Lightnews — Scholar-powered news

Ted Underwood

@tedunderwood.com

20K followers 5.6K following 18K posts

Uses machine learning to study literary imagination, and vice-versa. Likely to share news about AI & computational social science / Sozialwissenschaft / 社会科学 Information Sciences and English, UIUC. Distant Horizons (Chicago, 2019). tedunderwood.com

Posts Media Videos Starter Packs

Pinned

Ted Underwood @tedunderwood.com · Jun 25

New this morning, a Comment I contributed to Nature Computational Science on the interaction between large language models and the humanities. 🧪 🤖 #MLSky

rdcu.be/etk07

The link above will be open-access for a month — plus, I'll reply to this post with a link to a permanently open preprint. +

The impact of language models on the humanities and vice versa

Nature Computational Science - Many humanists are skeptical of language models and concerned about their effects on universities. However, researchers with a background in the humanities are also...

rdcu.be

14 53 160

Ted Underwood @tedunderwood.com · 1h

It looks like it would have been fun to ride, but the tipping risk makes that probably a no-go

Ted Underwood @tedunderwood.com · 2h

I think in the 1980s I may have seen someone towing a suitcase (long side down) with a strap. But it wasn't common. The vertical thing (short side down) with extensible handle is much easier. It seems to have been a late 80s invention, but I don't think I had one until the late 90s if then.

Reposted by Ted Underwood

Richard Morris @ahistoryinart.bsky.social · 7h

One critic, writing in 1911 (the date of this work) commented, George Clausen 'gives us out of very simple elements the most perfect nocturnes, these mark a new and definite advance in his power in taking complete hold of a subject and making it his own.'

3 17 100

Ted Underwood @tedunderwood.com · 7h

More often balancing a phone in my lap or glancing over at it on the passenger seat. All bad strategies.

Ted Underwood @tedunderwood.com · 7h

Wheeled luggage and Google Maps would be very high on my list.

A big improvement in the past three months, for me, is one of those phone holders you can attach to your dashboard so you don't have to try to hold/touch the phone while using Maps if driving solo.

2 3

Reposted by Ted Underwood

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 7h

Lot of people should read this exhaustive list of life improvements from the 90s: gwern.net/improvement

My Ordinary Life: Improvements Since the 1990s

A list of unheralded improvements to ordinary quality-of-life since the 1990s going beyond computers.

gwern.net

5 8 47

Reposted by Ted Underwood

Nathan Lambert @natolambert.bsky.social · 8h

Recurring frontier lab gossip:

OpenAI has best post-training/rl and has pushed it super hard on weaker pretraining.

Gemini has spectacular pretraining. Making a reasoning model was super easy for them & OpenAI folks were surprised

Anthropic? Secretive i guess.

1 1 42

Ted Underwood @tedunderwood.com · 9h

I can report that life feels different in fields that do have progress narratives. There's a lot of discipline-level chaos in information science (about which subfields deserve attention). But within the subfields I know best there's often a pretty clear sense of recent achievements & next steps.

1 3

Ted Underwood @tedunderwood.com · 10h

I think this is one of the answers for information science, but it doesn't cover CS. www.ischools.org/news/categor...

Jobs

www.ischools.org

Reposted by Ted Underwood

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 11h

One strategic thing about making this an appealing scientific community is to overshare and boost work from graduate students here

3 9 53

Reposted by Ted Underwood

chenhaotan.bsky.social @chenhaotan.bsky.social · 12h

What is the best way to find faculty job advertisements for CS/info students on the job market? It seems that students in my group are spending an unreasonable amount on this. I thought CRA is good, but seems not active any more.

3 3 9

Ted Underwood @tedunderwood.com · 11h

Not strongly? I guess Elvis is importing a genre to a new audience; the Beatles kind of invent "boy band," and then experiment with it.

TS is much more strongly about her own personal story, and I'm not sure the claim I would make for her is founded on formal or generic experimentation.

1 1

Ted Underwood @tedunderwood.com · 11h

We don't talk about the elder ones, Adam

1 6

Ted Underwood @tedunderwood.com · 12h

This is an excellent picture of how you'd build a research ecosystem if you didn't want AI to turn research into a winner-take-all game where only the biggest players have any chance of making recognized contributions.

It requires a lot more collaboration than most academics are used to.

chenhaotan.bsky.social @chenhaotan.bsky.social · 1d

This resonates very well. I think it should apply to all disciplines. I also really liked this recent piece by Channing and Ghosh: arxiv.org/abs/2509.06580

1 3 27

Ted Underwood @tedunderwood.com · 12h

Yep, this became a sentence in something I'm writing. Thanks again.

Ted Underwood @tedunderwood.com · 12h

Agreed. OP is trying to be analytical, or at least admit my own failure of objectivity. But basically I am a fan.

Ted Underwood @tedunderwood.com · 13h

"This season's colours are unlike anything in the normal spectrum. Indeed, it is only by analogy that we call them colours at all."

1 4

Reposted by Ted Underwood

Hannah Walser @hkpmw.bsky.social · 13h

Nordstrom Rack at the mountains of madness

A gently glowing blue sign on the wall of a department store reads “prices below reason”

4 9 46

Ted Underwood @tedunderwood.com · 14h

Yes, Christopher Newfield's mode of institutional engagement is also an important unifying project.

Both that project and the response to AI do feel a bit post-apocalyptic. It's like "we're working on locating new seed crops and finding some substance that can kill the Shamblers."

1 2

Ted Underwood @tedunderwood.com · 14h

Yeah, I didn't want to say it because it's an ego-centered interpretation, but articulating a rationale for resistance to AI does seem to be the unifying project.

1 5

Ted Underwood @tedunderwood.com · 14h

getting beyond that in a good, self-consciously restrained way, it could be positive. Unfortunately I suspect it's getting beyond it more in a Children of Men way.

1 8

Ted Underwood @tedunderwood.com · 14h

I've been wondering if I was just out of the loop, but yeah: I have no idea what theoretical questions organize debate in literary study right now.

That's not necessarily bad? The hypertrophy of theoretical branding (already going strong in the 90s) wasn't super productive. If the field were +

2 12

Ted Underwood @tedunderwood.com · 15h

Yeah, I don’t know. I’m struggling to make sense of it

1 4

Ted Underwood @tedunderwood.com · 15h

Jeez. You don’t pay that kind of money to get the guy, you pay it to kill the competition.

1 6

Ted Underwood @tedunderwood.com · 15h

It doesn’t get worse than malevolent recursive JSON.