Dan Björkegren
@dbjork.bsky.social
1.1K followers 170 following 24 posts
Digital data/machine learning/economics, focused on developing economies. Faculty at Columbia. dan.bjorkegren.com
Posts Media Videos Starter Packs
Pinned
dbjork.bsky.social
New working paper: Could AI leapfrog the web?

Only 37% of sub-Saharan Africans use the internet. Cost is the #1 constraint. In a study with 469 teachers in Sierra Leone, we find AI works better than the web--and is 87% cheaper.

arxiv.org/abs/2502.12397

Let's dive in... 🧵(1/X)
Reposted by Dan Björkegren
davidduvenaud.bsky.social
It's hard to plan for AGI without knowing what outcomes are even possible, let alone good. So we’re hosting a workshop!

Post-AGI Civilizational Equilibria: Are there any good ones?

Vancouver, July 14th
www.post-agi.org

Featuring: Joe Carlsmith, @richardngo.bsky.social‬, Emmett Shear ... 🧵
Post-AGI Civilizational Equilibria Workshop | Vancouver 2025
Are there any good ones? Join us in Vancouver on July 14th, 2025 to explore stable equilibria and human agency in a post-AGI world. Co-located with ICML.
www.post-agi.org
Reposted by Dan Björkegren
davmicrot.bsky.social
🚨The Agentic Economy🚨 new paper w/ 9 co-authors: what happens to economy with expanded agent use for both consumers & businesses? The architecture of agentic communication will determine extent to which generative AI democratizes (or restricts) access to
economic opportunity arxiv.org/abs/2505.15799
Reposted by Dan Björkegren
maxkasy.bsky.social
🤖 Interested in machine learning, economics, and the state of AI?🤖

In September, I will teach a 1-week intensive version of my course on foundations of ML (maxkasy.github.io/home/ML_Oxfo...) in our summer school.

Apply here: ouess.web.ox.ac.uk/september-su...

Spread the word!
Foundations of Machine Learning
Research on machine learning, experimental design, economic inequality, and optimal policy
maxkasy.github.io
Reposted by Dan Björkegren
mclem.org
This is alarming. The US president is personally stating plans to remove people he personally deems to be "home-grown terrorists" to indefinite imprisonment overseas

The same prison where he claims authority to send *anyone*, in secret and without criminal charges or opportunity for appeal
titonka.bsky.social
Before press came in — but while live feed was running on Bukele’s feed — Trump said to him: “home-growns are next. … You’re gonna need to build about 5 more places.”

x.com/nayibbukele/...
x.com
dbjork.bsky.social
dbjork.bsky.social
Also, search results are low quality. Only 2% of results were from in-country. We asked teachers to rate responses, without telling them how the response was generated. Teachers rate AI responses as more helpful, relevant, and correct than web search results.
Reposted by Dan Björkegren
reidhoffman.bsky.social
Very important research. AI is better & cheaper for communities that don't yet use the web –– meaning it can unlock exponential gains for entire countries.
dbjork.bsky.social
New working paper: Could AI leapfrog the web?

Only 37% of sub-Saharan Africans use the internet. Cost is the #1 constraint. In a study with 469 teachers in Sierra Leone, we find AI works better than the web--and is 87% cheaper.

arxiv.org/abs/2502.12397

Let's dive in... 🧵(1/X)
dbjork.bsky.social
The research conducted with my students at Columbia Jun Ho Choi, Divya Budihal, Dominic Sobhani, and chatbot creators Oliver Garrod and Paul Atherton
dbjork.bsky.social
This is together with the excellent folks at Fab Inc who created the chatbot and are coordinating work at AI-for-Education.org. Also @educaid.bsky.social in Sierra Leone
dbjork.bsky.social
AI can reformat knowledge to work better on small screens and costly, intermittent connections. Can it help catch up remote, low-income communities to the revolution of information?
dbjork.bsky.social
Also, search results are low quality. Only 2% of results were from in-country. We asked teachers to rate responses, without telling them how the response was generated. Teachers rate AI responses as more helpful, relevant, and correct than web search results.
dbjork.bsky.social
Because querying an AI is 3,107x more data efficient, AI is already 87% cheaper than loading a web page for our teachers in Sierra Leone.
dbjork.bsky.social
Why don’t teachers use web search? First, because search is slow and expensive. The average web page uses 3,107x more data than the corresponding AI response (!)
dbjork.bsky.social
What do teachers use AI for? Mostly for facts and conceptual questions.
dbjork.bsky.social
85% of sub-Saharan Africans have mobile broadband signal, but few use the internet. Internet users use WhatsApp—but seldom web search (see plot).

We study a GPT-based chatbot accessible through WhatsApp. Sierra Leonean teachers use AI more than web search (leapfrog!)
dbjork.bsky.social
New working paper: Could AI leapfrog the web?

Only 37% of sub-Saharan Africans use the internet. Cost is the #1 constraint. In a study with 469 teachers in Sierra Leone, we find AI works better than the web--and is 87% cheaper.

arxiv.org/abs/2502.12397

Let's dive in... 🧵(1/X)
Reposted by Dan Björkegren
asheshrambachan.bsky.social
Applications are open for the Machine Learning in Economics Summer Institute (MLESI) 2025 are open!

If you're a graduate student, come learn about ML/AI and its uses throughout economics.

Apply by March 28. The application and more info can be found here: www.chicagobooth.edu/research/cen...
Machine Learning in Economics Summer Institute 2025 (MLESI25)
<span id="docs-internal-guid-53f97f20-7fff-b55c-c89c-4f85d663a48c" style="color: #39393a;"><span style="color: #000000;">MLESI<span id="docs-internal-guid-394e21c4-7fff-1437-16ff-fdbb1ce1d34c" style="...
www.chicagobooth.edu
dbjork.bsky.social
PhD students: there is a tutorial on AI + digital economics Feb 12 too; apply here www.nber.org/calls-papers...
Digital Economics and AI Tutorial
www.nber.org
dbjork.bsky.social
Love AI and economics? The NBER Digital and AI spring meeting will be at Stanford the day before Valentine's. Two weeks left to send us your best work! ❤️
www.nber.org/conferences/...
Digital Economics and AI Meeting, Spring 2025
www.nber.org
dbjork.bsky.social
Congratulations!!
dbjork.bsky.social
We'll be thinking about the future to improve the policies and ventures we build today. The syllabus is here. Are there readings or topics I am missing? Suggestions welcome! dan.bjorkegren.com/syllabus_aii...
dan.bjorkegren.com
dbjork.bsky.social
We'll be wrestling with the work of the future, who should train machines, and what values AI should encode. We'll explore whether decentralized minibus networks in Africa can tell us about planning cities for self driving cars, algorithms that bridge rather than divide, and ways to govern.