New paper out in Proceedings of the Royal Society B: we apply linguistic tools to sperm whale vowels. The result: sperm whale vowels do not just look like human vowels. They also behave like them. We found several parallels. Like in Latin, whales have short and long vowels.

I love New Yorkers, entirely unperturbed by art that’s just some giant butts.

Who wants to see itty bitty printing presses?

I really can't believe how we don't talk about the incredible amount of bias baked into modern models nearly as much anymore, when egregious examples are _everywhere_.

Brutal.

Trump keeps ordering the DOJ to come to particular findings, so his administration can use those as a fig leaf to avoid future prosecution for those crimes. I’m not saying that courts won’t fall for this scam—I’m saying we have to overhaul our government so that they *can’t* fall for it.

If you'd like to hear @ftrain.bsky.social roleplay as Marion, a hunter struggling with internal demons, working to make friends at a fantasy tavern in Brooklyn...now's your chance. Thanks for having @hilarymason.com on the podcast, Aboard team!

The engineering management corollary of this is that token leaderboards are bad engineering management. (No exceptions)

MathCode MathCode is a terminal AI coding assistant with a built-in math formalization engine. Give it a math problem in plain language and it will automatically convert it into a Lean 4 theorem and attempt a formal proof. github.com/math-ai-org/...

CEOS SHOULD CODE, THEY SAID

This is how we did geo-based cell network selection at Google Fi: a convolutional neural net over a one-channel-per-(cheap)-telemetry-signal grid of S2 "pixels", used to predict (expensive) quality of service

S2Vec is a vector embedding for geospacial information! Very cool work, this makes a ton of sense conceptually but I imagine the data gathering/cleaning/validating effort puts it out of reach for most non-Google entities. research.google/blog/mapping...

This is a very cool debug

I really didn't foresee a discussion of the etiquette of approaching the flesh and how it differs between cannibals and vampires on my work agenda today.

Overheard a woman who I would guess was in her 70s explain to a friend that she “likes blue-skee it’s like twitter used to be” and her friend said “blue … what? Skis?” And why not, the whole thing was very sweet.

It’s a beautiful day, NYC, let’s get up and shine. #nokings

Same question, but for the English speakers -- do you get different results with slang? How about using proper punctuation vs a more casual style?

I'm curious if anyone who is multi-lingual has experienced different outcomes from using LLMs with prompts in different languages? Do gendered aspects of language help guide context? Does relative diversity of a language support or diminish clarity? Please share subjective experiences!

Dolores Huerta was the speaker at my college commencement. I don't remember exactly what she said, but I do remember being amazed and inspired by the conviction and power in her words. She's a hero. Imagine the strength it took, and while carrying the weight of what she endured.

TIL that Mr/Mrs Potato Head was invented in Brooklyn and was the first toy advertised on TV in 1952 www.instagram.com/reels/DVkjOK...

Turns out Swift is an aesthetically pleasing programming language? C++ is to Brutalism as Swift is to Scandinavian Modernism

I fell in love with this Jenny Holzer piece at the @brooklynmuseum.org yesterday, in the exhibit celebrating 200 years of this wonderful museum.

There's a parallel between people who "Make Content" and people who "Build with AI", and it's the loss of just writing things that are interesting, and making things that are interesting, to purely making things that can drive revenue, or worse, attention on this sick internet.

I so rarely get both nerd sniped and aesthetically sniped by the same thing, this is rad.

If you’re using 1M tokens a day for software dev or personal assistant LLMing I think you don’t know what you’re doing, not that you’re a great engineer. (With a very few exceptions)

WE’LL BE LESS ACTIVIST IF YOU BE LESS SHIT

If nothing else, #GDC has shown me that there's a huge need for support for how to make games that use LLM in the loop fast and scalable. And very few folks have real and transferrable experience doing this.

This is so very delightful.

Excellent metaphor.

I'm very grateful that Anthropic didn't give Claude a girl's name. (They are by no means a good or perfect company, but that's one obvious place that sexism could have been encoded and we were spared.)

Today's joy was an adventure to the Museum of the City of New York to see He Built This City: Joe Macken's Model If you also have a special interest in NYC, infrastructure, models, maps, etc, I would highly recommend it!

18 Photographs to Rekindle Your Love of Massive NYC Snowstorms

Wanting to understand why any of this horror is happening; what is their justification, and what’s really going on. Here’s one point of view.

In deals like the TMNT/MTG crossover, who is paying who? I really want to know those deal terms and how that came to be! Context: www.polygon.com/mtg-tmnt-bes...

Ilium and Olympos are books I have been thinking about since I first read them, probably 20? years ago. RIP, he connected so many beautiful ideas so creatively.

When academia's stars mistreat people, they're "punished" with relief from teaching, mentoring, and service responsibilities. This frees them to spend more time on the more valued work of research. And dumps less valued responsibilities onto colleagues, making it harder for them to become stars.

"It’s insane that “how much killing will we let the killer robots do on their own” is being hashed out as back-room handshakes between the military and its AI contractors. This seems like a matter of public policy if ever there was one. Have we got a legislature?" open.substack.com/pub/thebulwa...

Really valuable piece. It opens up a set of questions about the potential effects of AI on science that I have not seen widely discussed. And without pretending to determine whether those effects will be net-good or net-bad, it explains why metascientific *judgment* may become more important.

I love New York

Our @hiddendoor.co office is in Boerum Hill, and I commented a couple of months ago about discovering the original ExLax factory nearby: bsky.app/profile/hmas... And now the Neighborhoods newsletter has an even better telling of the story: theneighborhoods.substack.com/p/boerum-hil...

Everyone should read this -- @anildash.com explains the very human why of how so much of this horror has come to be. None of it is inevitable. We can choose something better.

Proud of my city: gothamist.com/news/nyc-are...

I went to The Sphere this morning to see The Wizard of Oz! It was a beautiful tribute to the original, a spectacular show of immersive theater, and evoked some questions of how to extend and care for a beloved film as it becomes a BIG experience; there were a few rough bits, and lots of wonder.

Over on threads someone just use ai;dr and we all need to adopt that right quick

Ask your nearest CEO what they’re favorite food is and see what they say. If they are all about Soylent and supplements as a way to fuel their productivity but have no interest in the beautiful sensation of food itself, BE WARY!!

I may be a small niche here, but which clothing brands sell beautiful gala gowns that are packable in a carryon, very comfortable, and give glamorous CEO vibes (vs just sexy vibes but not matronly vibes), and can be worn with flats?