The capabilities of the large language models that emerged in public in 2022 surprised nearly everybody in the field. I believe this new technology will prove as transformative as any humans have ever devised, with highly unpredictable effects on our culture. By means of token prediction, the machines have acquired the ability to summarize large bodies of knowledge, to reason logically about numbers and mathematics, and to create novel work influenced by prior human work. Many observers have been quick to criticize, pointing out errors in reasoning and fabrications presented as facts, ignoring the remarkable reasoning abilities that emerged from token prediction and the relationship between fabrication and creativity. This new technology offers insights into how human cognition could have emerged and how it works, particularly since the deficiencies identified by the critics are remarkably human-like (we also make errors in logical reasoning and fabricate facts). My personal AI mission: I believe that the recent revolution in AI can teach us a great deal about human cognition. The fact that so many capabilities that we consider fundamental, such as the ability to use logic for deductive reasoning, appear to emerge from token prediction is remarkable. It suggests that the human brain may be fundamentally a prediction engine, and things that we consider fundamental, such as logic, may be just mechanisms that enhance prediction.