Maybe I read it too quickly, but it doesn’t look any different to just using parallelizing in the individual game engines (AI, Physics and Graphics). Which should work a lot better considering the possibility of dedicated hardware for at least Physics and Graphics, requiring a more nuanced approach for best results.

Agree with aalpert, it doesn’t look as new idea. Most of big game projects are multi-thread.

Might still be an interesting read for some. Thanks for sharing, Gaston! :)

Probably they are talking about Havok…