Pinned
Thrilled to be featured (along with a few of the many rising stars of digital history) in this week's PNAS Front Matter: pnas.org/doi/10.1073/... "Historians Use Data Science to Mine the Past"
Jo Guldi
@joguldi.bsky.social
Professor of Quantitative Methods at Emory. Data scientist, writer, historian. Newest: The Long Land War (2022), The Dangerous Art of Text Mining (2023).
1.4K followers308 following149 posts
Pinned
Thrilled to be featured (along with a few of the many rising stars of digital history) in this week's PNAS Front Matter: pnas.org/doi/10.1073/... "Historians Use Data Science to Mine the Past"
For the first time: the culmination of 15 years of digital-history experiments—and what they teach us about navigating archives with AI. 1:30 ET, Baltimore + livestream: proflouishyman.github.io/ai_conferenc... @duttonbooks.bsky.social preorder: bit.ly/WarOverRent
Irish tenants used maps, ledgers, and court records to challenge rising rents. Their evidence helped create the world’s first Land Court. That history leaves a durable question: who controls the record? joguldi.substack.com/p/coming-soo... Subscribe for more essays.
Are we living in an “age of AI”—or deciding to call it one? Harvard Law asks historians how eras are made, why diffusion matters, and how historical memory cuts through hype. clp.law.harvard.edu/knowledge-hu...
AI didn’t save me from work. It replaced repetitive Canvas labor with cumulative intellectual work. Lesson for universities: use machines to reduce paperwork—and the savings to bring humans back into the room. computationalhistory.substack.com/p/how-ai-sav...
Well, this was dangerous. A preorder sale led me to @joshspecht.bsky.social on property, @jeannasmialek.bsky.social on Maria Edgeworth, and Andrew Whelan on destroyed records and power. @duttonbooks.bsky.social’s THE WAR OVER RENT connects all three. PREORDER25
Turning Points is out—and our Trento conversations are in print. My chapter asks: Was that anecdote typical? What did the archive omit? The digital turn didn’t make interpretation obsolete. It made its problems newly visible. www.degruyterbrill.com/document/doi... @toffolosandra.bsky.social
The idea that rent should simply go up and up is much newer than we tend to imagine. The last time it was promoted with abandon, it helped bring a country to the brink of revolution. Cover reveal today! joguldi.substack.com/p/the-rent-w...
As America approaches its 250th anniversary, we’re asking whether historians can tell a common national story. I think the better question is: Where do historians actually agree? #America250 #AI #History #HistoricalConsensus open.substack.com/pub/joguldi/...
What if climate change, AI, democratic decline, and inequality are all versions of the same problem? My latest essay: open.substack.com/pub/joguldi/... (computationalhistory.substack.com)
I'm so excited to announce, with Jim Clifford, a Working Paper Series on Critical Search. We'll be publishing our own and others' experiments with AI, design, and pedagogy in a format built to reward quick turnaround and rapid ideation -- forging a track from lab to the world of journals.
What if the first great generation of data activists were not in Silicon Valley, but nineteenth-century Ireland? open.substack.com/pub/joguldi/...
Historians warned against cherry-picking. But now others are making big historical claims at scale—and shaping public debate. We need a response. Critical search. Disciplined selection. computationalhistory.substack.com/p/beyond-che...
AI’s hardest problems — retrieval, ground truth, alignment — aren’t just technical. They’re historical. The future of trustworthy AI may depend on historical methods. computationalhistory.substack.com/p/ai-needs-h...
Digital history didn’t fail; it stopped short. It built archives, not cumulative research programs. Now AI is asking the questions DH left unresolved. Computational history is what comes next. computationalhistory.substack.com/p/the-lost-p...
Jo Guldi reposted
I can't stress enough how studying philosophy at a liberal arts school prepared me for my career. Logic problems, organizing abstract concepts, neural nets and models of cognition, language and conversation, the nature of experience. Honestly expected never to see a trolley problem again tho.
Jo Guldi reposted
Jo Guldi reposted
This is cool. Johannes Uhl and colleagues use historical maps and settlement data to generate a dataset of the approximate date of every road in the United States. Check it out: arxiv.org/pdf/2506.16625 @jhuhl.bsky.social
Jo Guldi reposted
So. Less ability to forecast hurricanes, and a crippled FEMA so you can’t get federal aid for the damage…
A huge blow was dealt to hurricane forecasters this week as a critical tool was abruptly terminated by the Department of Defense and NOAA. The immediate discontinuation of data from three weather satellites will severely impact hurricane forecasts this season and beyond. More ⬇️
Now if only they would publish the data on their findings so that we could understand the work of government and the sacrifices of public servants. Because every past government committee in a democracy committed to transparency publishes its findings, right??
“I personally was pretty surprised, actually, at how efficient the government was,” Sahil Lavingia told NPR's Juana Summers.
Profound personal statement about the power of the humanities, the risks associated with losing the university as a bridge to social mobility, and how much is at stake as we abandon the infrastructure of research. Thank you for writing this, Karin. We need the long form version.
"My father, an early and prominent computer scientist, passed away more than two years ago. Since then I’ve been trying to make sense of how rapidly the world that helped make him, and in turn the one he helped shape, is unraveling." 1/ scholarlykitchen.sspnet.org/2025/06/26/d...
Jo Guldi reposted
I wish more of my neighbors in Western North Carolina realized this.
Trump has gutted FEMA grants to build storm-resistant infrastructure in flood and hurricane-prone communities. Two-thirds of counties that have lost funding voted for him in 2024. Trump is even telling his own voters they're on their own. https://www.cbsnews.com/femagrantcuts/
I do love that it was helpful to a China scholar, Tom! I’ve been hearing the same thing from sovietologists as well. Thanks for weighing in!
Fascinating interview with @joguldi.bsky.social novaramedia.com/2025/06/13/h... part near the end on going beyond communism vs capitalism to classify modes of development during cold war is especially relevant to China
Super powerful Sunday meditation on America in despair by theologian Diana Butler Bass @dianabutlerbass open.substack.com/pub/dianabutle…
Listen to my new podcast/interview with the brilliant young British folks at NovaraFM, where I talk about participatory mapping, land reform, and their fascinating histories -- "How to Redistribute the Globe" novaramedia.com/2025/06/13/h...
Jo Guldi reposted
Jo Guldi reposted
Jo Guldi reposted
Check out the handout from the AHA’s Congressional Briefing on the history of deportation. Panelists @hidehirota.bsky.social, Ana Raquel Minian Andjel, & @yaelschacher.bsky.social shared policies concerning immigration & deportation throughout US history. 🗃️
My second talk at the University of Glasgow will connect the history of land politics to affairs in Scotland: “The Long Land War: Scotland and Ireland in the longue durée” eventbrite.co.uk/e/jo-guldi-t... (alt text: The Long Land War, 1881-1990)
I'm thrilled to be giving a series of lectures at Uni Glasgow in a few weeks! Here's one: "Will AI Kill Democracy or Spur a Democratic Revival?" eventbrite.co.uk/e/jo-guldi-w... (alt text: human silhouette with a question mark, robot head, and the slogan, "Crisis of Fact")
Jo Guldi reposted
AHA members @joguldi.bsky.social, Peter Roady, @laraputnam.bsky.social, @kalanicraig.bsky.social, Arlene Díaz, & Katherine McDonough were featured in an article in PNAS, one of the most-cited & comprehensive multidisciplinary scientific journals, on how historians use data science to mine the past.