I think the obvious reason for INT8 still sitting in development is the tech press itself. The problem is, AMD has just released it's latest "wanted by popular demand" hardware, the X3D3 CPU. Almost since the day the dual CCD X3D was released, users wanted the cache enabled CCDs on BOTH chips, due to scheduling issues and imagined gains. AMD knew it was a marginal upgrade at best for a number of reasons, which made them look even more like they were "hiding something" good and easy to do. After the big AI hardware vacuum appeared, I guess AMD decided "what the heck" and made one. The predictable result was that the "tech press" took a hatchet to this chip for it's "very small performance increase" and cost. While a few reviews did mention this wasn't AMD's idea, they still hammered home the value proposition. "What's the point..." seem to be the subtile for the reviews, which was probably obvious from the get go.
I can see the INT8 FSR getting the same press for the same reasons. It bites compared to the "real" FSR, too many artifacts, and my personal favorite "it's not as good as DLSS", especially when looking at screen captures under a microscope.
The problem with releasing a "not ready for prime time" product is that no matter how many times we're told it's not ready for prime time, the public and the press will quickly forget this during review time.