David Eagleman: Neuroplasticity and the Livewired Brain | Lex Fridman Podcast #119

Brain Structure

David Eagleman explains that intelligence in the brain is not localized to a single spot but emerges from the interaction of various parts. He compares this to the economy of a city, where no single location can be pinpointed as the source of economic activity 1. This interconnectedness allows humans to adapt and learn in ways that simpler organisms cannot. For example, humans can absorb and integrate new information to solve problems, unlike fish, which operate on more rigid, pre-programmed behaviors 1.

   

AI Limitations

Comparing human intelligence to artificial intelligence, Eagleman highlights the limitations of current AI systems like GPT-3. While GPT-3 can generate impressive text, it lacks the nuanced understanding and adaptability of the human brain 2. He points out that human creativity involves not just remixing information but also understanding context and relevance, something AI currently cannot do 2.

   

Learning Mechanisms

Eagleman discusses the unique mechanisms of learning and memory in the human brain. He mentions Ribbo's law, which states that older memories are more stable than newer ones, a phenomenon not seen in artificial systems 3. This stability allows humans to build on past experiences in a way that AI cannot yet replicate. The adaptability of the human brain, driven by curiosity and the ability to absorb information from the environment, sets it apart from machine learning models 4.