Ah, in the paper he says that he's not actually talking about free will. That makes it less confusing. But I still think he's making much too strong a connection between Knightian unpredictability and free will. I don't see any reason at all to think that the former is a necessary condition for the latter, and he doesn't really attempt to argue the point. (He promises that he's going to argue it in section 2.5, but then section 2.5 ends up only talking about problems of personal identity.)
https://arxiv.org/abs/1306.0159
Scott is making a pair of claims: If it's possible to build a fast, noninvasive, perfect predictor then what more can science say to show that people don't have free will?
If it's impossible to build such a predictor, for some fundamental reason, then what more can science say to show that free will exists.