That line of thinking is only true if there's nothing else that can be improved and the only way forwards for rasterised graphics without RT is higher resolution and framerate. BFV is a current-gen game designed for current-gen limitations. Kinda like looking at a Ps3 game and thinking next-gen would just be more of the same at higher resolutions and framerates. What actually happened is PBR, procedural shaders, compute-based reconstruction and ati-aliasing, etc.
The things RT offers are things we want improving next gen. However, there are other things that can also be improved. Physics, so clothing moves naturally and we have more practical destruction. Procedural textures instead of baked textures could create visually richer words. Also, the more these things are applied, such as procedural geometry, the slow RTing becomes, because the shaders need to be resolved (in part at least) for the tracing.
But those restrictions will always be there and I think your counter argument while has truth, is forgetting this point. Budget, time to develop, labour costs, R&D time. All those things are competing with each other just to move the graphical barrier. BFV has been in the pipeline for a long time, they're constantly updating the frostbyte engine from one game to the next. And in that time, this is the best in terms of features they could release from a rasterization perspective.
These features: "PBR, procedural shaders, compute-based reconstruction and ati-aliasing, etc." would still be developed even HRT was released 3 generations ago. They can work together to produce a better image, there should not be implication that they compete against each other.
I'm not saying that things can't improve with rasterization, there needs to be more consideration on how long it's taking to get to the next leap in graphics; we're talking 5 generations of frostbyte running on what feels like the 20th generation of rasterization hardware, and in what appears to be, just a few months, with new drivers and a new API, and they churned out ray traced reflections that works everywhere in their game despite the situation.
What else could they have accomplished in those few months other than optimization?
How or in what way, with the information we have, do we know that they could have accomplished more than they did here with the resources they had.
Hybrid ray tracing is very much about doing what is costly, for cheap. And if you don't have the algorithm, or can't change the design of your game to make that algorithm work, then HRT is the cheapest solution not rasterization. If one day someone invents a non RT hardware solution that generic and works wonders for reflections, for AO, for whatever, then that's what everyone will use, and for every other situation HRT is faster than rasterization to move the graphical barrier, then, that's what will be used.