Regardless of how much one may love the AWE32, let's be honest about this: the UltraSound was the first mass-market "wavetable" sound card that delivered decent sound and didn't cost a fortune. It is arguably the reason the AWE line existed at all. Love it or despise it, the historical significance of the UltraSound is undeniable.
As already mentioned, the idea of running GUS software using a card with an on-board EMU8000 as a replacement may seem natural, but would not be easy. Architecturally, the two are quite different. On the MIDI side, their instrument formats are sufficiently different that there is no simple sure-fire way to automatically convert from one to the other and be sure to get decent sound. More generally, their methods of interpolation, constraints on memory layout and alignment, and methods of avoiding clicks during playback are all quite different. Designing a TSR that would successfully account for all of these differences at all, much less seamlessly, would be quite a challenge.
Then, if we're discussing this because the objective is running the many GUS-only demos, we run into even more complications. Many of the people who write those things are fantastic at grabbing the hardware "by the ears" and getting it to do unexpected things. Developing a TSR that could handle all of the special cases that they are likely to have leveraged to get the results they wanted, and also work with the sometimes fully-custom DOS extenders in use would be an ambitious undertaking, to say the least.
Is such a thing possible? Maybe. Would it always work? Very unlikely. Would it be perfect? Definitely not. Given the challenges in developing such a thing, the price of a GUS really starts to look less insane. If you think you're such a great coder that you'll be the one to do it, feel free to take a stab at it, just don't blame us if you have less hair at the end of your attempt than when you started. 🤣