Machine versus organoid

Garry Kasparov was 33 years old when he faced Deep Blue, IBM’s chess-playing supercomputer. Willis Gibson is 13 years old and he just beat Tetris. The difference is that Kasparov lost and Gibson won, but those small details aside, we must ask ourselves, what does that tell us? That Tetris is easier than chess? That the human race is evolving faster than machines? That games are diverting our attention from the real issues as to what constitutes intelligence? Big questions.

To begin with, Tetris is no easier than chess, quite the opposite. Chess is a defined game in which the knight moves in an L shape, the bishop diagonally, and the castle in its specific fashion and no other. To each of your moves, your opponent has a limited number of responses. As a defined game, it is relatively easy to compute. Tetris, however, is a mathematical horror. It belongs to a group of particularly tricky problems (NP problems, in the jargon), where every tentative answer is easy to confirm or discard, and yet there is no systematic way to solve it in a reasonable amount of time. A good metaphor is a jigsaw puzzle, which takes a lot of time to solve, but only a second to confirm.

In this day and age, it is comforting to say that artificial intelligence is not here to replace us, but rather to help us. But the truth is, the war between living and silicon brains has been raging since the days of Kasparov and Deep Blue, and has only intensified over the last few years. Scientists at Deep Mind, a London firm that was acquired by Google years ago, have created systems that beat humans at chess, the Chinese game of Go and, even more challenging, at poker, where the cards in your hand count less than the cards your opponent thinks you are holding. Artificial intelligence has created high-level strategies over the years that centuries of grandmasters never imagined. ChatGPT is the only the best-known of an entire generation of systems that are turning knowledge upside-down.

There’s another way in which machines and humans are locked in a race, and it comes from the world of neuroscience: organoids, human organ-like bodies made from stem or precursor cells. The latest is a cerebroid, or mini-brain (about the size of a grain of rice) created from fetal cells. They are complex and brain-like enough to study neurological diseases from them. But we know that human neurons spontaneously connect to form circuits, and circuits can do computations. There is interesting research on hybrids of neural tissue and digital circuits that can do some computations. Science fiction authors are going to have a field day.

Nor is it true that humans are evolving faster than machines. The Gibson kid is the first human to beat Tetris, but AI systems had already beaten him to it. Since there will be no moratorium on AI, perhaps we need to speed up neurology.

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition