I’m not sure that’s quite true. Say the proof of proposition P requires a minimum of N symbols. You could prove it in one paper that’s N symbols long and nobody can read, or you can publish k readable papers, with an average length on the order of N/k symbols, and develop a theory that people can use.

I think even if N is quite large, that just means it may take decades or millennia to publish and understand all k necessary papers, but maybe it’s still worth the effort even if we can get the length-N paper right away. What are you going to do with a mathematical proof that no one can understand anyway?