Lucas Poy

Looking forward to presenting my work at @weatherheadcenter.bsky.social in Harvard next spring!

wigh.wcfia.harvard.edu/event/global...

A thoughtful essay worth your time. I’ve been wrestling with the same questions and have much still to learn.

yes, exactly. Of course this has the problem of proportions, not always is possible to have such a ratio teacher:students, but I think there are alternatives that work. For instance convert a 10-15 people class into a "workshop" in which students "defend" their work in a roundtable format.

how do Oxford tutorials work?

sí

Yes, during my BA in Argentina most of the final exams were oral. I wouldn't say I enjoy them but the process of studying for them was certainly useful.

Agree! My essay is quite limited in scope to my own experience teaching in university in the field of history.

you are the overwhelming minority!

The last days of the take-home essay as we knew it “Not to laugh, not to weep, but to understand”.

I wrote about the take-home essay and whether it has a future in the age of AI. What are we actually testing when we assign them? What alternatives do we have? You can read my thoughts here: lucaspoy.substack.com/p/the-last-d...

sí, es que me parece la explicación más probable. Otra posibilidad es que haya sido un chat muy largo en el cual dejó de usar el contexto grounded del RAG y se basó solo en el LLM.

Rent Strikes Since the nineteenth century, working-class families have predominantly relied on tenements for housing, with rents often consuming a large portion of their household budgets. There is a long and cont...

Ever wondered how tenants organized across different continents and periods?

𝑅𝑒𝑛𝑡 𝑆𝑡𝑟𝑖𝑘𝑒𝑠: 𝐴 ℎ𝑖𝑠𝑡𝑜𝑟𝑦 𝑜𝑓 𝑐𝑜𝑙𝑙𝑒𝑐𝑡𝑖𝑣𝑒 𝑡𝑒𝑛𝑎𝑛𝑡 𝑎𝑐𝑡𝑖𝑜𝑛𝑠 𝑎𝑐𝑟𝑜𝑠𝑠 𝑡ℎ𝑒 𝑤𝑜𝑟𝑙𝑑, edited with @hannesrolf.bsky.social, is coming November 2025!

Preview the table of contents and contributor details: uclpress.co.uk/book/rent-st...

es raro... será que el OCR que hizo no es bueno y reemplazó lo que no se leía bien por cosas "plausibles" pero no reales? Probaste subir el mismo PDF de nuevo, pero con un OCR bien hecho con otra herramienta?

pero la in-line citation te tiene que dirigir a uno de los PDFs que subiste, o no?

Using NotebookLM for research with digitised historical newspapers This is the third post in a series of reflections about how historians can engage with generative AI.

I wrote a post in my Substack about experimenting with generative AI for historical research with digitised newspapers. It looks at how it works with structured sources, where it helps, and where it falls short.

lucaspoy.substack.com/p/using-note...

Using NotebookLM for research with digitised historical newspapers This is the third post in a series of reflections about how historians can engage with generative AI.

New post on Substack: how I’ve been testing NotebookLM with digitised newspapers. In short: great for discovery (much better than keyword searching), but weak and risky for analysis.

lucaspoy.substack.com/p/using-note...

Using NotebookLM for research with digitised historical newspapers This is the third post in a series of reflections about how historians can engage with generative AI.

New post on Substack: how I’ve been testing NotebookLM with digitised newspapers. In short: great for discovery (much better than keyword searching), but weak and risky for analysis.

lucaspoy.substack.com/p/using-note...

Gracias!!

These guiding principles of the AHA seem written a year ago, and by people who don't really use nor know much about AI tools.

I wrote about it here substack.com/home/post/p-... and here substack.com/home/post/p-...

De American Historical Association bakt er blijkbaar niets van als het gaat om AI. De reacties uit de Anglofone historische wereld die hier voorbij komen zijn voorspelbaar opgewonden en polemisch: 'AI is de duivel'. Ik wijs dus nog maar eens op de inzichtelijke blogs van @lucaspoy.bsky.social.

I'm quite shocked about the sort of echo chamber that Blue sky seems to be: a cacophony of people who don't know very well what they are fighting against. I hope I'm wrong!

The focus on hallucinations misses the point that RAG tools like NotebookLM don't hallucinate.

These guiding principles of the AHA seem written a year ago, and by people who don't really use nor know much about AI tools.

I wrote about it here substack.com/home/post/p-... and here substack.com/home/post/p-...

Historians: We are having the wrong conversation about AI This might be the first in a series of posts exploring how historians can engage with AI tools.

Are we framing the AI discussion wrong in history departments? I think we're too focused on student “cheating” while missing the bigger picture. I've been thinking about this for a while and decided to write it down.

lucaspoy.substack.com/p/historians...

Proofs!

NotebookLM: The end of reading as we knew it? This is the second in a series of posts exploring how historians can engage with AI tools.

Some more thoughts about generative AI and our work as historians (in particular about using NotebookLM for secondary sources)
lucaspoy.substack.com/p/notebooklm...

First piece I’ve read where I see *possible* academic value coming from an “AI” tool. Not saying I’m a convert by any means, but definitely going to give NotebookLM a try.

“del éxtasis a la agonía oscila nuestro historial, podemos ser lo peor o también lo mejor con la misma facilidad”

I agree, but I still think that most students do want to learn

NotebookLM: The end of reading as we knew it? This is the second in a series of posts exploring how historians can engage with AI tools.

Some more thoughts about generative AI and our work as historians (in particular about using NotebookLM for secondary sources)
lucaspoy.substack.com/p/notebooklm...

Interesting post about AI tools, and useful reminder that a Large Language Model (LLM) "...is not designed to be accurate but to be plausible. It has no concept of truth or falsehood, only of linguistic patterns"