Kameron Decker Harris

Hey AI, please rewrite my shitpost in the style of Emily Brontë’s “Wuthering Heights.” 🔥

That would mimic a harmonic input, but you'd also need a network with an architecture/training meant for sound input. Like a speech to text network or something

Yessssss. 💯% agree

Advancing regulatory variant effect prediction with AlphaGenome - Nature AlphaGenome, a deep learning model that inputs 1-Mb DNA sequence to predict functional genomic tracks at single-base resolution across diverse modalities, outperforms existing models in variant effect...

I had intended to post something about this new Google DeepMind paper that appeared yesterday in Nature, but the press coverage has added to what there is to say. So this is a long 🧵
www.nature.com/articles/s41...

Breaking the Spell of Vibe Coding – fast.ai Sinister variations on the positive state of flow

www.fast.ai/posts/2026-0...
"AI coding agents can produce syntactically correct code... they don’t produce useful layers of abstraction nor meaningful modularization. They don’t value conciseness or improving organization in a large code base. We have automated coding, but not software engineering."

I dunno, but my cat did this twice in the night and woke me up!

A prompt for the phrase “feminist music” which Spotify says “I can’t help you with that topic.”

My first prompt for “feminist high energy good vibes workout style music” which was refused by Spotify with the error “we can’t work with that prompt”

Spotify prompted me to try their new “prompt a playlist” tool and they BLOCK THE WORD “feminist”.

WTF Spotify!

In case anyone wants a little break, here are a few of my favorite facts that I learned in 2025, in no particular order.

Some trees benefit from, and may even encourage, getting struck by lightning because it kills their nearby competitors nph.onlinelibrary.wiley.com/doi/abs/10.1...

Neat, we're going to explore using it during my group meeting tomorrow

🚨 #CCN2026 Proceedings submissions are open!
CCN 2026 again features an 8-page Proceedings track (alongside extended abstracts). Accepted papers will appear in CCN-Proceedings (CCN‑P) with DOIs on OpenReview.

Very cool, congrats Dean!

New paper out at PNAS: www.pnas.org/doi/10.1073/...
Revisiting the high-dimensional geometry of population responses in the visual cortex with @jpillowtime.bsky.social. The review took forever because a reviewer was doubtful our new estimator can infer eigenvalues beyond the rank of the data! (1/6)

Beautiful spot!

Only one way to find out....

This reminds me of the recent work out of Colin Bredenburg and others on top-down/bottom-up model of hallucinations: elifesciences.org/reviewed-pre...

Totally, or maybe higher-level (?) circuitry that identifies figures. I once had a hypoxia-induced hallucination of tiny vizcachas (rabbit-like animals) on a high-altitude Chilean volcano.

Diisopropyltryptamine produced auditory hallucinations, strange since it's similar to other classical psychedelics in the tryptamine family, including psilocybin.
en.wikipedia.org/wiki/DiPT

a cartoon of alice from alice in wonderland surrounded by mushrooms ALT: a cartoon of alice from alice in wonderland surrounded by mushrooms

It's not the only drug or even only mushroom with stereotyped hallucinations. Muscimol, from Amanita muscaria, commonly produces distortions of perceived body size.

"endoscopic crust with Ramanujan filling"

Yes please. It's hardly possible to review the technical content, not to mention the citations

TIL about the Hamantash spectrum!
From "Ramanujan bigraphs and applications"
math.huji.ac.il/~parzan/bigr...

Shrinking the number of people you advise will lead to better mentorship, too.

I’m going to halve my publication output. You should consider slow science, too If we don’t slow down, the research enterprise is going to crash, argues Adrian Barnett.

If you have tenure consider publishing less... doi.org/10.1038/d415...

Those pickles are so incredible, I agree

The upper and lower bounds here are the same but shouldn't be. Interestingly, for my student the overview mistakenly swapped the (correct) upper and lower bounds.

Subtle little typos in Google's AI overview for the search "approximation of sum by integral" found in office hours yesterday. I was leading the student through the derivation then we went to check online, and I found the AI answer was wrong.

Congrats, that was a fun preprint to read.

- DeMarzo et al. (2003) "Persuasion Bias, Social Influence, and Unidimensional Opinions", QJE. academic.oup.com/qje/article-...
- DellaPosta et al. (2015) "Why Do Liberals Drink Lattes?, " AJS. www.journals.uchicago.edu/doi/10.1086/...

Hah, nice take. I need to try more AI coding