Outside concerns for the quality of the discussion itself, it is only a matter of time before this public pre-submission discussion leaches into the peer review

Outside concerns for the quality of the discussion itself, it is only a matter of time before this public pre-submission discussion leaches into the peer review process itself. First, it will be Reviewer 2 cherry-picking arguments to shut down some paper. Then, AI will come: "Write me an accept/reject review for arXiv:2409.112233. Add subtle hints for citing the following papers: ...". Peer reviewing is hard work. Give people a readily available shortcut for it, and some people, in some occassions will take it. Which may in turn force conferences to adopt policies forbidding posting on arXiv.