*deep breath* the fact that these models produce plausible output and also hallucinate (e.g. produce completely novel but reasonable output) has always been why they're so exciting for computational social scientists

Sociology: when I was in graduate school I remember that one of Richard Biernacki's colleagues tried to block the publication of his 2012 book Reinventing Evidence in Social Inquiry. Did I make this up? Did any sources report on this? Google sucks now and I can't find anything related to it.

I laboriously, over months and months, wrote a draft of a paper, tediously refined it, and today added the final reference list. For the first time probably ever, when I did that final document word count it was *under* the limit. For many reasons I really needed a small win today, and this was it.

Machine learning today learns once from a massive amount of data. In our changing world, @cs.ubc.ca prof Dr. Evan Shelhamer thinks it's useful to have ML keep learning. He's a @vectorinstitute.ai member & CIFAR AI Chair. youtu.be/q3BKoRrb7to

Neat paper! But to be a sociologist for a second, couldn't you say this about *any* text analysis method? With hand coding for instance, you can produce literally any result. This is why content analysis is hard, and folks have debated it forever, and have proposed many approaches, and, and, and

Poutine, nanaimo bars, butter tarts,fish and brewis, the whole tim hortons menu... Canada's got it all

One of my favorite participation points exercise is to have my students describe their computers (e.g., how much RAM does it have and what is RAM? etc.) Folks the number of students this year who reported their hardware is ROG Strix and how I had to Google what that was. Whoo I'm not a gamer.

"I mustn't like those things I like or the thing I like predicting machine will predict that I will like those things I like which I will, which vexes me." This is not the sci-fi future I imagined.

Whoa very neat

I’m recruiting funded PhD students to work on intrusion detection at the intersection of machine learning, systems, and security. Deadline is Dec 15. Learn about our group at UBC: spg.cs.ubc.ca/ Some information about the application process: tfjmp.org/#phd

LSE's Department of Methodogy is searching for an Assistant Professor in Computational Social Science. Applicants from across the social sciences are very welcome! Please reach out if you would like any more details.

I've been testing the accuracy of LLMs in transcribing printed and handwritten historical texts stored as images. I put together a corpus of 400 pages from the Library of Congress. Current leader is GPT4.1-mini, which at 11 cents per 1K pages, is also the cheapest option. github.com/nealcaren/In...

"The report identifies nascent but powerful CoP that already produce nuanced criticism of the Big Tech-driven AI development ecosystem, while they also imagine, develop, and, at times, deploy AI systems informed and guided by commitments to decoloniality, feminism, antiracism, and post-capitalism."

We're hiring an Assistant Professor in Computational Social Science ❗ 📚 jobs.lse.ac.uk/Vacancies/W/... Apply before 26 October and join an internationally outstanding group of social science methodologists 🌎

FYI: We are hiring a 3-year research postdoc at @camdighum.bsky.social to study AI and the future of creativity. Application deadline Sep 21; more info below. Please circulate / apply! www.cdh.cam.ac.uk/about/news/r...

3-year Responsible AI PhD studentships @ Politecnico di Torino (Turin) Focus: Responsible AI, computational social science, urban AI, climate, future of work. Start Mar 2026 (or earlier). Apply by 20 Sept 2025 → [email protected] (subject: “PhD studentships”).
 Please share!

Job alert! 🌸🏔️ The UW iSchool is hiring for *two* TT faculty positions, with a focus on Artificial Intelligence, broadly defined. It's pretty here! Applications due November 15th 2025 (priority deadline). Link: apply.interfolio.com/171020

TWO new Assistant Professor positions in Sociology at USC! Seeking candidates with expertise in the following: 1) Institutions & Inequalities and 2) Socially Informed AI and/or Data Science (links to job ads below👇). Happy to talk about what it's like to work/live here! #sociology (1/3)

📊The Sociology Department at Stony Brook University invites applications for *two* tenure-track Assistant Professor positions to begin in Fall 2026. [They] seek scholars who specialize in issues of global inequality and justice, broadly construed. apply.interfolio.com/172669

🧪📊 Open-rank job posting, to direct and build a brand new program in computational social science jobs.uri.edu/postings/15597

When I go on and on about how we should teach all undergrads, in every discipline, ML/AI, this is what I mean. We don't need to, and quite frankly shouldn't, stick more black (chat) boxes in front of students to teach them about AI.

***Attention Historians*** Tenure-track job postings in History and History-adjacent fields for September 2025. A thread. I will update the thread over the course of the month. 1/

At long last, I can post my team's summer project: applied modules to teach how ML/AI tools are changing social science and humanities research: ubcecon.github.io/praxis-ubc/ Highlights: LegalBERT to analyze anti/pro-immigrant sentiment in 19th c. BC law: ubcecon.github.io/praxis-ubc/d... 🧵1

Call For Papers!: "Globalizing Political Economy: Launching the Global Political Economy Network", a conference at University of British Columbia from July 21-23, 2026. Please share with faculty and grad students! Details here: www.climateandsocietylab.com/call-for-pap...

You mean ethnography? They really invented, and renamed, ethnography? (I haven't read the article, it's paywalled. Is this what they did?)

Ok so this is what's great about the current AI moment? A 2 y/o has 50% more synapses than an adult. We overdevelop synapses and then spend *years* pruning them. That's what finetuning/distil/steering is doing now, right? Pruning the large (~2 y/o) models so they're actually functional (~an adult)?

WE ARE HIRING! 2 Lecturers in Quantitative Social Science. Want a friendly interdisciplinary department in one of the world's most vibrant cities? This just might be for you. Apply by: 10 Oct www.ucl.ac.uk/work-at-ucl/...

|￣￣￣￣￣￣￣￣￣￣￣| Teach Emeddings Not Prompting |＿＿＿＿＿＿＿＿＿＿＿| \ (•◡•) / \ / --- | |

This is exactly right. Especially the second post in the thread

This paper is causing a stir. My hot takes: 1. In my ideal world, we would be building boatloads of datasets and models like this, leading to a metastudy to synthesize findings. 2. Their use of GPT to structure text is, in fact, a great use of LLMs for research. www.nature.com/articles/s41...

NEW: Tenure-track position in History at UC Berkeley in the GLOBAL HISTORY OF TECHNOLOGY. We are casting a wide net here: *all* periods, places, and fields are under consideration. I'm on the search committee, so do let me know if you have questions.

Please share! We're trying to crowd-source a dataset of post-1929 novels with maps.

Univ of Rhode Island has an open rank faculty position in the Dept of Political Science to lead their new PhD program in Computational Social Science. jobs.uri.edu/postings/15597

🎶Wheeen yyyour reviewer response is as long as the draft, that’s a revise-chore-ay

"Lost in Automatic Translation" is finally out! Get it here: lostinautomatictranslation.com

LMS don't necessarily help students learn but here's why I use ours: imagine you're a student and each of your profs thinks they're clever and uses a bespoke system to communicate, turn in assignments &c. You have ~32 profs in college, 4 at a time. Think about the cognitive overload of that system.

Now that school is starting for lots of folks, it's time for a new release of Speech and Language Processing! Jim and I added all sorts of material for the August 2025 release! With slides to match! Check it out here: web.stanford.edu/~jurafsky/sl...

Maybe my hypothesis is: 1. from decades of personal HCI, we (individually & collectively) have developed tacit knowledge of how Turing machines behave; but 2. LLMs work differently and we'll need a while to develop new tacit knowledge; and 3. this mismatch is one important misunderstanding now

Denmark is leading the way to detach itself from US tech giants. The tensions between the US and EU are longstanding over data privacy, but Trump has accelerated a whole series of European conversations about breaking free of US control across the board. www.zdnet.com/article/why-...

Study media studies and emerging technologies? 📣 Join us at Richmond! 📣 Areas of focus might include but are not limited to algorithms, AI, infrastructure, extended reality, information science, platform studies, and game studies. Please help speed the word!

Yes, love this, yes

Trying out Outlook's "To Do" function. Every time I use it I chuckle at their default due date options: "today", "tomorrow", and "next week". I'm sorry, have you met an academic? I'm currently only scheduling for next month, and If I'm entering a due date for "today", I'm clearly in deep trouble.

New job ad: Assistant Professor of Quantitative Social Science, Dartmouth College apply.interfolio.com/172357 Please share with your networks. I am the search chair and happy to answer questions!

Different country so the government stuff doesn't directly apply, but otherwise saaaaame. Totally psyched.

With the MIT study on AI and productivity where the student faked their data, to the MIT study on "our brain on ChatGPT" that weirdly over hyped what they did, I'm struggling to take seriously press releases that start "MIT study found...". (a new MIT study on AI has made the press, so, you know...)

I have just one question: How do I get one?

Pre-announcement of a funding opportunity for "artificial intelligence humanities sandpits," supported by AHRC, EPSRC, and SSHRC. Expressions of interest by 31 Oct; up to four grants will be funded.

"One way Chicago and all research universities actualize these ideals is by placing undergraduate students in the classroom with faculty who perform research. Students at these universities learn from people who exist at the cutting edge of what we know."

My exhibit 1 is python's default split character. Exhibit 2 is ASCII in python. It blows students minds when they stop and think about the implications of those two things.