What’s New in ICML 2026 Peer Review; ICML Blog (blog.icml.cc)
ICML published some new rules for their 2026 peer review. Most notable are measures to combat AI slop and other ways of peer review abuse. They also mandate participating in organization of the conference if you submit at least four papers. Also, they will provide “advanced reasoning” LLM feedback before the submission deadline for authors.
It is clear that AI is causing stark changes in the research landscape. Right now it’s focused on AI slop, peer-review abuse, and so on. But I believe this is only the beginning, and we will have to deal with the broader impact of AI upending research as it was before.
Highly Opinionated Advice on How to Write ML Papers (alignmentforum.org)
Supreme advice on how to write papers for submission to ML conferences. It covers not just the writing, but also what to look for before writing and when/how to start writing.
It also covers some less nice parts of modern academia, like the randomness of review processes (see the table below for how two groups of reviewers rate the same papers) and that peer review does not maximize truth-seeking.
Scholar Maps: a journey from LiRA to the universe of human knowledge
Visualizing all of research.
Academic Conferences
"I'll be on Whova before the conference starts and cold-message random people like a cringy salesperson about why they have sushi in their interests. And that usually turns into a great conversation"
Python Dependency Management
Proper dependency management can prevent software engineering problems that could result in scientific problems
Simple Ideas
That's high-school level math, but it's really the whole secret. Just do this and it will get you a nobel prize in chemistry.