It's not that easy to compare. Tokyo Metropolis includes various remote islands, so the official density is lower than NYC's. If we consider just the 23 wards, the population density is higher than NYC's but none of the wards seem to have a population density as high as Manhattan's.

👋 from Germany

you got it boss

I wonder how they plan to deal with the problem of hallucination. For human authors, we generally assume certain norms of scientific responsibility. For AI-generated papers, I think we will need reproducibility and other mechanisms to validate claims.

This requires a community-wide solution. And such a solution was indeed developed in 2008 in the DB community: a proper hybrid journal-conference model with monthly deadlines. ARR copied many of the ideas, but still major differences.

www.vldb.org/pvldb/volume...

sigmodrecord.org/publications...

What is transformative about LLMs:

Plurality (majority?) of students are no longer reading or writing papers

Soon the majority of all known writing will be synthetic

Videos of people now indistinguishable from the real thing

Net new types of fraud with no known counter

etc

Take it seriously!

I get to inbox zero all the time by turning actionable emails into to-do items that can then get sorted/prioritized along with other to-do items.

With this you can easily get to inbox zero, which feels great, though you will still need strategies to manage a growing to-do list :)

Since the traditional 2nd tier ones still exist, I guess your point is that that they haven't grown at the same rate? Indeed, the majority of papers that don't make it at the top ones probably just remain on arXiv. Might be a good idea to have something like Findings for the rest of AI.