Who’s Adam? The Most Outrageous NeurIPS Review Unveiled}
A recent NeurIPS review has gone viral for its humorous and critical tone, highlighting the often bizarre peer review process in top AI conferences, sparking widespread discussion.
These days, everyone has received their NeurIPS 2025 review results, right?
Usually, it’s the time to start complaining about the review comments.
But just a few hours ago, we saw the most outrageous NeurIPS review comment this year on Weibo.
From Yiping Lu, an assistant professor at Northwestern University’s Department of Industrial Engineering and Management Science, and a Peking University alumnus.
Within hours of posting, it had been viewed over a hundred thousand times.
The reviewer’s comments are as follows:
Both architectures use Adam optimization. “Adam” — who/what is that? I believe this is a serious spelling mistake that the authors should have removed before submission.
Yes, this is exactly the review comment from Professor Lu’s NeurIPS paper.
Even Professor Dan Roy couldn’t resist criticizing: “NeurIPS reviews are a total mess.”
Doesn’t this pain you?
With the hotness of AI conferences, the surge in paper submissions and the quality of reviews seem increasingly incompatible.
This year, NeurIPS received nearly 30,000 submissions.
It’s impossible to review all manually.
Can we rely on AI for review?
In fact, AI-assisted review has become common in academic conferences today. Xuandong Zhao, a postdoc at UC Berkeley, said: “Two years ago, about one-tenth of reviews felt AI-assisted. Now? It seems nine out of ten reviews are edited or generated by AI, including grammar corrections and even complete comments.”
AI now seems to be involved throughout the entire process of writing, reviewing, and publishing papers.
After witnessing these outrageous reviews, everyone should still prepare to rebut properly. A blogger from 2020 was recommended, which might help.
More conveniently, AI Review Tips from 2020 are also summarized in a Chinese article: “18 Tips to Convince Reviewers and ACs When Your Paper Scores Low”.
Finally, after the review results are out, NeurIPS authors, don’t forget to submit your papers to our AIXiv column!