Algoritmo XY

ALT This philosophy rests on the mistaken assumption that machine intelligence and human intelligence are fundamentally similar. Humans are “one-shot” learners. We form hypotheses from a few examples, simulate possibilities in our minds, and refine our understanding through a social process of trial and error. Hence, children learn language by imitating a few words, generalizing, and adjusting their speech based on how others respond. We are not very good at absorbing massive volumes of information or sifting through unstructured data for relevant patterns. By contrast, AI models thrive on enormous training sets and excel at pattern recognition at scale, but they have yet to demonstrate genuine creativity. They have no experience of real-world embodiment, nor any capacity for trial-and-error learning through interactions with the physical and social world (except in a limited way when there are clear rewards for reinforcement learning in specific domains).

ALT When two things are different, you should not—and typically cannot—use one to mimic the other. The results would be suboptimal. It would have been a colossal mistake if Phil Jackson, the legendary coach of the Chicago Bulls in the 1990s, had pushed Michael Jordan to mimic everything that Scottie Pippen and Dennis Rodman were doing. The team went from championship to championship precisely because these players worked together and complemented each other’s skills. The same applies to AI and human skills. Using AI to do what humans cannot do, so that humans can expand what they do, is more productive than mimicry. In a future scenario where AI increases, rather than displaces, human capabilities, electricians would be aided by AI diagnostics, nurses would consult AI in interpreting symptoms, and teachers might use AI to personalize instruction for each student.