Reza Salehnejad
@rezasalehnejad.bsky.social
Reposted by Reza Salehnejad
And of course @akorinek.bsky.social's excellent JEL paper www.aeaweb.org/content/file...
www.aeaweb.org
July 9, 2025 at 7:50 AM
And of course @akorinek.bsky.social's excellent JEL paper www.aeaweb.org/content/file...
Reposted by Reza Salehnejad
New easy-access source of economic data -- like FRED but from Stanford & MIT and with different series.
#EconSky
www.industrydatanavigator.org
#EconSky
www.industrydatanavigator.org
Industry Data Navigator
Democratizing data on investment, employment, production, prices and trade to understand the 21st century economy.
www.industrydatanavigator.org
June 14, 2025 at 1:07 AM
Reposted by Reza Salehnejad
🤖💡 Are LLMs really developing social behaviors—or just reproducing familiar patterns from training? In our new paper, @pettertornberg.com and I take a closer look at recent claims of emergent conventions in AI and offer a different perspective... 🧵
June 2, 2025 at 6:41 AM
🤖💡 Are LLMs really developing social behaviors—or just reproducing familiar patterns from training? In our new paper, @pettertornberg.com and I take a closer look at recent claims of emergent conventions in AI and offer a different perspective... 🧵
Reposted by Reza Salehnejad
June 7, 2025 at 11:45 PM
Reposted by Reza Salehnejad
Reposted by Reza Salehnejad
Curious how accurate multimodal LLMs are at digitizing diversely formatted historical tables?
➡️ pretty accurate!
➡️ helps a lot with harmonization
➡️ far cheaper than manual digitization
➡️ but iterative, disciplined prompt training is important!
Check out our new WP - arxiv.org/abs/2505.11599
➡️ pretty accurate!
➡️ helps a lot with harmonization
➡️ far cheaper than manual digitization
➡️ but iterative, disciplined prompt training is important!
Check out our new WP - arxiv.org/abs/2505.11599
June 2, 2025 at 2:18 PM
Curious how accurate multimodal LLMs are at digitizing diversely formatted historical tables?
➡️ pretty accurate!
➡️ helps a lot with harmonization
➡️ far cheaper than manual digitization
➡️ but iterative, disciplined prompt training is important!
Check out our new WP - arxiv.org/abs/2505.11599
➡️ pretty accurate!
➡️ helps a lot with harmonization
➡️ far cheaper than manual digitization
➡️ but iterative, disciplined prompt training is important!
Check out our new WP - arxiv.org/abs/2505.11599
Reposted by Reza Salehnejad
Meanwhile at the UVA Center for Teaching Excellence yesterday, we had a writing retreat to prepare new collections for the UVA Teaching Hub. Expect to see several new collections posted this week! First up is one with my favorite resources on in-class polling: teaching.virginia.edu/collections/...
In-Class Polling for Student Engagement — UVA Teaching Hub
<p>Instead of asking our students "Any questions?" and hoping for a response, we can use polling technologies to enable and invite all of our students to share their questions and respond to ours. The...
teaching.virginia.edu
June 3, 2025 at 2:00 PM
Meanwhile at the UVA Center for Teaching Excellence yesterday, we had a writing retreat to prepare new collections for the UVA Teaching Hub. Expect to see several new collections posted this week! First up is one with my favorite resources on in-class polling: teaching.virginia.edu/collections/...
Reposted by Reza Salehnejad
Original paper (paywalled): www.nature.com/articles/s41...
Response paper (open): www.nature.com/articles/s41...
Response to response (paywalled): www.nature.com/articles/s41...
Response paper (open): www.nature.com/articles/s41...
Response to response (paywalled): www.nature.com/articles/s41...
An empirical investigation of the impact of ChatGPT on creativity - Nature Human Behaviour
The use of ChatGPT helped generate more creative ideas for various everyday and innovation-related problems, compared with not using any technology or using a conventional Web search (Google). This ef...
www.nature.com
May 14, 2025 at 6:24 PM
Original paper (paywalled): www.nature.com/articles/s41...
Response paper (open): www.nature.com/articles/s41...
Response to response (paywalled): www.nature.com/articles/s41...
Response paper (open): www.nature.com/articles/s41...
Response to response (paywalled): www.nature.com/articles/s41...
Reposted by Reza Salehnejad
1/ The rise in larger firms' share of private sector US employment over the past 3 decades has come mostly at the expense of small ones (5-49 employees).
May 13, 2025 at 6:21 PM
1/ The rise in larger firms' share of private sector US employment over the past 3 decades has come mostly at the expense of small ones (5-49 employees).
Reposted by Reza Salehnejad
A new study with almost 300 Chinese student subjects replicated the main findings of our Management Science paper (Ederer & Manso 2013) that short-term pay-for-performance and lack of tolerance for early failure inhibit innovation.
So happy to see this experiment still making waves!
So happy to see this experiment still making waves!
May 13, 2025 at 6:47 PM
A new study with almost 300 Chinese student subjects replicated the main findings of our Management Science paper (Ederer & Manso 2013) that short-term pay-for-performance and lack of tolerance for early failure inhibit innovation.
So happy to see this experiment still making waves!
So happy to see this experiment still making waves!
Reposted by Reza Salehnejad
Ohh I had this idea. Guess it didn't work.
May 8, 2025 at 6:28 PM
Ohh I had this idea. Guess it didn't work.
Reposted by Reza Salehnejad
Study in Nature: “Across 30 out of 32 evaluation axes from the specialist physician perspective & 25 out of 26 evaluation axes from the patient-actor perspective, AMIE [Google Medical LLM] was rated superior to PCPs [primary care docs] while being non-inferior on the rest.”
(& AIME is an older LLM)
(& AIME is an older LLM)
May 4, 2025 at 1:27 PM
Study in Nature: “Across 30 out of 32 evaluation axes from the specialist physician perspective & 25 out of 26 evaluation axes from the patient-actor perspective, AMIE [Google Medical LLM] was rated superior to PCPs [primary care docs] while being non-inferior on the rest.”
(& AIME is an older LLM)
(& AIME is an older LLM)
Reposted by Reza Salehnejad
May 4, 2025 at 7:27 PM
Reposted by Reza Salehnejad
New preprint from @lauraknelson.bsky.social, @mattwilkens.bsky.social, and myself tests different ways of simulating the past with LLMs. We don't fully answer the title question here—just show that simple strategies based on prompting and fine-tuning are insufficient. +
Can Language Models Represent the Past without Anachronism?
Before researchers can use language models to simulate the past, they need to understand the risk of anachronism. We find that prompting a contemporary model with examples of period prose does not pro...
arxiv.org
May 2, 2025 at 12:47 PM
New preprint from @lauraknelson.bsky.social, @mattwilkens.bsky.social, and myself tests different ways of simulating the past with LLMs. We don't fully answer the title question here—just show that simple strategies based on prompting and fine-tuning are insufficient. +
New preprint from @lauraknelson.bsky.social, @mattwilkens.bsky.social, and myself tests different ways of simulating the past with LLMs. We don't fully answer the title question here—just show that simple strategies based on prompting and fine-tuning are insufficient. +
Can Language Models Represent the Past without Anachronism?
Before researchers can use language models to simulate the past, they need to understand the risk of anachronism. We find that prompting a contemporary model with examples of period prose does not pro...
arxiv.org
May 3, 2025 at 3:25 PM
Reposted by Reza Salehnejad
Paper on the economics of AI competition shows new AI models are adopted very rapidly but are not always substitutes for each other (better models in the same family often substitute, but others expand the pie). Also, people use a mix of models. andreyfradkin.com/assets/deman...
May 2, 2025 at 12:39 PM
Paper on the economics of AI competition shows new AI models are adopted very rapidly but are not always substitutes for each other (better models in the same family often substitute, but others expand the pie). Also, people use a mix of models. andreyfradkin.com/assets/deman...
Reposted by Reza Salehnejad
April 22, 2025 at 4:31 PM
Reposted by Reza Salehnejad
🚨 New working paper 🚨
Demand for LLMs: Descriptive Evidence on
Substitution, Market Expansion, and Multi-Homing
A key question for the business of AI is the extent to which LLMs are differentiated from each other. I use data from OpenRouter to take a first look.
andreyfradkin.com/assets/deman...
Demand for LLMs: Descriptive Evidence on
Substitution, Market Expansion, and Multi-Homing
A key question for the business of AI is the extent to which LLMs are differentiated from each other. I use data from OpenRouter to take a first look.
andreyfradkin.com/assets/deman...
April 22, 2025 at 3:35 PM
🚨 New working paper 🚨
Demand for LLMs: Descriptive Evidence on
Substitution, Market Expansion, and Multi-Homing
A key question for the business of AI is the extent to which LLMs are differentiated from each other. I use data from OpenRouter to take a first look.
andreyfradkin.com/assets/deman...
Demand for LLMs: Descriptive Evidence on
Substitution, Market Expansion, and Multi-Homing
A key question for the business of AI is the extent to which LLMs are differentiated from each other. I use data from OpenRouter to take a first look.
andreyfradkin.com/assets/deman...
Reposted by Reza Salehnejad
There is a bunch I disagree with here, but it makes an argument for why AI adoption may be less rapid and less immediately world-changing (& dangerous) than people expect.
Well worth reading, as it is a useful base case where AI is just another technology. knightcolumbia.org/content/ai-a...
Well worth reading, as it is a useful base case where AI is just another technology. knightcolumbia.org/content/ai-a...
AI as Normal Technology
knightcolumbia.org
April 20, 2025 at 1:28 AM
There is a bunch I disagree with here, but it makes an argument for why AI adoption may be less rapid and less immediately world-changing (& dangerous) than people expect.
Well worth reading, as it is a useful base case where AI is just another technology. knightcolumbia.org/content/ai-a...
Well worth reading, as it is a useful base case where AI is just another technology. knightcolumbia.org/content/ai-a...
Reposted by Reza Salehnejad
Large scale job displacements due to AI are likely to occur more slowly than a lot of people talking about AI and work might suspect.
Translators were paid more and in more demand a year after GPT-4 than they were the year it launched. Change is likely to come, but the timing is hard to predict.
Translators were paid more and in more demand a year after GPT-4 than they were the year it launched. Change is likely to come, but the timing is hard to predict.
April 15, 2025 at 10:34 PM
Large scale job displacements due to AI are likely to occur more slowly than a lot of people talking about AI and work might suspect.
Translators were paid more and in more demand a year after GPT-4 than they were the year it launched. Change is likely to come, but the timing is hard to predict.
Translators were paid more and in more demand a year after GPT-4 than they were the year it launched. Change is likely to come, but the timing is hard to predict.
Reposted by Reza Salehnejad
How Tariffs Affect Trade Deficits...
... a new paper with amazing trade economist Arnaud Costinot.
Many politicians and the general public expect tariffs to reduce imports lower imports and thus work to close a trade deficit.
Economists typically say "not so fast"...
🧵1/n
... a new paper with amazing trade economist Arnaud Costinot.
Many politicians and the general public expect tariffs to reduce imports lower imports and thus work to close a trade deficit.
Economists typically say "not so fast"...
🧵1/n
April 14, 2025 at 5:25 PM
How Tariffs Affect Trade Deficits...
... a new paper with amazing trade economist Arnaud Costinot.
Many politicians and the general public expect tariffs to reduce imports lower imports and thus work to close a trade deficit.
Economists typically say "not so fast"...
🧵1/n
... a new paper with amazing trade economist Arnaud Costinot.
Many politicians and the general public expect tariffs to reduce imports lower imports and thus work to close a trade deficit.
Economists typically say "not so fast"...
🧵1/n
Reposted by Reza Salehnejad
What is a recent #EconSky economics paper that has a core result that is
1. Important
2. Surprising (sign or size)
3. Decently simple
Self nominations r welcome.
1. Important
2. Surprising (sign or size)
3. Decently simple
Self nominations r welcome.
April 11, 2025 at 7:50 PM
What is a recent #EconSky economics paper that has a core result that is
1. Important
2. Surprising (sign or size)
3. Decently simple
Self nominations r welcome.
1. Important
2. Surprising (sign or size)
3. Decently simple
Self nominations r welcome.
Reposted by Reza Salehnejad
Retaliation hurts! Check out our new paper to see how costly it can be for the U.S. economy.
🚨 We just released a paper examining the long-terms economic impacts of "Liberation Day" tariffs.
In summary: while these tariffs may help reduce the trade deficit, they will impose significant costs on the U.S. economy after retaliation by trade partners.
In summary: while these tariffs may help reduce the trade deficit, they will impose significant costs on the U.S. economy after retaliation by trade partners.
April 10, 2025 at 7:37 PM
Retaliation hurts! Check out our new paper to see how costly it can be for the U.S. economy.
Reposted by Reza Salehnejad
@manuelacollis.bsky.social @clemvaneff.bsky.social this is such great work!
People are willing to forgo 15-30% of their wage to avoid hostile work environments.
Women report a stronger distaste for exclusive workplaces and environments with sexual harassment and value hybrid work twice as much in the presence of sexual harassment.
Collis & @clemvaneff.bsky.social
Women report a stronger distaste for exclusive workplaces and environments with sexual harassment and value hybrid work twice as much in the presence of sexual harassment.
Collis & @clemvaneff.bsky.social
April 10, 2025 at 12:52 AM
@manuelacollis.bsky.social @clemvaneff.bsky.social this is such great work!