to the point. Even if some details are incorrect (which I think they are), it seems that the main idea MIGHT WORK. But how exactly is that possible? Well, the idea is just an attempt to exploit a well-known loophole of the Bell theorem: the SUPERDETERMINISM loophole. Namely, if everything, including our "free" decisions, is actually determined by physical laws, then, AT LEAST IN PRINCIPLE, it is possible to get Bell correlations without nonlocality. The standard Bohm interpretation is also superdeterministic, but it still does not contain sufficiently many hidden-variables to avoid nonlocality. To overcome this, Shaterland adds ADDITIONAL hidden variables, the wave functions psi_f. There is no doubt that you can avoid nonlocality by adding a sufficient number of superdeterministic hidden variables. The difficult part is to do it in a relatively simple way, and that's what Shaterland attempts to do. His attempt can be seen as a combination of transactional and Bohmian interpretation.
This approach can be compared with that of 't Hooft, who is trying to construct local superdeterministic hidden variables that, at first sight, do not even resemble QM.