Tags: * against review load proportional to submissions * soften sticky meta/reviews
"I am not convinced that this author-reviewer automatism is constructive. When a large lab submits,
say, 20 papers to ACL, the lab head is the one who already needs to work on/review all 20 of them before
submission, and then gets ""punished"" by receiving another 20 papers to review, while this experience
would be much more beneficial to postdocs and senior PhDs... I would assume that senior faculty is also
the most frequent source of ghosted reviews? I am also wondering if ARR shall be optional, or allowing
reviewers to see the previous reviews shall be a free choice of the authors. Sometimes the reviews are
non-fixably negative (the type of ""didn't use BERT -> strong reject"", ""used only BERT, not novel
enough -> strong reject""), and having people seeing those in the next reviewing round doesn't bring
much beyond a permanent negative bias."