And yet... Check out the bodies displayed at 2:08. A good range, but nowhere near the range in effect for real people - and at least half of them look noticeably wrong to me.
The cloth displays aggressive horizontal wrinkling, in a way that makes no sense. Check out the center-most figure, whose shirt curves in below the stomach with extensive wrinkles that gravity should remove. All of the figures display confusingly broken folding on the horizontal plane - the leftmost model stretches the cloth across the chest, then displays bizarre smooth-to-wavy patterns at stomach height as you move from front to side. And as far as I can see, all of this is weak to multi-layered, stiff, or tailored cloth.
So yeah, I appreciate this, but I think it's woefully inadequate to modeling even cheap, popular clothing. That's going to take a few more breakthrough as large as this one at least, so Magic Leap's VR abilities aren't nearly enough to justify this business model.
Tends not to be a linear progression at all; if the attention of more than a couple researchers is applied to any of these things then they grow very quickly - couple this with improvements in parallel processing power and it becomes quite viable. (see: physically based lighting, rigid body -> full crumple simulation, fluid dynamics, Photoshop's "smart fill", or really any game tech)
Apropos of nothing: Same applies in terms of hardware - what we can do with light field cameras now was science fiction just over a decade ago, but within a couple years of crossing a research tipping point we went from 10x10' camera arrays to <$100 hand-held commercial products.
source: graphics tech and SIGGRAPH nerd for the last 23 years
actually there was an extremely slow build up of this tech, with a sudden and rapid escalation - but between 1908 and 2010 was a very shallow incline :)
https://www.youtube.com/watch?v=RaWeWtG0tJQ