Like almost any concept, the argument over free will really becomes semantic (and pedantic) when pushed to academic extremes. At a certain point it shifts to "is there a difference between free will and the apparent ability to choose what we do in any given moment?"

This scientist claims that the inability to tease any choice from the infinite variables that affect that decision means that the decision isn't ours. It is an equally valid conclusion that you don't need to know every single thing that influences you in order to have agency among those influences.

Moore's take on the Cartesian question of "how do we know we exist?" is similar. It points out that the debate actually has nothing to do with existence, but what it means to "know" something, and that "knowing," like anything, can of course be made impossible with philosophical and academic contortions (e.g., arguments like "but what if this is a simulation and there is a "great deception" that only convinces you that you exist?"). It is not that some form of knowing cannot exist, it is that people are capable of imagining fantasies in which knowing cannot exist, and Moore denies that we should let the ability to conceptualize something beyond the intended context of our language (i.e., perceived reality) pervert our ability to see and accept something concrete.

Is Moore right? Who knows, but he gets at the point that the answers to questions of free will, existence, ontology, etc. have more to do with how the questions are framed academically and philosophically than with how the same concepts actually operate in real life. It will always be possibly to frame a question (or to define the words within a question) in a way that denies the possibility of knowing or agency. But the ability to do so doesn't mean that other methods of asking or knowing are impossible.