Try olmOCR 2 in the Ai2 Playground, check out our blog for more info, & download the weights and data from Hugging Face: ▶️ Playground: buff.ly/pJhoWXN 📝 Blog: buff.ly/8uUabID 🤗 Model & data: buff.ly/EUYHIuN

✨Version bump✨ txtdown v0.3.1 Minimal markup for Latin text collections Improved document validation; improved quotation handling; hierarchy-aware citations github.com/diyclassics/...

My goodness the Fantastic Futures conference in DC in Sept looks good www.conftool.org/fantastic-fu... Catch @amsichani.bsky.social and my paper "Data-Driven AI without Data" on Wednesday the 16th

Everyone should read this, just to see how they implemented a perceptron (a very basic neural network) in Age of Empires II arxiv.org/abs/2605.31514 The artificial neurons pass activation energy in the form of goats.

In visual culture & media theory, images from diffusion models are often analyzed through data critique: comparing what's produced to what's ingested. In this pre-print, I propose to look at the digestive system too: the decisions built into the system that automate specific ideologies of seeing.

I've been offline more than usual because of paper and grant deadlines plus work travel, but I'm back with my first real newsletter today! Click for rambling about infrastructure, benchmarks, Babylonian tablets, and more.

Check out this awesome zine from colleagues in @morphss.bsky.social among others!

New favorite (anti) AI book is @carissaveliz.bsky.social's PROPHECY. Two thirds of the way through and the philosophical zingers just do not stop! AGI to underground bunkers to Hannah Arendt on destruction is *chefs kiss* www.penguinrandomhouse.com/books/759692...

Finally! An article about the AMOC circulation that isn’t just clickbait / hype, and provides some real physical insights about how the ocean works. It turns out the ocean may be far more resilient than clikbait stories suggest. A very worthwhile read! www.science.org/content/arti...

1. Static word embeddings might model culture. But they're fixed. Boo. 2. Dynamic embeddings are contextual! 3. But encoding is computationally expensive. Boo. 4. Decoder only models are efficient! 5. But they seem to collapse all perspectives. Boo. What, then, ought the sociology of culture do? 👇

I made an artwork about the World Cup 12 years ago with Teju Cole and Mario Klingemann: www.jerthorp.me/timeofthegame Teju spoke so eloquently about the World Cup as "public time" and that idea has stuck with me.

Got a digitised collection that needs OCR? uv-scripts is a set of single-file Python scripts that OCR a whole image dataset to markdown in one command — 20+ open VLMs to pick from, nothing to install but uv. github.com/davanstrien/...

An LLM will grade a paper according to the criteria through which it is designed to produce a paper. As such, the LLM grades humans based on the criteria engineers use to assess the LLM. mail.cyberneticforests.com/its-not-just...

Yes! So glad the @inquirer.com has picked up on this awesome story. Congrats @nicosiamarissa.bsky.social !

Built a useful GLAM model this week: flags blank index cards before expensive OCR. Cheap, scales, adapts. IMO the technical barrier is now very low, the main one is knowing how. So I wrote this up as a chapter: danielvanstrien.xyz/ai-patterns-for-glam/patterns/index-card-classifier.html

If you're wondering why Chris Olah, the cofounder of AI company Anthropic, was sitting on the stage with Pope Leo today as the pontiff unveiled his new encyclical on AI, well ALLOW ME TO EXPLAIN. religionnews.com/2026/05/22/w...

This is a good thread, but this part in particular is key

FWIW, when Fish keynoted our DH conference at Oregon around this time, he was engaged and interested. I think because he realized that there were “non-insurgent” forms of DH work. Or maybe he is just very polite in person.

Re-upping for the evening crowd. Sharing an art bot that I programmed to post pieces from our art collection. Once a day. Forever. At the moment, it's a little lonely, so give it a follow! #bsmuseums #artbot #pennstate

NuExtract3 (4B, Apache-2.0) does OCR *and* structured extraction. Point it at a dataset of scanned index cards + a JSON schema → clean catalog JSON One command on huggingface Jobs. (or skip the schema for plain Markdown OCR) Script + dataset 👇

Preprinted! Been working on this for ~1 year w wonderful coauthors Nancy & Mark: > Understanding Artificial Neural Networks: Mysterianism about Known Mechanism is Mysticism doi.org/10.5281/zeno... Have you noticed many experts say they don't understand AI? What's going on? 🤯 Let us tell you... 1/

Goodbye Google Chrome, nice to have met you …

CfP: Revolutions and Reflections: A Conference for Editing and Recovery Practitioners elaboratories.org/event/revolu... August 8-9 at UVA #PublicHistory #archives #DH #preservation

I re-shared this earlier before I had finished it but it merits doing so again now that I have. @plach.bsky.social has a singular voice here that helped clarify a lot for me about how we might chart a path for AI resistance that pushes back against vendor control and imagines different futures.

Here in NYS, I think public institutions need to see if state guidance changes, since we also have local mandates coming down... In the meantime, here is the online version of the Federal Register's ADA Title II compliance extension

I thought this essay was spot-on in comparing vibe coding with gambling rushes and a “sinister flow state.” They’re not a staunch critic, but describe a disorienting experience quite well.

Btw my press funded a small study into best-practices for alt text for illustrations, especially of illustrations of art. The resulting guide is short and clear and available free to download.

What makes Pennsylvania perfect… @mwoluchem.bsky.social and @livgar.bsky.social of @datasociety.bsky.social on our latest Mystery AI Hype Theater 3000 episode, “Data Centers Go Nuclear”. Check out their latest piece in @newint.bsky.social. newint.org/science-and-... (w @emilymbender.bsky.social)

Ditto his comments on libraries

This story has everything, a focus on desal that led to overlooking other water sources, a city manager making $400K/yr saying don’t panic, a former water system manager saying residents should absolutely panic, and magical thinking: “The last hope to avert disaster was a 20- to 30-inch rainfall.”

The Price Lab’s Critical Approaches to AI Working Group has released a white paper in which we advocate for AI-free instruction in reading, writing, & research. These are fundamental skills in the humanities (& in general), & with decisive action we can keep teaching them well in the age of AI!

«The user experience and applications of the technology may change, but the foundational incapacity for accountability, and the underlying ideological and material infrastructures of the AI industry, remain.» Great text by @eryk.bsky.social

Here we go again. Excellent cover story of @thecontinent.org in 3 parts. "The US-Europe love pact rekindles their colonial kink" "This time around, the colonisers have AI" And: "The African Union is paralysed" Read it!

"It's not artificial intelligence. It's African intelligence." Michael Geoffrey Asia, the secretary general of the Data Labelers Association in Kenya tells @jasonkoebler.bsky.social about the notoriously brutal and underpaid work of training AI. Watch now: www.youtube.com/watch?v=QH65...

"The Protocol distinguishes between retrieval-based AI systems, which keep source materials under organizational control, and general-purpose model training, which absorbs knowledge into model weights irreversibly." 📜 libraopen.lib.virginia.edu/public_view/...

@kmcdono.bsky.social @danielwilson.bsky.social and I have a new OA article out: eur01.safelinks.protection.outlook.com?url=https%3A... It’s about the fragmented landscape of historical data, and what we can do about it to improve discoverability, sustainability and reuse.

The promise was always that digital humanities would be fully subsumed into the humanities - make the "digital" redundant. But the AI age has really accelerated the the "D" over the "H" in the field, causing big splits between "digital" and "public" humanities instead of "digital" vs "traditional"

Art History will continue to be seen as an elitist discipline if the public apparatuses for communicating it refuse to acknowledge the scholarship that has drawn attention to art's role in serving and subverting power. Read the righteous review by Sherry Lindqiust here. hnanews.org/hnar/reviews...

Docs is an open-source alternative to Google Docs for collaborative writing, developed by a French-German government collaboration: impress-preprod.beta.numerique.gouv.fr/home/

Visual art media can be made accessible and tactile. The link here describes a photography exhibit that was made for folks with visual impairments, including tactile versions of the photographs with braille descriptions. www.rte.ie/news/ireland...

Searching for any form of "magnus" followed by any noun—just one example of flexible pattern matching possible with LatinCy Readers `find_sents` call... from this demo notebook: github.com/diyclassics/...

Beautiful thread from my ride-or-die @charliejane.bsky.social about the organization that our @humblebundle.com is benefitting. Buy some books (cheap!!) and help the amazing Miss Major Alexander Lee Black Trans Cultural Center!

We have another exciting CFP! "Noisy Systems: Aesthetics, Epistemology & Computation examines noise across technical, social, and cultural domains, bringing together machine learning, critical AI studies, media theory and archaeology, art history, philosophy, and practice-based research."

Really looking forward to exploring this open access special issue of The Lion and the Unicorn dedicated to the intersection of Children’s Literature and Digital Humanities co-edited by @drfitzphd.bsky.social muse.jhu.edu/issue/56268

University of Arizona Libraries is hiring for a RARE BOOK LIBRARIAN! 1st review of applications February 9, 2026. arizona.csod.com/ux/ats/caree... We have great collections and great people here. Not on the committee, happy to chat.

Proud to have Dine weaving in my personal collection, this recent video about weaving is really great and personal and moving: www.youtube.com/watch?v=MWis...

Our tribute to Natalie Zemon Davis, in the Archiv für Reformationsgeschichte. Please share widely! www.degruyterbrill.com/journal/key/...

Paste any IIIF manifest → model classifies every page locally → see where illustrations appear. Part of small-models-for-glam: small, efficient models for cultural heritage work. Not everything needs GPT-4! Try it: huggingface.co/spaces/small-models-for-glam/iiif-illustration-detector