[godelski]

> If only we had better ways to distribute fixes to papers.

I'm not sure what's wrong with OpenReview?

Anyone can upload, like arxiv. Comments are allowed and threaded so can be tracked. Revisions are possible.

But as a ML person, I post to arxiv a lot too. They have improved a bit, allowing for links to Github and such. Which we could make it common place to upload tex source files. Which are already completely accessible via arxiv fwiw.