The race toward artificial general intelligence is often presented as a contest over who has the most compute or the strongest benchmark scores. Beneath those measurements sits a less settled question: what do the participants believe intelligence actually is?
In The Six Thrones: A Philosophical Map of the AGI Race, I read the major AI laboratories as competing schools of thought. The essay examines the tension between expecting intelligence to emerge from scale and building structures that let a system model, anticipate, and interact with the world.
Within that map, OpenAI represents an emphasis on emergence, Anthropic on explicit behavioral principles, DeepMind on learning through environments and simulation, and xAI on confrontation with a turbulent world. Meta’s competing research and industrial ambitions reveal tensions within a single organization. The discussion of Chinese AI strategies extends the view toward intelligence as infrastructure.
These are interpretive lenses for the landscape described in the article, not fixed identities for every researcher or project. The essay makes room for overlap, especially where consumer products and industrial systems meet.
What interests me is how choices about training, control, and deployment express different expectations of the future. Before asking which lab will reach AGI first, the full article asks whether they are imagining the same destination, and what those imagined destinations tell us about ourselves.