I think that's a very premature conclusion to draw (though I know you're just repeating what the video said). As with the last Potato Masher video I saw, there are multiple errors and omissions of fact. Whether these issues are due to bias, ineptitude, or a lack of care I of course can't say. Carelessness is evident in the fact that he mislabels a piece of footage early on, but it doesn't seem sufficient to explain all the problems.
For example, the video claims there's no AO on the PS4 version, which is incorrect. It also goes unremarked that, while there are fewer reflections on PS4, the ones that are present are higher quality. Also unmentioned is the greater draw distance on PS4. Further, the video claims that reflections on cars are missing, but itself includes footage that shows this to be untrue. In addition, JERM does explain that the PM is using temporal filtering, but evidently doesn't understand (or simply doesn't say) that this means it's running below 1080p. Finally, the onscreen fps counter during some PM sections shows that the game is occasionally dropping frames while standing still or walking, and dipping as low as 27fps during regular driving. Digital Foundry found the PS4 to be locked 30fps during these activities.
So the actual comparison is something like this:
Potato Masher - ~763p +reconstruct, framerate 27-30fps; higher-resolution shadows, a few more reflections, selectable sharpening
PS4 - 1080p, framerate locked 30fps; higher-resolution textures, better reflections, higher draw distance
Interpretations may vary, but to me that looks more like a PS4 advantage rather than a draw, much less a Potato Masher advantage.
The PM Pro video shows a much clearer gap over the PS4 Pro, with actually higher resolution, all settings as good or better, and an identical performance profile. There is an error in the narration, but it's a common one--JERM does not understand what CBR is--and it's not material to the conclusions.