A Computational View of Life and Intelligence

This discussion will offer a novel, unified perspective on both, as described in Blaise Agüera y Arcas’ new book, What Is Intelligence?: Lessons from AI About Evolution, Computing, and Minds (MIT Press and Antikythera, September 2025). He'll begin with the striking idea first made by John von Neumann around 1950: that general computation is at the heart of any living system. 

Speaker: Blaise Agüera y Arcas, VP/Fellow, CTO of Technology & Society at Google 

About the talk: Experiments in artificial life (ALife) have now allowed us to observe how life can emerge from randomness (abiogenesis), and suggest that “symbiogenesis” can play a central role in evolution. When computational entities enter into symbiotic relationships, they must model each other, and when they merge, “computational parallelism” enables them to become both more complex to model and more powerful as modelers.

Recent research from our group on multi-agent reinforcement learning sheds additional light on how this form of modeling results in greater cooperation, as opposed to the selfish dynamics of classical game theory. These cooperative dynamics can spark “intelligence explosions,” which are familiar to evolutionary neurobiologists who have explored the relationships between troop size and brain scaling among highly social animals. 

In this framework, the evolution of life, the major evolutionary transitions (including the emergence of nervous systems and brains), intelligence explosions among highly intelligent species, and even the development of advanced technologies like AI, can be understood as stages in an ongoing process: computational symbiogenesis.