note that his talking about the junction, where the P and N types meet, not the whole thing. So around the point where they meet, positive holes from the P will move to the N, and electrons from the N to the P, to fill up the gaps.
However the majority of the N type silicon remains negatively charged, and the P type silicon, positively charged.
This has the effect of creating a potential difference across the junction, so when an electron is freed by light from the N type, it can't go straight to the P type through the junction, since at the junction, the P type is negatively charged.
so to get to the P type, the electrons have to go through the external circuit to get to the P type. Then the electron is able to return back to the N type, through the junction.
I think that's what I found the most difficult thing understanding.