Teddy Roland

@teddyroland.bsky.social

Postdoc @ School of Information Sciences, UIUC American Literature, Media Theory, Data Science I publish under "Edwin Roland" but don't tell anyone.

I would just add that, from where I sit, this doesn't look like turf that can be defended or a tide that can be held back. There's already interdisciplinary collaboration on this topic and we can all work together to ensure the work is done well.

Federico Pianzola@fpianz.eurosky.social · last mo.

some of us saw this coming: computer scientists are doing a lot of research on fiction and narrative. 💥 63 papers + 2 SemEval tasks out of 6,422 publications in the latest @aclanthology.org Full list: federicopianzola.me/blog/index.h... #ComputationalHumanities @tedunderwood.com

Looking for something to do this summer? We have a new batch of documents up at our Zooniverse site, Corresponding with Quakers. We've added materials from the Huntington's Shackleton Correspondence to UCSB's Ballitore Collection and we need help transcribing! www.zooniverse.org/projects/rac...

Corresponding with Quakers | Zooniverse - People-powered research

Investigating race, gender, class, and religion in 18th- and 19th-Century Irish Quaker Documents

zooniverse.org

Conservatives too dislike UaaS and no one knows what to do "[To] enact the ideals of the civic-education movement, you wld need more than a new college: you wld need to reimagine the entirety of our education system and the society into which students graduate [...] a whole new ideal of the person"

Asheesh Kapur Siddique@asheeshks.org · 2mo ago

Everyone needs to read this before handwringing about ‘declining humanities major’ or getting fascism-curious about a GOP or Koch-brothers funded effort to install a civic center on your campus: harpers.org/archive/2026... 🗃️

New work from my lab! @teagrjohnson.bsky.social built a 12-dimensional narrative framework, annotated Dolma (no small feat given its extreme diversity), and analyzed narrative features across pretraining subsections. Highlight: pretraining data space displays strong narrative organization!

Figure 1: A web passage scored across our 12 narrative dimensions. Agency and setting dimensions are rated on a 1–5 Likert scale, temporal sequencing and causal density are passage-level proportions (0–1), and event density is the rate of event triggers per token. This passage scores high on agency and event features but low on setting, a “narrative profile” commonly seen across first-person web narratives.Figure 6: UMAP reduction of SBERT embeddings for 20,000 randomly sampled NARRADOLMA documents, colored by PC1 score (interiority). Labels are based on manual examination. Overlays for all three PCs appear in Fig. A5.
Teagan Johnson@teagrjohnson.bsky.social · 2mo ago

1/ LLMs learn narrative from their pretraining data but what narrative content is actually in there? It turns out narrative is wildly unevenly distributed across sources and topics. New preprint with @andrewpiper.bsky.social @elliottash.bsky.social @mariaa.bsky.social:

This image depicts the proportion of each Dolma category in the top quartile for the first three principal components.

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?

I can confirm that using claude to write rust works pretty well. RustMallet is a whisker faster than Java Mallet at training topic models, and adds some tweaks that I have avoided in Mallet for stability reasons (multiple saved states by default). github.com/mimno/RustMa...

GitHub - mimno/RustMallet: Rust implementation of sparse Gibbs sampling LDA, following the SparseLDA scheme from MALLET

Rust implementation of sparse Gibbs sampling LDA, following the SparseLDA scheme from MALLET - mimno/RustMallet

github.com

Even constraining the AI impact to the individual layer however, will change the nature of work. More experimentation and more exploitation of novel solutions for instance will change the complexity of code committed if not the quantity. Some evidence for this coming out of public repo analysis too

The AI encyclical is bringing something to a head for me, as someone who came up through the Catholic school system. In a nutshell, I think that a lot of cultural critics -- in AI and other fields -- rely on mystification to make their arguments just as much as any religious dogma does.

Wow!! An undergrad I mentored this spring had a paper accepted at ASIS&T. Really really impressed by their curiosity with the project and motivation to turn it into a formal paper. These kids!

My tiny thread yesterday is now a whole op-ed! A quick read, connecting Pope Leo's apology for colonialism and slavery to his warnings about AI — it's all related. Working with editor @mpgphd.bsky.social was an absolutely pleasure!

Religion Dispatches@religiondispatches.org · 3mo ago

the latest from RD: @roopikarisam.bsky.social on what makes Leo XIV's first encyclical so radical -- tracing AI back through histories of enslavement and exploitation, and repenting for the RCC's complicity in both religiondispatches.org/2026/05/26/l...

The annual Academic Jobs Report by the AHA is here. For me, the two big takeaways are (i) a roughly 15% drop in the total number of History faculty since 2010, and (ii) a 20% drop from the previous year in tenure/tenure-track job ads in last year's hiring cycle. Not irreversible, but pretty grim.

The 2025 Academic Jobs Report – AHA

The latest available data on higher education hiring proves that history departments are shrinking.

historians.org

Good news: the formula this story reports is about state funding only. Long term structural change is possible! Bad news: Federal funding got bumped too during COVID and it goes away next fall. OUSD is scrambling to cover enrichment programs and teacher aids, which impact current students.

The Oaklandside@oaklandside.org · 4mo ago

California’s spending is also now ranked as one of the most equitable in the nation. The data, from the 2022-2023 school year, reflect Covid spending boosts.