Intelligence is closely linked to complexity, emerging from networks of adaptive agents and challenging anthropocentric assumptions.
Understanding intelligence involves dimensions of strategy, inference, and representation, with collective cultural artifacts playing a crucial role in enhancing it.
Deep dives
The Nature of Intelligence and Complexity
Intelligence is closely linked to complexity, as intelligence emerges from networks of adaptive agents. The brain, for example, is a network of neurons that collectively generate intelligent behavior. There is an ongoing debate about whether intelligence should be defined based on its structure or output. Modern approaches to understanding intelligence utilize complexity science techniques such as information theory and network neuroscience. The study of intelligence has a history of being anthropocentric and has often overlooked intelligence in non-human organisms.
The Evolution of Intelligence and Falsification of Assumptions
The historical study of intelligence has revealed that early notions of intelligence as exclusive to humans were false. Darwin's assistant, George Ramanas, defined intelligence as systems capable of adaptive flexibility, which could be found even in bacteria. The distinction between humans and the rest of the living world based on flexible, learned, recursive, and self-aware intelligence versus innate and inflexible behaviors has also been debunked. Examples such as the learning abilities of invertebrates have challenged these assumptions, highlighting the need to reevaluate the definition and understanding of intelligence.
Different Dimensions of Intelligence: Strategy, Inference, and Representation
Intelligence can be understood through three dimensions: strategy, inference, and representation. Strategy involves pursuing objectives and is found in various forms of life, including viruses. Inference encompasses the ability to reason, deduce, and induce. Representation refers to encoding and operating on information, enabling problem-solving and simplification of difficult tasks. The shift from circular orbits to ellipses in planetary motion, the use of different number systems, and the development of mathematical concepts like calculus exemplify the role of representation in making hard problems easier to understand. Collective cultural artifacts, such as language and maps, play a significant role in enhancing intelligence and should not be overlooked in the study of AI and machine learning.
With the recent release of GPT-4, now seemed like a good time for our episodes on intelligence. And not just artificial intelligence, but intelligence in general. To help us on this journey, we're joined again by David Krakauer, President and William H. Miller Professor of Complex Systems at the Santa Fe Institute.
This episode is part one of our two-part conversation with David about intelligence. In part 2, David is going to cover artificial intelligence. But in this episode, we're going back to basics and David asks, what is it that makes us intelligent?
