Mihai Codreanu

@p_e_t_e_r_s_e_n LLMs (cloud.google.com/use-cases/ocr)

What are people using for OCR of historical documents these days?

Today, we’re continuing to push the boundaries of AI with our release of Gemini 3.1 Pro. This updated model scores 77.1% on ARC-AGI-2, more than double the reasoning performance of its predecessor, Gemini 3 Pro. Check out the visible improvement in this side-by-side comparison, showing Gemini 3.1 Pro’s crisp animation built with pure code. Read more about today’s 3.1 Pro update: blog.google/innovation-and…

The first cohort of Stripe econ of AI fellows: (+thread, with some thoughts on the state of the field)

The past 18 months have seen the most rapid change in human written communication ever By. September 2024, 18% of financial consumer complaints, 24% of press releases, 15% of job postings & 14% of UN press releases showed signs of LLM writing. And the method undercounts true use

🚨 New research: We analyzed 1.5M+ documents to track LLM-assisted writing adoption across society from 2022-2024. The results? 📊By late 2024, LLMs assist in writing: - 18% of financial consumer complaints - 24% of corporate press releases - Up to 15% of job postings (esp. in small/young firms) - 14% of UN press releases All showing similar adoption trajectories despite diverse contexts. PDF: arxiv.org/abs/2502.09747 🧵

@andrei_mntn @fabolange Congrats, Andrei!!! :)

Thank you @fabolange (and others) for the advising leading up to this. I came to econ very late and it was my first-ever applied micro paper (started 4th yr PhD). Not a humblebrag. I only mean that I had a lot of ground to make up in little time and could never have done it alone

@isadesisto This looks like a very interesting project!! You should try the Surya OCR open source model. twitter.com/VikParuchuri/s…

OCR & historical data gurus: I have scans of biz directory for interwar Romania. I want to digitize the data for Bessarabia (Moldova) to look @ economic conditions & class composition of villages. Anyone have ideas abt how to do this w/o hand typing? Ex attached & pls share!

@JacobCConway @ChicagoBooth Congratulations, Jacob! 🎊

I'm thrilled to share that I'll be joining @ChicagoBooth as an assistant professor! I am grateful for the incredible support I've received from my family, advisors, and friends over the years, and am very excited for this next step.

Entrepreneurship PhD students: apply for the 2024 @nberpubs Entrepreneurship Research Boot Camp (week of July 15th). My student experience 10+(!) years ago was instrumental to my career. Thanks to @NBER and @robinson_fuqua for keeping it going. conference.nber.org/confsubmit/bac…

"Rationalizing Entrepreneurs’ Forecasts" by Mihai Codreanu @m_codreanu Nick Bloom @I_Am_NickBloom & Robert Fletcher @StanfordEcon @Instacart

@vlad_surdea Nice work!!

The consequences of mass repression remain ambiguous, with strong evidence for both inhibitory and mobilising effects on future dissent, depending on the political opportunity structure. However, there is virtual consensus that high-threat environments constrain dissent.

Sorry to learn that Max Corden has passed away. His book "Trade Policy and Economic Welfare" was an inspiration, and the idea of effective protection is a lasting part of the trade canon

Conference on Field Experiments in Strategy (CFXS) is off to an amazing start @HarvardHBS with @hyunjinvkim walking through her design choices for our PhD workshop.

@M_De_Nardi Took a while & a bunch of phones and emails but yes!! Hopefully some reimbursement will come too :-)

@m_codreanu What a nightmare! I hope you got your lost luggage in the end, at least?

@felicia_ionescu @TurkishAirlines It more than did!! Such great organisers in @MCopaciu and Cristian Litan (and also amazing photographers!!)

@BBCNews What wonderful news!!! Congratulations Prof. @DianeCoyle1859!

Ian Wright, Davina McCall and Ken Bruce are among the famous faces in the King's first Birthday Honours bbc.in/3PfXF4a

@MarkDVerhagen Thanks, Mark!! Looks super cool, and relevant, added on my reading list :)

@m_codreanu Cool stuff @m_codreanu ! We once did a similar descriptive exercise using Scopus data (so excluding preprints) and by discipline. Let me know if you'd be interested to learn more!

1/4 Will AI get better over time? Besides more compute, and data/user feedback, there is one additional reason: ↑ research Using ArXiv data, we see exponential growth in the number of papers in AI & related topics. AI growth rate is significantly surpassing non-AI CS related

4/4 Came across this part of a project on academia-industry ties in @erikbryn AI class (w/ Aguilar, @jdlight, Dias) and my wider agenda w/ @theJohnBonney Also check out my related paper w/ @I_Am_NickBloom, Davis, Zhestkova on WFH endogenous technical change More updates soon!

3/4 Is this research "crowding out" old methods? For now, let's look at NLP "emerging" (e.g. contextual embeddings), and "declining" methods (e.g. bag of words) There doesn't seem to be changes in trends for research on conventional thought "declining" methods. At least for now!