This pointer led me to track down a 1993 essay of his, Virtual Environments and the Emergence of Synthetic Reason, which seems to make arguments on similar subjects, and might (?) provide some sort of a preview.

The arguments in that essay, at least, remind me of some of the dynamicist positions in artificial intelligence, which similarly focus on population- and system-level properties like attractors and stability as the right level of abstraction for talking about complex systems. They tend to make a similar anti-reductionist argument, though out of more practical concerns, that a focus on the literal operation of systems at their code or symbol level isn't very useful for understanding how they operate, because the many of the interesting macro properties of the system as a whole are emergent effects not visible (or at least heavily obscured) at that too-far-down level of description.

I agree with Mark. From the video it sounds like he's recapitulating the standard alife critique of AI, with a focus on dynamical systems approaches (e.g. people like Stafford Beer). In the philosophical literature, this would continue in the tradition of autopoesis, and the work of thinkers such as Maturana and Varela (e.g. Autopoiesis and Cognition).

Sounds like more asocial alife'rs who imagine they can capture the sublime features of intelligence with barely legible arguments amongst themselves. Pfft. Son of Kurzweil. They are "2012", BTW. And Skynet. That too.