Late to the party on this, but I like how this ends: “Confronting the material effects of AI on inequalities of wealth and power is even more critical, and the outcome of that struggle is hard to predict. What I do feel certain about, however, is that we need an active strategy.” Yes!

Hi DH friends, join us on Nov 10, 10-11 am CT, for “New Book History Research with Internet Data”, a hybrid panel sponsored by SHAR, to explore challenges and opportunities of using Internet data and digital methods for book history research. More info in the poster attached and comments :)

Kimi K2 Thinking and announcement tech blog is now live.

Our School of Information Sciences is running four searches, and we can hire more than four candidates. Here’s the first link, for a job that might appeal to people working in history of information, history of science, or digital humanities. +

DataRater can continuously+adaptively learn which examples will improve an ML model the most during training. Delighted to have worked on this with @luisazintgraf.bsky.social, @dancalian.bsky.social, @gregfar.bsky.social, Iurii Kemaev, et al. (see full list at the of Luisa's thread).

This forum is very worth reading. But the focus on students does remind me of one place I resisted the Chronicle's (generally excellent) advice on my own piece. When I said AI's been adopted by faculty, students, and staff, my editor added "—especially students!" Which was funny, but I deleted it. +

Straw men organizing to protest their abuse by Republicans. "We are made out of straw! We can't take much more!"

Come work with us @DHSS! 📢

it really cannot be overstated how good a tacit knowledge database is "is the best practice to do it like this, or like this?" unlocks soooo much in any domain

“But I do feel confident that the strategic move is not to hope AI never learns to solve our problems, but to demonstrate that we are the ones who can teach it how. And then to invent new, more difficult problems to solve together.” An invigorating vision!

Flying tomorrow and I am definitely packing an extra change of clothes for anticipated fuckery on the return leg

You know how the existence of cars and power tools makes it necessary to have “exercise” and “manual crafts” as a deliberately chosen part of a balanced life? We’re probably going to need something like that cognitively, for adults.

A new paper with Bogdan Georgiev, Javier Gomez-Serrano, and Adam Zsolt Wagner: "Mathematical exploration and discovery at scale" arxiv.org/abs/2511.02864. Further discussion is at terrytao.wordpress.com/2025/11/05/m...

why intern at Ai2? 🐟interns own major parts of our model development, sometimes even leading whole projects 🐡we're committed to open science & actively help our interns publish their work reach out if u wanna build open language models together 🤝 links 👇

🆕 from us at #EMNLP: Are LMs better at answering questions about Germany in German than in French? Is national knowledge linguistically contingent? Interestingly, only for some multilingual models is this true. Aya knows China best in Chinese, but LLaMA's best in English always.

In fact already many people at CUNY doing this exact work! cc @mkgold.bsky.social @lukewaltzer.bsky.social @lmrhody.bsky.social etc

Digital humanities researchers often care about fine-grained similarity based on narrative elements like plot or tone, which don’t necessarily correlate with surface-level textual features. Can embedding models capture this? We study this in the context of fanfiction!

Our research downstreams into pedagogy! My talk on Language, AI, and Education has been featured by Texas Language Center. I present pedagogical approaches to creative writing, technical and qualitative interpretability techniques, and narrative capacities of LLMs. www.youtube.com/watch?v=_9Ql...

Skip the first section. The rest is quite helpful.

Gen AI & the novel: new report from @clemicollett.bsky.social examines how literary creatives are using (or not using) GenAI, how their work & income are affected, how this shapes their hopes & fears, & what measures govt might take to protect UK literary scene. Launch 20 Nov see linked post!

Very much looking forward to presenting at this tomorrow. I will be making my usual pitch that datasets are the foundational infrastructure for cultural heritage to benefit from and create useful AI models and tools. Be warned, I did fire up the meme generator for my slides...

Oil is the most Lovecraftian thing that actually exists.

I've seen an alternative argument going around that academics should instead step back and essentially see how things play out with AI, having confidence that if it sticks around (and doesn't go the way of the MOOC) it'll be adapted to academic needs and vice versa, like email, the internet +

I love Zohran y’all, but posting “Proven: politics not about compromise and we never have to compromise again” because the Democratic nominee won the election for … mayor of NYC … might slightly overstate available evidence? Still, 💕 and hugs.

Wonderfully sensible essay about Higher Education and AI by @tedunderwood.com.

As the cold reality of Cuomo’s defeat sets in, this is a moment for centrists to ask themselves: Have we lost touch with the common voter? Get out and spend time in the tattoo parlors and dispensaries. The drag shows, community gardens and outsider art galleries where Real New Yorkers congregate

"[I]f disciplines move proactively, they can dramatize their central role, and show that our campuses are not just lovely places to scribble in blue books but sites where students discover things that aren’t yet contained in any model." ♥️ Approach AI with curiosity and a little wonder, not dread.

What I liked about that speech was a) It was hopeful and forward-looking but also b) frankly angry about billionaires & corruption. Obama gave us (a) but this is a time in our history when you can't be honest without (b). Also c) seeing Mira Nair and aloof wife on stage at the end.

I needed this night for basic mental health tbh

ByteDance's Ouro (Open-weight) A family of pre-trained Looped Language Models (LoopLM) that instead build reasoning into the pre-training phase through (i) iterative computation in latent space, (ii) an entropy-regularized objective for learned depth allocation, and (iii) scaling to 7.7T tokens.

In the year one million and a half Humankind is enslaved by giraffe Man must pay for all his misdeeds As the treetops are stripped of their leaves

N-gram novelty is widely used as a measure of creativity and generalization. But if LLMs produce highly n-gram novel expressions that don’t make sense or sound awkward, should they still be called creative? In a new paper, we investigate how n-gram novelty relates to creativity.

"We need to be bolder: It is our duty to ensure library collections remain open to the public in a form that empowers 21st-century readers. If our intellectual heritage gets enclosed in proprietary tools, we will find ourselves making the same bad bargain we made with scientific publishers"

A very reasonable take on AI in education

Wrote a short piece arguing that higher ed must help steer AI. TLDR: If we outsource this to tech, we outsource our whole business. But rejectionism is basically stalling. If we want to survive, schools themselves must proactively shape AI for education & research. [1/6, unpaywalled at 5/6] +

Glad to see this commitment by Anthropic. Preserving models is a low-cost move that could have safety and welfare benefits. Hopefully we'll see other companies commit to this as well www.anthropic.com/research/dep...

Indeed, the whole point of building models is *to* illuminate mechanism, or more specifically, to test theories about mechanism(s), and see where and how they fail. LLMs have failed nicely, and are clearly not adequate models of human behavior, but have provided useful information en route.

I’m not sure existing methods will always suffice for this. But sometimes they may, and in any case we’re not limited to existing methods. It’s a good challenge.

My new field guide to alternatives to standard LLMs: Gated DeltaNet hybrids (Qwen3-Next, Kimi Linear), text diffusion, code world models, and small reasoning transformers. 🔗 magazine.sebastianraschka.com/p/beyond-sta...

Introducing OlmoEarth 🌍, state-of-the-art AI foundation models paired with ready-to-use open infrastructure to turn Earth data into clear, up-to-date insights within hours—not years.

The world is now full of more disgusting villains. Your masterworks of fraud and violence are forgotten, like trunkless legs of stone in a desert. Still, you played your part in getting us here: so this is for you, Dick Cheney.

feeling like the end of ongoing ai copyright wars. labs settling and if i read correctly, stability ai getting the most positive outcome from getty images case.

Every image this account posts would have sold me a paperback book in the 1980s. I would have been unable to resist these covers.

This essay from @bschmidt.bsky.social on how history rejected computational methods, & so "quantitative history" ended up in the social sciences, & "digital humanities" in literature, with no historians doing computational work, is fascinating, & worth a read: dhdebates.gc.cuny.edu/read/computa...

Interesting chart where service based sectors are using AI more (even though, e.g. the US has way less trust or optimism in AI than a place like China) could be a resounding advantage in a willingness to fund the endeavor as it gets even more expensive in the next couple years.

Halloween is about laughing at death. We send children off with treats because children are how we laugh at death. Death gets all of us, but it can't get us all.

It's almost as if letting the ultra-rich own the entire mainstream press and also all the big social platforms had a downside.

blocklists are an imperfect solution to the fact that the biggest problem of social networks is finding the content you want to see and filtering out what you don't. if you invented more effective tools to express preferences and coordinate voluntary socialization, they would have the same effect