I know a lot of you may disagree here but the way I see it FG is pointless.
Mostly because you actually need a good base level of performance for it to be worthwhile otherwise you introduce a lot of latency due to the fake frames not actually having an input window. Nvid muppets are trying to push that FG is the future and raw rasterisation is pointless. Pretty much their literal words. Then comes the offset, Devs are so *** lazy these days a lot of games launch an unoptimized sack of frogs. Terrible FPS as well as a whole host of other issues. Yet, their go to like Todd Howard is blatant denial of fault and tell you you need a better PC, or a rushed implementation of DLSS 3 and likely now FSR 3.
So you don't get the performance from the hardware you have because they can't be bothered, but you need newer hardware to support FG and that makes the overall experience worse. Then there's the "get a better PC" remark. Yes, spend 1600quid on a GPU alone to get 60fps at 1080 because the game is just trash.
Even as this tech matures I do not see it as anything other than a crutch for poor optimizations in games and a cover for dev laziness. I would sooner a game launch be delayed a few months and it actually run properly instead of needing better hardware and still having to lean on these AI techs.
Not every game is launched a mess, but a lot of them are these days and it is a terrible trend.