But yeah something like this would be truly interesting. Risky though. They are just building prettier versions of the same game and it sells super well.
That's a bit unfair in my opinion. From ps360 to ps4one, Rockstar did balance well a bit of improvement on every area. Next gen got pretier pixels + higher population density + more variety of pedestrians and vehicles + longer view distance + more physics + more songs on radio + more graphical effects + better textures + addition of First Perso + etc...
GTA is not a single one of these elements, its all of them together, and R* knows this better than anyone, and they did split their efforts acordinly.
Had they kept graphics equal and only turned population/car density to 11, there'd be people saying "They are just building busier versions of the same game and it sells super well" in fact, they might not have said it was selling super well, because maybe the game would sell a little less had they gone that rout.
Puting it as if its a lazy profit thing is very short sighted too. Increasing density and leaving graphics untouched (which includes new content: reworked textures, models, new shaders...) is much cheaper development wise. -Get that variable that defines the amount of character to be sapawned on this area at this time of day, change it to 10x its current value. Done.
Finaly, compare GTA to Saints Row for exemple, and you'll see that R* isn't taking, by any means, the lazy route. GTAV could have come a year or more, earlier, with half the map size, a bunch of repeated generic buildings all over the town (each building is unique and inspired on real LA buildings
with remarkable atention to detail), no three character dynamic, a fraction of
the peripheral content that one in every 100 players ever ends up encountering, and the thing would have sold just as well uppon release.
R* didn't because they wanna release a good game. Might not be you cup of tea, and that's fair, but for what they've set out to create, i think they did a damn good job.