Oh no .... The first rounds of ICLR in 2013+ had no page limits (and the submission site was .... arxiv!). At some point I reviewed a 47 page paper which looked like an intern report "we did this, then that, and here is a table, and some python code, and then this didn't work so we did this ..."
Comments
I mean, eg, let's take a paper / idea everyone clearly thinks is impactful. Transformers... But, have you, umm, read the Attention is all you need paper? Impossible to understand.
Imho there's no clear answer, but the other extreme, like Shinichi Mochizuki's unchecked works, has downsides