PART 2: LETTERS 1. "warn" people ahead of time that you are going to ask them for letters, it takes a while to write these. 2. Some ppl think more letters = better. I don't. Letters are not additive, they average. If you have 4 letters that correspond to [8, 8, 9, 6], just send in the top 3.

It’s grad school application season, and I wanted to give some public advice. Caveats: -*-*-*-* 
> These are my opinions, based on my experiences, they are not secret tricks or guarantees 
> They are general guidelines, not meant to cover a host of idiosyncrasies and special cases

for various reasons I had to briefly explain the plot of Hamlet to my seminar class yesterday (turns out none of the students had read or seen it), and so I got to use the phrase "Hamlet is Big Mad about this"

mid-semester shout out to Andrew Shtulman's book "Learning to Imagine"; I'm teaching a seminar on imagination and pretense, and the book has fun and useful, and leads to interesting discussions

Nature suggests you use their "Manuscript Adviser" bot to get advice before submitting I uploaded the classic Watson & Crick paper about DNA structure, and the Adviser had this to say about one of the greatest paper endings of the century:

as above, so below

מסורת האלווין: בבוקר אחרי, הילדים מוסרים את מה שנשאר תמורת מתנה, וזה מוסדר בחוזה. (דרך ללמוד על פרצות בחוק. בגרסאות קודמות של החוזה לא היה כתוב איזה מתנה הם יקבלו ומתי, למשל)

a relevant False Knees

[moments after creating my AI digital clone] ME: alright clone, do my chores CLONE-ME: no US: my god, it worked

"and how wide do you want your stairs?" "oh, about 1 Yuki" "...what"

if I were a VC I would interview potential employees just on their ability to categorize AI startups as [potentially legit] / [impossible] / [nonsensical] / [tractable but dystopian]

I got an invitation to use the "Nature Research Assistant", put up a paper on intuitive physics and asked it for more references to relevant papers; can you guess why I marked those parts in red

קיימברידג׳, הגזמת

🚨 New preprint 🚨 How do people's mental models shape memory, prediction, and generalization? We find that people spontaneously construct goal-dependent causal abstractions that compress experience to privilege relevant information. 📃 osf.io/preprints/ps... 🔗 github.com/cicl-stanfor...

overheard in coffee-shop convo now between two tech bros.: "I don't want to think about AI 24/7, you know? I want to think about other stuff! Like...like...stoicism, you know?"

In case you don't know already, the journal Open Mind has a Bluesky account that automatically posts new papers: @openmindjournal.bsky.social The journal is diamond open access (free to read, free to publish) thanks to the support of MIT Press, Harvard Library, & MIT Library.

(niche nonsense) I've been reading on taboos and unsavory exchanges (part of a reading group on vagueness/indirect speech) The notion is that some exchanges cross moral boundaries, e.g. buying a kidney or a child; so you ways around that

אני מבין למה אנשים מתבלים את העברית שלהם באנגלית כשחסרה להם מילה (או ביטוי), אבל אני לא מבין שימוש כמו: ״וואנס שאתה מבין את הבעיה...״ ״זה דבר שהוא אקרוס כמה תחומים״ ״הייתה מסיבה רצינית, היו שם את כל ה הו אנד הו״ (הכל דוגמאות אמיתיות שלא נאמרו בבדיחה)

academic paper titles :

there's a lot of interest in making AI more "human like" -- but what about the flip side? when real humans seem like robots/puppets/NPCs? nnnnnewwww pre-print: "The Development Of Sensitivity To Automatic Behavior" (by Bass, Bonawitz, & me) doi.org/10.31234/osf...

Going to the Ren Faire and shaking my head the whole time so the people in the King's Field know I disagree with it

someone talking loudly with a friend at the coffee shop: lovely, the fun of being human, the joy of a coffee shop, good for them someone talking loudly over a zoom conversation at the coffee shop: you are the worst of us, how dare you, this is not what we Do Here

embedded video paused at the right time

It's been 15 years since Edna Ullmann-Margalit passed away, and I keep going back to stuff she's written. I highly recommend 'Normal Rationality', which collects her essays. If you're looking to start, maybe look here: bit.ly/4qk2GZS bit.ly/46XudIV bit.ly/4nkfLQc bit.ly/3KUIiOy

the best things in life are freeequently around 20$

Halloween stores sections are organized by "decorations", "props", and "costumes" when they should be organized by "spooky", "scary", and "creepy"

New preprint! "Non-commitment in mental imagery is distinct from perceptual inattention, and supports hierarchical scene construction" (by Li, Hammond, & me) link: doi.org/10.31234/osf... -- the title's a bit of a mouthful, but the nice thing is that it's a pretty decent summary

"we present Work(Around)-Bench, a benchmark for assessing whether LLMs are actually doing a task or simply finding silly short-cuts around it. So far every model is at 99.6%"

are you at #COLM2025? Check out this spotlight talk by Sonia Murthy! @soniakmurthy.bsky.social (room 520B @ 11:30am) are you NOT at #COLM2025? Check out this paper by Sonia Murthy! paper: arxiv.org/abs/2506.20666 Kempner Deeper Learning blog feature: kempnerinstitute.harvard.edu/research/dee...

(Daughter, 4 years old, crying her heart out) Me: "What's wrong, tutu?" Daughter (moving her hands on the sofa): "If my fingers were markers they would ruin the sofa!" Me: "But your fingers...are not...markers?" Daughter (peak distress): "I said IF!"

I had a great time visiting the Northwestern Institute on Complex Systems today (much better than that time at the Southeastern Collective in Simple Entities)

is there an accepted technical name for the noises people make while doing some task in the presence of other people, e.g. pulling up some code and meanwhile going 'do do doo dee doo"?

NEW on our #DeeperLearning blog People balance being kind vs. being honest — and #LLMs should too. New research shows training choices often favor informativeness over kindness, but prompting can induce sycophancy. Read more: bit.ly/3Wqrtxl

there's a kind of argument I've taken to calling Theseus' Motorcycle: You point to the ship and say "this ship cannot roll on land" "It can too", says the other side. repeated tests show the ship cannot roll on land,

so sick of the explore-exploit dilemma, how about something new?

if you happen to be in the Harvard area over the next few months, maybe check out The Gloomy Gallery, a small show featuring the work of Edward Gorey (the curators clearly had fun with it)

It's officially been 75 years since the proposal of the Turing Test, a good time bring up 'The Minimal Turing Test': www.sciencedirect.com/science/arti...

a fun and thought-provoking read

daughter (9) with a contender for the greatest opening line of the past century

Q: Why did the LLM cross the road? A: We're not sure, but it achieved 94.7% on CHIKENBench-Large

in 1952, the BBC held a discussion between Turing and other academics, on "Can automatic calculating machines be said to think?". The sound is sadly lost, but the transcript is here, and it's very cool: turingarchive.kings.cam.ac.uk/publications...

Dear Editor, Thank you for submitting your review request to Tomer Ullman. As you may know, we decline a substantial proportion of reviews requests without sending them out for further evaluation. After careful consideration, we regret to say we cannot offer to take on this review.

Shorthand version of the paper: telling people that "Themaglin is associated with an increased probability of Pneuben” makes people think that Themaglin causes Pneuben. (or maybe it is just associated with an increased probability of that inference?)

funding orgs studying AI risk seem to be using LLMs in their review process, which poses the risk that a dangerous AI will downvote proposals that may detect it, but my grant to study this problem keeps getting rejected :(

paltering זה מונח בלי תרגום מדויק לעברית שמתייחס לסוג ספציפי של הונאה בה מספקים מידע אמיתי כדי ליצור מחשבה כוזבת. אז, זה נכון ש״רק״ 44% בסקר בבריטניה תומכים בהכרה...אבל גם רק 18% מתנגדים. לחלק גדול אין דעה, ובקרב בעלי הדעה היחס הוא 70\30 בעד. en.wikipedia.org/wiki/Paltering

THE PROPHECY