pfp by @ShizzyAizawa. Post-Newtonian heterodox economist. Inventor of the JvN hypothesis for LLM scaling limits. World champion of econ/finance/food ragebait
🚨 NEW PAPER 🚨
Excited to announce this working paper (co-authored with @hongyaoma, @ykanoria, @rajivatbarnard) looking at the extent of wash trading on @Polymarket
We used all historical trade data from @Polygon blockchain (~70 million trades) and developed a graph-based algorithm for estimating wash volume.
My student Michael Cuna is on the econ job market this year. He has an excellent paper showing the impact of the hidden curriculum on educational outcomes, particularly in the case of first-gen students. The paper also demonstrates the potential for AI tools to close these gaps.
Please do not hesitate to reach out to me or the rest of the committee if you have any questions.
@CunaMichae77590’s JMP studies the hidden curriculum—unwritten rules for success. Combining data and field experiments, he finds first-gen students engage less in key actions, e.g. networking, due to low awareness.
Read more: economics.uchicago.edu/direc…#UChicago#EconJobMarket
Another few months pass, and still my theory holds: the physical upper limit to intelligence in the universe is John Von Neumann; frontier AIs are already at his level; and no entity in the universe will ever be strictly more intelligent than Von Neumann
🚨 JMP from our PhD candidate Marco Zanotti @_zanottiM shows how funds use data analytics to attract capital to both existing and newly created funds. Very interestingly, fees ↑ despite no chg in performance. Carefully executed. Check out his JMP/exciting research pipeline!
What brings money into mutual funds? This interesting JMP by Marco Zanotti cleverly shows that funds - especially those focused on retail customers - do data analytics on their customers, and this seems to contribute to fund flows!
Thanks @alz_zyd_!
Understanding how data availability reshapes asset management—beyond its role in improving portfolio allocation—is key to thinking about the future of the industry.
You can find my JMP here: marco-zanotti.com/files/JMP_…
What brings money into mutual funds? This interesting JMP by Marco Zanotti cleverly shows that funds - especially those focused on retail customers - do data analytics on their customers, and this seems to contribute to fund flows!
This is obviously not true in reality, and Marco's paper nicely shows that (my paraphrase) funds that are bad at investing, but good at marketing, unfortunately still manage to attract funds in equilibrium.
See the full job market paper on @_zanottiM 's website!
Tracking technology seems to be good at getting inflows, funds that do data analysis also increase expense ratios. Not good for consumers!
This paper is interesting because we sometimes naively think of funds as competing for capital mostly based on their performance
The paper is based on a super clever measurement strategy: if a website uses Google analytics or other tracking technologies, this shows up in their source code. So Marco just looks at fund websites to see who is using tracking tech
What brings money into mutual funds? This interesting JMP by Marco Zanotti cleverly shows that funds - especially those focused on retail customers - do data analytics on their customers, and this seems to contribute to fund flows!
Visited a startup accelerator in Palo Alto last week
It was 10pm and the founders were still coding
“Do you have a government permit to be working at this hour?” I asked them
They looked at me and said “bro what are you talking about” in a typical condescending American way
I explained that in Europe we would require special authorization from the labor ministry and overtime compensation forms filed in
“Bro no wonder you guys ship so slow”
I immediately called the police reporting there are people doing unlicensed labor after business hours
The officer laughed and hung up
I left for the airport that same day and flew back to Berlin where work happens between 9am and 5pm, with 1h lunch break, as intended by law
No wonder American startups fail so often
Joking aside, here's my base case for thinking about AI's impact on GDP growth. I think we'll keep growing at 2%. Stuff like AI that comes once in a while how we managed to grow at 2%/y as we've done for millennia since the first industrial revolution.
There’s no sense in which we are doing something "extra". This is just what we need to do to keep growing as we have. Otherwise, we stagnate. In other words, “AI” = “2% GDP growth in long run equilibrium”. It's just like all the technological shocks that have propelled us forward thus far.
Adding a good quote from @danielrock:
"We need innovation to grow at the base rate."
It's probably fine for students to use ChatGPT in schools. The students are thinking less, but the total amount of thinking in schools is increasing, since the LLMs are thinking more than the students ever did
At higher levels of GDP, societies which produce higher GDP aren't actually any better at moving mass around. They're just better at producing dopamine, basically, since at the top of Maslow's hierarchy we pay not for calories but for dopamine
Once everyone is fed and clothed, GDP starts measuring weird stuff, like whether you eat fancy branded yogurt with expensive ads, or generic chemically identical yogurt. Or how much time you spend staring at a screen playing videogames
The economic idea of GDP is a decent measure when society is on the lower levels of Maslow's hierarchy. There, GDP measures essentially whether each person is eating 1 pound of rice or 2 pounds of rice, which basically measures how good society is at moving mass around
After this comped and then un-comped cake was tantalizingly flashed in front of our eyes, the server then came by with the bill.
Americans. Why. Why do you put up with this???
I was at a steakhouse today, my wife and I spent $360 in total. We reserved for 7:45pm, got seated at 8:20pm. Towards the end, a delicious-looking comped cake arrived. As we were about to eat it, our server rushed over saying the cake came to the wrong table, and whisked it away
Did a quick class poll and 80%+ of students want to have take-home exams. What is the point of having take-home exams in 2025??? you just put the exam in ChatGPT and get 100%
you might ask. What is the point of learning anything in 2025 if ChatGPT can do everything possibly worth doing? and my answer would be, idk I guess. Learning is fun maybe? hard questions
Did a quick class poll and 80%+ of students want to have take-home exams. What is the point of having take-home exams in 2025??? you just put the exam in ChatGPT and get 100%