…you should also reread what I actually meant. My argument wasn’t about shader use cases - it was about the current engine architecture and prioritization. But since you interpreted it as an argument against shaders, I’ll respond accordingly:
You say compute shaders aren’t niche, but then immediately bring up post-processing effects, which are definitely niche in the context of Roblox - especially given that we already have built-in post-processing instances. More advanced effects, like the one you mentioned, are obviously impossible under the current engine, but realistically, only a specific subset of games would even attempt to use them.
And when it comes to non-graphical GPU computation, you’re venturing into GPU-specific architecture and coding (e.g., mesh shaders, GPU particles), which isn’t portable and would introduce a whole new set of platform compatibility challenges - especially across mobile, where Roblox still maintains a large footprint.