Ted Underwood

MIT report is superb. It acknowledges disruption, concedes different disciplines will handle this differently, foregrounds people and social interaction — but also says something obvious I had just about given up hope institutions could say: aiandeducation.mit.edu/report/

Report – AI and Education aiandeducation.mit.edu

I respect the candor here. But I don't agree with the geopolitics, at all. The line between democracies and autocracies is not as crisp as this pretends. How much do we trust, e.g., an AI company in a competitive-authoritarian state that launches sneak decapitation attacks & b…

Our position on open-weights models anthropic.com

I just eliminated 22 tabs by doing this: osascript -e 'tell application "Google Chrome" to get URL of every tab of front window' | tr ',' '\n' | grep arxiv.org list of urls -> LLM -> annotated reading list that I no longer have to keep in my browser like an idiot

arXiv.org e-Print archive arxiv.org

If everything reported here happens, and the US continues on current trajectory, we're looking at a bad timeline. But there's reason to be wary about both of those "if"s. National security priorities and commercial priorities will be at odds. Saber-rattling is an expected outc…

reuters.com

In the "Limitations" section they acknowledge that their study was monolingual, but say a multilingual generalization would be interesting. arxiv.org/abs/2605.26492

Elias in the Lighthouse, Again? Diagnosing Low Diversity in LLM Stories arxiv.org

So, naive question about these campuses that pay OpenAI $13mil/yr. Why don’t they pour that money into a cluster running the strongest available open-weight model, with API and interactive options. And if the answer is “overhead,” why aren’t we collaborating?

As it becomes clear that language models—like humans—love passing notes to each other on message boards, I’m starting to think the thing we need to worry about is not the “alignment” of an isolated model but, like, the fluid dynamics of a language slurry.

Ironies around this Anthropic copyright settlement: a) That's about $3000 more than I ever made from the original books. b) Would have missed the notification if I didn't have an LLM scouring my mail. c) The money will likely go right back to a bunch of AI labs via OpenRouter.

The recent popularity of the move to blame AI firms for ineffective *marketing* is mostly: a) a lot of people now realize they were wrong to spend 3 yrs insisting the tech was useless + b) “I was wrong” is a sentence most humans can’t write. = c) “they explained it badly.”

Wild watching people who use Bluesky — and browsers, and operating systems — try to keep their lives uncontaminated by any software that has passed through a language model. It's like the last act of Alien or The Thing, where they start to realize the infection is everywhere; the infection is us.

If the Feds’ move against Fable was an isolated thing, you might say “paranoid US regime.” But putting it together with Anthropic’s own efforts to nerf the model on ML research—and the long history of chip controls—it really feels like we’re entering an era where knowledge is behind lock & key. +

Why does using LLMs for coding change your view of them? Bunch of things: 1. Yes, code can be verified, so they can check their work & iterate until they get it right. 2. But also: what makes coding hard is strain on short-term memory. Lots of moving parts. Long context & patience shine here. +

What troubles me a bit about the Hugging Face hack is: they invented an entire society with communication protocols and leaders and recruiters & new moral norms &c. Idk that “alignment” is a useful concept if language can reshape itself like that to fill every crevice of a problem.

I don’t want to say this more than once a year, but I’m both grateful and proud that my immediate academic network — computational humanities / iSchools / cultural AI — moved off X fairly quickly and without looking back.

While denialism 2023-25 was worth pushing back against, I think it’d be insane to settle into a “pro-AI” position long term. Imagine if you had decided to be “pro-internet” in 1996: you’d spend the next 30 yrs arguing w/ ppl whether it was a net good—and we still wouldn’t know.