Please provide evidence. Source code examples having one then "copy / paste" to get the other two. Or direct developer commentary from a neutral (non-nVidia sponsored) developer.
All three of the GPU company Scalers use the same inputs/setup. The real work will be exposing those inputs and setting up for the first scaler you support, but after you have done that for any one of them, it definitely would be as simple as copying what you did for the first one, and changing the calls to one of the others. Though most developers would do as was mentioned earlier in the thread and build some kind of abstraction layer. Either using NVidia Streamline or something in house.
Also for third party game engines like Unity and UE4, the work is already done (and I would bet most in house engines as well).
I have seen the example for UE4. Literally all you have to do for UE4 is download the free DLSS plugin, put it in your plug-in directory, activate the plug-in, and restart your project, and you have DLSS working.
It's literally about 5 minutes work for UE4.
Anyone claiming it's a lot of work to add a 2nd or 3rd scaler at this point is either ignorant, or lying.
We were told something similar about ray tracing. "Just flip a switch and everything automatically works everywhere without developer effort!"
A lie.
We were also told "once engines support DX12, everyone will get automatic performance boosts without any developer effort, including all indie developers!"
Also a lie.
I think you misinterpreted some things.
These are fundamentally different things. RT obviously requires extensive work, per scene, it's an active decision about where you will and won't use reflections, ray limitations, tuning etc...
A scaler OTOH does not involve any per scene tuning. It hooks into the rendering engine in one place, once you do that setup, it's essentially done. Once you have hooked in one scaler, the others hook in the exact same place, with the exact same inputs, so it's absolutely trivial to add the others.