From: lexfridman
The Thousand Brains Theory of Intelligence is a groundbreaking concept introduced by neuroscientist Jeff Hawkins in his book “A Thousand Brains: A New Theory of Intelligence.” The theory proposes a novel perspective on how the human neocortex functions to create intelligence.
Core Concepts
The Thousand Brains Theory
The central premise of the theory is that intelligence is not a unified process but rather consists of multiple parallel processing units running simultaneously, which Hawkins defines as “cortical columns.” According to Hawkins, each column in the neocortex operates as an independent model of the world to support perception, movement, and other functions. There are approximately 150,000 such columns in the human brain, each working as an individual system but collaborating through a voting mechanism [00:10:01].
Independent Modeling Systems
Hawkins describes these cortical columns as independent modeling systems that converge by communicating through long-range connections, referred to as voting neurons. These connections allow various parts of the brain to reach a consensus on perceptions and reactions to stimuli [00:12:06]. The individual columns validate each other’s conclusions, leading to what we consciously perceive as a single coherent view of reality.
Intelligence and Learning
Definition of Intelligence
In Hawkins’ view, intelligence is the ability to create a model of the world within the brain that can predict and interact using this model. These internal models are not static; they continuously update based on new sensory inputs and experiences. Hawkins emphasizes the critical importance of movement in learning and forming this model, pointing out that these systems learn concepts like the structure and functionality of objects through interactions, including touch and manipulation [00:15:15].
Predictive Models and Reference Frames
A key element of the Thousand Brains Theory is the reliance on predictive models and reference frames. The brain uses reference frames to predict future events based on past interactions and sensory information. Predictive capabilities are deeply intertwined with how these reference frames are used to gauge and respond to ever-changing external environments. Hawkins argues that predictions are not just an aspect of intelligence but an inherent property of these internal models that aid in constantly updating them [00:18:14].
Evolutionary Perspective
Evolution of Intelligence
Hawkins ties the evolution of intelligence to early life forms that initiated movement and developed simple navigation capabilities. Over time, this evolved into the complex neocortex structures we have today, enhancing the ability to model the world at an advanced level [00:24:22]. He draws a parallel between animal navigation systems and the sophisticated modeling systems present in humans, suggesting that when these mechanisms became generalized, they evolved into what we now understand as the capabilities of the neocortex.
Implications and Future of AI
Application to Artificial Intelligence
The theory has significant implications for developing artificial intelligence (AI), suggesting that creating AI requires understanding and replicating the modeling capabilities of the neocortex. Hawkins envisions AI systems that mirror these thousands of parallel processes to approach or even exceed the flexibility and learning capacity of human intelligence [00:56:07].
Conclusion
The Thousand Brains Theory of Intelligence offers a revolutionary way of understanding human intelligence by suggesting that it arises from the collective work of thousands of small, independently functioning yet interlinked modeling systems within the neocortex. This perspective not only deepens our understanding of the brain but also provides a blueprint for developing future AI systems that could emulate or exceed human capabilities.