the world

Key features

• Emulates brain-like cortical columns to learn temporal patterns and structure.
• Learns continuously and incrementally without needing massive datasets or retraining.
• Uses sparse encodings and homeostasis to minimize memory and compute.
• Modality-agnostic design supports text, image, audio, video, and sensors.
• Builds direct, indirect and abstract world models for better generalization.
• Prototype includes sensor retina, thalamus, pose system, and transition pool.