Okay, look, suppose for the sake of argument that you're right about this. I don't think you are, but I don't need you not to be to make this point.
Now, I know you're arguing against considering game mechanics too much, so I'll try and respect that. But can you at least concede that CP are purely a game-mechanical abstraction? That is, there is no in-universe thing that behaves like/explains/justifies CP. We good on that?
Okay, so CP are a game mechanical abstraction. Then how should CP costs be determined? Presumably with the goal of creating good gameplay, since their lack of correspondence to reality prohibits assignments based in realism. I would argue that the best gameplay is obtained by maximizing the number of viable options present, which would suggest that more-effective options should be assigned higher prices for game balance.
This suggests that if experts really are as important as you claim, then they will be hysterically expensive in terms of CP. You follow? My argument is then that such a massive expenditure would be inefficient because the benefits are too spread out, and we should instead focus on specializing in a particular direction, thereby getting more net gain in that direction at the expense of others.
From here you could argue the relative benefits of specialization versus a generalist approach, of course, but thus far I have not seen you doing so.
.
