What is your favorite finding from any paper that involves the application of computational methods to natural language data?

A neat paper I worked on, begun over 2 years ago when our concerns were more myopic, was just published today. It came out an excellent set of workshops at the Santa Fe Institute, led by the indefatigable Joshua Garland. The gist: sometimes inefficiency is good for you www.nature.com/articles/s44...

Really great new paper using agent-based modelling to show how an exploratory childhood can lead to innovation in the population at large. royalsocietypublishing.org/rspb/article...

If you're interested in a short behind-the-scenes of the retraction of a mega-metascience paper—the first of its kind (the retraction, not the paper)—this might be for you! Had to gloss over so many details to observe allotted time but happy to share an extended version of my slide deck with links.

Cells can sense the geometry of their environment, from tiny subcellular features to whole-tissue patterns. How can we engineer environments to systematically probe these responses? 🧵 I'm @lhinderling.bsky.social from @olivierpertz.bsky.social lab, let’s do a short tour on microfabrication!

Is there a term of art for the materially real aspects of fictional entities via their effects in the physical world? Ex: Santa stories are related to people putting presents in stockings so that this action becomes a real-world manifestation of Santa stories.

Super proud of this paper with @apvelilla.bsky.social and @babeheim.bsky.social, now out in Psych Review. Non-paywalled version (preprint) here: osf.io/preprints/so...

Introduce yourself with five concerts you’ve seen: * The Jesus and Mary Chain * Leonard Cohen * Pixies * Lou Reed * Sigur Rós

My latest column for @newscientist.com is up for subscribers (including library card holders)! This was a fun and hard column to write because I had to explain quantum field theory in like 3 sentences 😹 I describe how unsolved problems drive creativity 🖖🏽 www.newscientist.com/article/mg26...

Wishing @helendecruz.net were here to see the pope recognize wonder as the key human ability. You should read their book Wonderstruck if you haven’t already.

Pretty much no one can make their ideas clear to others on the 1st attempt. The two main reactions are "let's painstakingly clarify this with some back and forth", and "doesn't matter as long as the various interpretations are themselves interesting", leading to the 2 main branches of philosophy.

A distributional semantic comparison of articles in Grokipedia vs Wikipedia could be interesting…

"Because science rejects claims to truth based on authority and depends on the criticism of established ideas, it is the enemy of autocracy. Because scientific knowledge is tentative and provisional, it is the enemy of dogma. "

Today's offering from the #NaturalPhilosophy Symposium: Kevin Zollman @kevinzollman.com on Refutation and Models of Social Organization. Commentary by Henry Farrell @himself.bsky.social. www.youtube.com/watch?v=ybOr...

New paper on formats of representation in LLMs. I defend pluralism both about the vehicles of representation in ANNs (e.g. neurons, polytopes, embeddings) and about the formats of representation (e.g. nominal, analogue, structural). Hope it’s useful to someone. philpapers.org/rec/MALFOR-2

As we put it here, “the shape is the meaning”: arxiv.org/abs/2509.00248

What problem is explainability/interpretability research trying to solve in ML, and do you have a favorite paper articulating what that problem is?

I’ve been preparing a presentation on this paper and trying to more concisely get the main point across. I’d say it comes down to this: To make the interpretive lens of a particular modeling decision visible, you have to contrast it against a population of alternative decisions. 1/5

if attention mechanisms in transformers create graphlike transformations across text then it is actually an incredible pun that groups of sentences are called "para-graphs"

"Helen de Cruz (1978–2025): an academic and personal tribute" *Religious Studies* Published online 2025:1-4. doi.org/10.1017/S003... @helendecruz.net

I really enjoyed talking to @nsaphra.bsky.social about her thoughts on what much language model interpretability research misses. My latest in @quantamagazine.bsky.social:

i think regardless of who has authored a paper, we (the scientific community) can and should stop criticizing the work on lack of perceived novelty. we can do better than perpetuating harmful myths about individual scientific papers having to present a discovery no one has ever thought of before.

Forgive one last point to vent about: The reviewer’s only other criticism was a lack of novelty. The obsession with novelty has made a lot of ML work insufferable so it’s unfortunate to see it here. But here’s who is apparently already making our points about how ML is used for humanities research:

How to quantify the impact of AI on long-run cultural evolution? Published today, I give it a go! 400+ years of strategic dynamics in the game of Go (Baduk/Weiqi), from feudalism to AlphaGo!

Well, this got rejected today as a short position paper from CHR so I suppose I can share it now: arxiv.org/pdf/2508.00095 Here’s a hopefully coherent ramble about what we’re trying to make clear:

Well, this got rejected today as a short position paper from CHR so I suppose I can share it now: arxiv.org/pdf/2508.00095 Here’s a hopefully coherent ramble about what we’re trying to make clear:

Out now! Bride of Frankenstein [film|minutes]

A quick summary for busy people: We propose a mathematical framework for learning representations by extracting the algebra of transformations of worlds from the agent's perspective. 1/n www.sciencedirect.com/science/arti...

💻 We are very excited to announce the launch of our special issue Computational Formalism, edited by Tess McNulty and Laura Alice Chapot, that reconsiders the intersections of computation and form at this emerging technological and critical moment. culturalanalytics.org/issue/12788

“[Something] unpredictable, spontaneous, unformulable and ineffable is found in any terminal object. Standardizations, formulae, generalizations, principles, universals, have their place, but the place is that of being instrumental to better approximation to what is unique and unrepeatable.” (1/2)

A GoFundMe for the instructor who was fired at Texas A&M!

I have a thing under review right now that argues translations across research domains (as epistemic “languages”) are useful insofar as translators can give a theory of their work. Auto-translation necessarily hides interpretive choices and so can only be useful in very narrow cases.

Why do societies reliably develop strikingly similar traditions like dance songs, hero stories, shamanism & justice institutions? In a new BBS target article, I propose a theory for such "super-attractors" + cultural evolution more broadly. Now open for commentary: www.cambridge.org/core/journal...

Great bit from Leslie Lamport in this video: “Coding is to programming what typing is to writing. Writing is something that involves mental effort. You’re thinking about what you’re going to say. The words have some importance but in some sense even they are secondary… youtu.be/rkZzg7Vowao?...

Here's a nice new paper yet again showing the benefits of drawing for better comprehension and understanding (which is not surprising from a multimodal language view): "Benefits from sketching for improving comprehension monitoring from illustrated texts" link.springer.com/article/10.3...

Well, there’s the answer to my student’s question

You will never be able to make "data driven" decisions until you realize just how much your decisions are driven by emotions: fear of missing out, defensiveness of your favorite idea, apprehension over something unusual. That's because there's no such thing as "the data" - only interpretations.

Seeing that @matthematics.com liked a post is a nice way to find out he’s here on Bluesky. I’ve been watching his abstract algebra videos over the summer and they are unbelievably good. Highly recommend.

Left out a crucial step (to not do): (2a) Evaluate the process by which the numbers in (2) were computed according to a metric that has no relevance to (1).

How to not do computational humanities: (1) Lay out a question/hypothesis about a complex, cultural domain. (2) Compute numbers that were inspired by (1) but without sufficiently formalizing (1) so as to meaningfully link to it. (3) Interpret numbers to mean whatever your prior vibes were.

Chess 2 *finally* got released.

plant_pol_kato: Kato plant-pollinator web. A bipartite network of plants and pollinators from Kyoto University Forest of Ashu, Japan, from 1984 to 1987. 772 nodes, 1206 edges. https://networks.skewed.de/net/plant_pol_kato Ridiculogram:

At long last, I can post my team's summer project: applied modules to teach how ML/AI tools are changing social science and humanities research: ubcecon.github.io/praxis-ubc/ Highlights: LegalBERT to analyze anti/pro-immigrant sentiment in 19th c. BC law: ubcecon.github.io/praxis-ubc/d... 🧵1

I get the sense that a fair number of people attribute a kind of strong emergence to LLMs such that there’s a hard distinction between an LLM as giant function and the simple functions that constitute it.

Is anarchist science possible? As an experiment, we got together a large group of computational neuroscientists from around the world to work on a single project without top down direction. Read on to find out what happened. 🤖🧠🧪

Hiroki Sayama announced that CCS2026 will be held in Binghamton, USA — see you there! 🇺🇸 #CCS2026