So far, I've got

St Petersburg's paradox
Gambler's fallacy vs regression to the mean
X vs Y and Y vs X having the same slope in lr (equivariant)
Zero probability in continuous random variables

Though I'm surprised nobody's really put together a take on US-China trade war. But I'm also unclear what radicalizing means in this context haha

To me, it seems EAF/LW writers lag behind the most recent current events unless they already have a preset take. Considering most of the tariffs were walked back today, I wouldn't be surprised if stuff is now stuck in drafts for 90 days until the pause expires

Quote dunking = using the quote repost for snide remarks against the original poster.

Unfortunately, the closest label we got is trans rat murder cult 😔

Truly bizarre to see the term "neo-Marxist" on a government document

Thanks for the link! Still a bit of a headscratcher though because it's pretty unclear how they go from their buzzwords to their totals. I was curious about studies that control for confounders, for instance, which aren't in the examples but would likely contain some mix of "DEI" terms

Wrote it down for funsies

Thanks! My bad for not thinking to check Dec 2024.

Yeah, I figured, considering "corr" and "causal" didn't bring up anything useful in these articles haha

Coffee and pure juice is vegan and good, apparently

It might be me, but it seems every news article about this doesn't link the original study. The closest thing I could find is this (based on Google Scholar results for "uk biobank milk depression" from 2025) which has its own incredibly detailed infographic to speculate over

The article even points out "idiosyncratic weaknesses" like hallucination, brittleness, letter counting, and tic-tac-toe.

But (my reading is) betting on these known weaknesses shouldn't underpin regulatory strategy. E.g., we can't assume logic puzzles can be the new CAPTCHA

This claim is disputable, but valid concerns about OOB performance, generalizability, multimodality, etc. aren't super relevant for this piece's purpose.

It's mostly a rhetorical transition to saying "assuming AI can't equals won't is risky and baseless thinking in regulation".

I may be misremembering this, but back when Harvard was scouting for Presidents (pre-Claudine Gay), the search committee took recommendations from students.

There was some whispers to get a group push for Avi Loeb, in no small part because it would be really funny.

(Should have linked this first, whoops)

FOMO from NYU faculty is referring to a "Reply All-pocalypse" where it seems that an email somehow hit a mass mailing list that looped in all the faculty into a gigantic email chain.

Would a crow be close enough?

Shoutout to voiceless wanderers in a beautiful and unsettling world gotta be one of my favorite genders fr fr

The original post is rage bait based on a bad machine translation of Chinese. An alternative reading is that the CEO had no idea what the GotY standards were (in a lighthearted "I'm new to this" way) and the acceptance speech was a team pep talk.