The fundamental problem of modern science is the problem of innovation. Where does novelty come from? This problem shows up in physics, biology, artificial intelligence, and economics.
Within physics, the problem is how to account for the fundamental constants of reality. They are all precisely tuned to make sentient and intelligent life—life that can learn about itself and the universe—possible through science.
Within biology, the problem is accounting for the source of highly complex genetic sequences that express finely tuned biological functions.
In artificial intelligence, the challenge is identifying solutions that are relevant to a given scenario. In economics the problem is identifying the right products for the market.
What do all these situations have in common? In each case, an independently specified target is a very small portion of the space of possibilities. Think, for example, of an effective treatment for COVID-19. Many ideas might sound good but not work well. Thus, a very high level of information is required to identify a treatment strategy that really does work consistently and hit that target.
The source of this information puzzles the respective fields. The puzzle is due to the fact that, while each field can describe the target precisely, using its axioms, none of the fields can describe how the target came to be hit.
In each case, the question is not answered by underlying fields either. Physics cannot explain where the complex genetic sequences came from in biology. Biology cannot explain why people choose one product over another in a market. Reversing the hierarchy does not provide much more insight either. Biology does not explain the fundamental constants of physics and economics does not explain the functional complexity of biology.
The one area where we do not encounter this mystery is engineering. In engineering, the cause of purposeful arrangements of parts is well known. This cause is engineering innovation. Engineers create the technical inventions that run our economy. However, once we get into the engineer’s mind, the mystery reemerges.
The field of artificial intelligence, which tries to reproduce human intelligence with computers, cannot figure out how to make computers come up with the sorts of novel solutions that engineers generate regularly.
One counterargument is that brains are not like computers so we shouldn’t expect computers to be able to do what brains do. However, everything physical can, in theory, be modeled with perfect accuracy via a computer. If so, nothing the brain does is beyond the capability of a computer.
We are back where we began: Taking a step back, major fields have a consistent problem with explaining the source of information necessary to hit very small targets. In engineering, we do have an identified cause of this information, which is innovation. Yet, when we reduce the engineer’s mind to a computer, the source of innovation disappears. Because computers can simulate anything physical—yet cannot derive the source of innovation—it appears that the problem how to account for innovation cannot be solved by anything built upon the laws of physics.
You may also enjoy these articles by Eric Holloway:
Will AI or fighter pilots win the 2021 dogfight? Or both? The outcome of future warfare will be decided, not by AI alone, but by finding and optimizing the tradeoff between human and artificial intelligence.
Current artificial intelligence research is unscientific. The assumption that the human mind can be reduced to a computer program has never really been tested.