Annual review of Mixtape Sessions Workshops! More cats, plus a new place to live
This is one of my personal updates. Skip it or don’t! Hope all of you are well.
First so that everyone sees this, I discussed yesterday a new paper showing since ChatGPT release, Upwork freelancers earnings and employment have declined until the end of the sample, using only a single days scrape. But Upwork CEO was on a show and said demand for any freelancer with any AI skill is soaring. Fairly substantial discrepancy between the study’s findings of large 2-3% reduction in total employment and 5-10% reduction in earnings in their event study versus the ceo statement and I’m wondering if it’s due to needing to focus on specialized skill? She says those with any AI background, versus their study which is the entire dataset scraped on that one day. Also, it could be the single day scrape that is why there’s such a difference. Maybe you can’t scrape one day to get future trends but could to get pre trends? Definitely want to think about that more.
But let me move on. Hear that? Listen again closely. That’s the sound of the door closing on a successful year of Mixtape Sessions. 2024 is right around the corner, and we just had our last workshop of the year. There were 22 workshops and 2800 people who attended. In addition to me, there was Jon Roth, Peter Hull, Rocío Titiunik, Brigham Frandsen, Chris Walters, Mark Anderson, Dan Rees, Brant Callaway and Alberto Abadie — all teaching applied econometric theory and practice to people all over the world. The number of those low income country folks grew substantially once I added India to the list.
It was quite exciting for me to have a chance to be involved with this kind of large scale, inexpensive teaching effort. It has been a real blessing for me in a lot of ways, but one way is I feel like I saw the need based on a conjecture of mine and decided to be the hub and bridge I felt was missing. I wanted to make connecting causal inference further out is still holding together. Many people know I hypothesize causal inference has mainly spread through labor markets and placements in applied micro field courses, less so textbooks. Even still while not true like it once was, I bet it’s still a bit true. So Mixtape Sessions is meant to address that, even if it’s just at the margin. So I’m pleased about the progress. I think adding two years together, it’s probably 5500 people attended.
So what’s next? Well with Tymons paper finally written up in my deck, and with my decision to reorganize my slides on uncondoundedness one more time, I think I have finally gotten what I want for this unconfoundedness chapter of my book. It took a long time but I think I’m finally ready to wrap up the chapter. Hopefully I can finally finish it. I still have a lot to do, and I realize that I finished the book originally because it was all I did for a summer. I’d been working on it more slowly before that summer as I was living in atlanta with my brother spending time with a family member in the hospital during visiting hours then writing at a coffeeshop the rest of the time. But now I’ve split my time up a lot, and am behind. I have to get it fixed so I can finish it and move onto another one of two books.
Keep reading with a 7-day free trial
Subscribe to Scott's Substack to keep reading this post and get 7 days of free access to the full post archives.