AP in Econ at @USCDornsife. I do Information Economics & more. Best friend of Spinoza (Nosik) Petrov-Libgober.
Los Angeles, CA
Joined September 2016
- Tweets6.5K
- Following1.7K
- Followers2.5K
- Likes34.8K
Pinned Tweet
New paper!!!
How does complexity relate to organization design?
And what does the answer say about how organizational hierarchy might change with new technologies like AI?
Thread below. If you don't have time for that, the picture basically summarizes the main idea.
Jonathan Libgober (same handle on other places) retweeted
Good managers are matched with bad workers to ensure all production lines meet deadlines with important buyers. In the September issue, by Achyuta Adhvaryu @achadhvaryu Vittorio Bassi @bassi_vittorio Anant Nyshadham @AnantNyshadham and Jorge Tamayo zurl.co/jfBTB
I feel like Grok's writing is aesthetically the most AI-ish but also the kind I most prefer. So predictable and structured, I know exactly what I'm going to get. Dependable.
Claude/ChatGPT's writing feels like it's more trying to fool me. Unnatural but harder to pinpoint.
Jonathan Libgober (same handle on other places) retweeted
Managerial misperceptions and heuristics lead to inefficient contract design, reducing firm performance by over 55 percent and challenging key assumptions of optimal contracting, from @fannan2316 and Collin B. Raymond nber.org/papers/w35788
Jonathan Libgober (same handle on other places) retweeted
Can AI conduct qualitative interviews with thousands of people? In a new REStud paper, Friedrich Geiecke and @XJaravel develop an open-source approach that combines the depth of qualitative interviews with the scale of surveys.
1/6
Jonathan Libgober (same handle on other places) retweeted
"We characterize equilibria in a canonical class of economies with endogenous production networks, showing that log-prices equal a min-max over networks of a primitive affine function of log-productivities."
NEW in @nberpubs: nber.org/papers/w35909
1000%.
I generally have no problem with AI if it's being disclosed; in many cases it can be a positive (e.g., someone used refine/refine competitor check).
But obfuscating AI use is frankly an insult--in the same way that failing to write/present as carefully as one could is.
Jonathan Libgober (same handle on other places) retweeted
Very happy to share a new paper, “Technology Bottlenecks: Measuring Where AI Could Raise the Value of Work It Cannot Yet Do.”
AI exposure measures ask about the work AI is capable of doing. We ask instead where the marginal value of work AI cannot yet do may rise as AI speeds up the rest of the job.
To do so, we employ 288 million U.S. job postings mapped line by line to their closest O*NET tasks and GPT-4 feasibility ratings from Eloundou et al. (2024) to construct a measure, “bottleneck pressure,” for each task in each occupation that posts it.
Using a simple task-based framework, we derive this pressure at the task level, showing that for an AI-infeasible task it is the exposure of the occupation that posts it, so the same task can face very different bottleneck pressure in different jobs.
Reporting bottleneck pressure alongside AI exposure shows both the scope for AI to perform work and where improving the work it cannot do may pay off more as capabilities expand.
Jonathan Libgober (same handle on other places) retweeted
Apropos of nothing, a friend received a rejection from a journal this week. R1’s ref report was obviously AI generated. 100% according to Pangram! Very, very likely that they didn’t even read the paper. And the editor just went with it.
What are we even doing here? If you want AI feedback, you could put the paper through it yourself. We go to journals, presumably, because they have access to experts. “Experts” delegating their thinking to AI is very clearly a violation of this social compact. What is the point of paying hundreds of dollars and waiting months for feedback from referees who pretend that they are contributing something?
I’m not anti-AI. I’m not against using it in refereeing either, if it’s done thoughtfully. But there are too many who are acting like it’s augmenting their expertise, when in fact they are letting it replace their thinking entirely. That’s a very dangerous norm to establish. If you replace your thinking entirely with AI, eventually you will be replaced.
Jonathan Libgober (same handle on other places) retweeted
I’m very opinionated. So it’s nearly impossible for me to hear someone’s arguments and *not* find something I disagree with. (This includes reading things I myself wrote years - or weeks - ago.) And yet the amazing Grant Sanderson @3blue1brown has put out a long, detailed video on the future of mathematics, in which I cannot find a single flaw. I’ll be sending this to every PhD student and postdoc I know, and strongly recommend everyone watch it.
youtube.com/watch?v=LKrwhjOI…
(Of course I tried to say many of the same things in my conversation with Thomas Koberda. But Grant’s version, unsurprisingly, does an infinitely better job at communicating those ideas!)
In the future, referees will be still be used by all major academic journals as part of regular review leading to article acceptance.
65%Yes
3%Within 2 years, no
10%Within 5 years, no
23%Who knows/show results
31 votes • Final results Jonathan Libgober (same handle on other places) retweeted
AEA RCT preregistrations often feature vague primary outcomes and frequent sample size deviations, while p-value distributions look similar for registered and unregistered studies. A dynamic signalling model shows banning late registration can actually raise registration rates.
Jonathan Libgober (same handle on other places) retweeted
In the current @EJ_RES ‘Research Registries and the Credibility Crisis: An Empirical and Theoretical Investigation’ by Eliot Abrams, Jonathan Libgober, John A List doi.org/10.1093/ej/ueaf120 @JlibDoesEcon @Econ_4_Everyone @RoyalEconSoc #EconTwitter
Finally in print!
In the current @EJ_RES ‘Research Registries and the Credibility Crisis: An Empirical and Theoretical Investigation’ by Eliot Abrams, Jonathan Libgober, John A List doi.org/10.1093/ej/ueaf120 @JlibDoesEcon @Econ_4_Everyone @RoyalEconSoc #EconTwitter
Not that it's my place, but here is some free advice for journal editors:
If a referee report is provided as a text response, rather than as a PDF, put the report through pangram to make sure the AI percentage is below 95%
Jonathan Libgober (same handle on other places) retweeted
Today the paper I am most proud of in my whole career came out as an NBER Working Paper. It took us a decade; all of our hard work is evident in its quality. I'm giving an ungated link to the pdf.
"Correcting for Heterogeneity in the Use of Subjective Scales"
@Dan_J_Benjamin is my only coauthor on X.
Here is the NBER page:
nber.org/papers/w35925?utm_c…
Here is the ungated pdf:
dropbox.com/scl/fi/fpgdvfoj6…
Jonathan Libgober (same handle on other places) retweeted
Today is my last day as mayor of the City of Evanston. These last 5+ years have been an incredible privilege, and I have learned so much from the people of this great community. Thank you, Evanston.
danielbiss.substack.com/p/wh…
Jonathan Libgober (same handle on other places) retweeted
1/8 There is lots of noise in the AI debates. One thing those debates do not need is pejorative framing: labels like “AI slop” and “plagiarism machine” that invite dismissal before evaluation. Academics need to have higher standards.
