Exploring the World of Charles Manski with an AI Generated Podcast Doing a Deep Dive on His Vita
Exploring the World of Charles Manski with an AI Generated Podcast Doing a Deep Dive on His Vita
The notebookLM thing that Google does where you upload a document and it generates outlines and summaries, but also an AI generated podcast with two AI generated speakers explaining any particular document to one another as well to their so-called audience is pretty amazing. But yesterday I had an idea—I wonder how far you could get with it if the only thing you gave it was someone’s vita. After all, there’s not a ton of information in a vita about people’s life and research, but on the other hand, it does function a bit like an autobiography of one’s career. What could these podcast hosts do if that’s all you gave them and it turns out—they can do a lot.
So, after spending yesterday uploading a lot of my friends’ vitas into it and sending it to them, as well as my own—which by the way is a great way to give someone who visits your website, or even your family, a bit more of a bird’s eye view of who you are, not to mention how the hosts of the show do a great job putting all of your life’s work in their own framing and connecting a lot of dots that maybe even you don’t see—I thought well who do I want to do this with in econometrics? Who is it that I think it would be great for them to simply review their vita and do an entire show about it? After all, they do it in complete sincerity, so that might be interesting. And I decided I wanted to do it with Charles Manski, an esteemed econometrician at Northwestern.
Charles F. Manski is a highly influential figure in econometrics and economics, holding the title of Board of Trustees Professor in Economics at Northwestern University since 1997. His academic journey began with a B.S. and Ph.D. from MIT, leading him to prestigious positions at institutions like the University of Wisconsin-Madison, Hebrew University of Jerusalem, and Carnegie Mellon University. Manski’s work is known for pioneering approaches in econometrics, especially around partial identification—a statistical technique that navigates uncertainty by identifying bounds rather than exact figures, broadening our understanding of complex, data-limited scenarios in public policy and decision-making. His prolific publication record, including Public Policy in an Uncertain World and Patient Care under Uncertainty, reflects a dedication to tackling ambiguity in real-world applications and public policy, which has earned him memberships in the National Academy of Sciences, the Econometric Society, and other esteemed organizations.
The podcast in the above embedded audio bubble is going to be a “deep dive”, like I said, by two AI hosts of an unnamed podcast. One is a female and one is a male, and you can tell by how they speak and the approach to the show that it fits a particular genre of podcasts out there. Think of the Binge Mode at the Ringer, or Reply All. Something about the voice, the background research, the person coming to the other team to explain what they learned but which sounds like they’re explaining something to one another for the sake of the audience too, but also as if they audience isn’t there.
The only information I gave them was this document—Manski’s 2024 CV. The accessing of outside sources seems pretty limited, to be honest, because if you ever just ask in the user interface things beyond what you uploaded, it does seem sometimes like it either cannot or is pretending it cannot access other information. I know that can’t be 100% true, because how else are they able to speak so fluidly and intelligently? I don’t know—I just know that I only uploaded Manksi’s CV and in addition to generating a summary and outline of it, it also generated the podcast which is a shareable link here, but it’s also a podcast I downloaded and embedded.
As far as Stephen Johnson’s involvement, he announced the release of notebookLM last October 2023. You can read that on his substack here.
So, I hope you enjoy this explanation of a person’s career by two AI podcasts on an unnamed show doing a deep dive on Manski’s CV and piecing together who he is and what he’s been about.
First post as a paid subscriber, but I'm a long time beneficiary of your Mixtape works, Scott.
I have to say, I'm impressed with NotebookLM (and have played a bit with it a while back), but does anyone else get serious Kristen Wiig and Will Farrell vibes when llistenlistening to these (worth a watch if you haven't: https://youtu.be/XDvZrubaEwA?si=cSnHjw1DRyEq82jV -- once you see it you can't unsee it). Regardless, it's a powerful tool.
That said, I was delighted by this episode on Charles F. Manski, though I want to hear you interview him. He's a hero of mine. I consider myself to be among of his intellectual "descendants," though we've never met in person--I sometimes think meeting him would kinda' be like Buddy the Elf showing up on his dad's doorstep. Though I'm more like a grandkid, as I was a pupil of two of his students, Jörg Stoye and Francesca Molinari at Cornell.
Francesca served on my special committee and is a letter writer as I head back on the job market this year. My dissertation research was originally based on some methods I developed based on an observation he made in passing in an article he wrote titled "Adolescent Econometricians" about how the individuals in our models (e.g. the Mincer equations on the returns to education) themselves face the same inferential problems we do as econometricians--I ran with this idea and this led to a couple projects with my favorite acronyms: Behavioral Econometric Estimators (BEEs) and Highly Integrated Vital Estimators (HIVEs). I founded a software firm called Hivemind Analytics in grad school to store capital I raised to execute on the idea and me and a team of CS friends developed an app called Predictament that streamlined the process of running prediction tournaments with strategic survey questions. His "Law of Decreasing Credibility"--about how the credibility of our inference decreases with the strength of the assumptions we maintain--is part of my credo and something that I've considered getting tattooed somewhere ;).
But, none of this is what I'd talk to him about if we were to meet or talk. I would actually just thank him for his work on medical decision-making.
It has been this set of work, above all, that I want to thank him for. In my 3rd year of graduate school, just as I was hitting my stride and working on my dissertation research, I got a call that changed everything.
Ended up patching together some other papers I had co-authored for my dissertation, moved to be closer to one of the country's best treatment centers and finished my degree in absentia. I never went on the job market (it's my first time, now), but was fortunate to find a position within a day's drive of the girls doctors. I backburnered the BEEs and HIVEs. Stuff happened, became a half-decent respiratory tech/CNA, and I had to close down Hivemind.
Our family has now been navigating life with two kids with an extremely rare variant of a terminal illness for 10 years.
This comment is already too long. Point is, Manski's work has helped my family make decisions in the face of all the uncertainty that has come. And, it's probably time I call him and thank him.
Anyway, if you would ever like a co-host if/when you interview him, let me know. I'd love to be a fly on the wall. I bet we could have a conversation that would give NotebookLM a run for it's money.