Very cool and very interesting! I've always wondered why it never appears feasible to utilize a compression algorithm to shrink the size of the proof in this case, or compress data inside blocks in the case of btc.

I think the average data can be shrunk via compression is often by a factor of 1/3rd. From an amateur perspective, being able to fill roughly 3x the amount of data per size would seem to be a worthwhile approach. Especially as compression isn't particularly resource intensive, at least not in comparison to mining cryptographic functions. Adding a compression abstraction layer woudn't appear to bottleneck the process. But I would guess I'm missing some important point here as crypto engineers never seem to bother with compressing data?