While incredibly cool and another example of the awesome potential of data driven driven systems, this one would be a bit more difficult to scale due to the bounds limitations (which also is true of Ubisoft's data-driven physics
cloth btw, but less of a limiting factor there). I'd love to see it happen though.
Alternatively I see Ubisoft's data driven cloth sim as having a clearer path to viability in AAA games. It could help address 4 large pitfalls of current runtime cloth solutions: performance, quality, scalability, and simulation pop. Also lines up well with Ubisoft's technical challenges with Assassins Creed. Not hard to imagine a push internally to bring this one out of the prototype stage and apply it to production.
Anyways, very exciting times ahead. Raytracing may be the shiny new toy right now, but there are other equally impressive ways games could evolve in the near future.