I still maintain that the feedback loop demonstrated in Bret Victor’s videos is _significantly_ quicker than anything you can achieve with any TeX engine in its current form.
Think of the part where he draws the fish. Or where he drags a slider. You need update rates of a few miliseconds, tops.
What we can achieve with TeX is in the order of seconds, at best — for simple documents.
I am definitely not convinced this lives up to Bret’s standards :-).
However, Bret's stuff doesn't exist, and there's a reason for that: it's really hard. What these guys are doing is trying to make something that actually works, and it looks like they're doing a pretty decent job of it. The fact that it integrates with all the normal tools (in particular, LaTeX) makes me inclined to overlook it's imperfections.
Additionally, I think that's actually the right way to do it. The result should be optimized for viewing, not for editing, and the editing environment should be optimized for editing, not for viewing. If the two can be merged into a tight feedback loop, then maybe we can get something of the best of both worlds.