@BestSell - All of your basic points are valid. In fact, I don't think we're really in fundamental disagreement about anything here.
However, falling upon the tendencies of present-day GMs is not as strong an argument as you might suppose.
For example, it is well-known that the Advance Variation of the Caro-Kann took a hit due to Nimzowitsch's loss to Capablanca at New York in 1927. In that game, Nimzowitch started with 1. e4 c6 2. d4 d5 3. e5 Bf5 4. Bd3 Bxd3 5. Qxd3 and subsequently lost, quite badly. However, in that game he made a number of positional mistakes which actually go against his own rules! Among them are 6. Nc3 (6. c3 protects the base of the pawn chain), 8. dxc5 (capturing away from the center, which My System tells you not to do), and 11. Nxc5 (trading a knight that moved 3 times for a bishop which moved once), all of which combine to cost white his initiative. Meanwhile, Capablanca made a number of moves which My System would immediately suggest, including 7. ..c5 (attacking the base of the pawn chain), 13. ..Nf5 (blockade), 32. Kg6 (blockade), and 41. ..Qc1 (infiltration of the back ranks on the open file).
This is what I mean when I say Nimzowitsch didn't follow his own rules as faithfully as Petrosian did. There's nothing wrong with White's position after 5. Qxd3, the problems come after that. Capa is a god to many chess players to this day (and not without good reason), but the errors of his opponent in this case are quite obvious. Modern GMs use an artificially narrow range of openings, and often treat single innovations and results as being more definitive than they really are. This is simply to keep the preparation time reasonable - if what they use produces satisfactory results, they feel no need to explore other openings.
Going into the games of past GMs and knowing where to deviate is difficult even for the very best players. So, as a matter of course, modern GMs don't even make a serious attempt to do so, even when the room for improvement is clear. If they don't understand a given position, they simply avoid the position altogether.
Modern GM play is shaped as much by who is and has been World Champion as it is by how well they actually play. The stark stylistic differences between them (clear to any player of even average experience) show just how many ways there are to play chess effectively.
There are many good reasons to give up tempo, just as there are many good reasons to give up other advantages in order to gain tempo.
yep