“Memory is the residue of thought” and fixing that when working with AI agents
Yesterday was day one of a two day new faculty orientation at Harvard Kennedy school.
Scott's Mixtape Substack
Scott’s Mixtape is dedicated to teaching causal inference, econometrics, and applied empirical research with a focus on how AI agents are transforming how researchers work. Practical tutorials, live research, and documentation of what works and doesn’t.
Live Last read · last published · next check
Latest posts
Saves to your Listen queue, to pick up on another day or another device.
Online version of The Remix is now here
My new book is called Causal Inference: the Remix and it comes out August 25th. Like my first, its publisher is Yale University Press. And Yale, as with the first, has been gracious enough go agree to there being a free online version.
Long overdue closing tabs on a Saturday morning
Oh what fool I’ve been.
Which stages of research production should we let AI Agents loose
Six months ago, it was both original and true to say that AI agents were capable of autonomously, start to finish, producing a manuscript of contemporary quantitative empirical research.
The Art of Loving
On Monday night I attended Olivia Dean’s concert at the Garden in Boston. Me and my friend had floor seats, and once the stadium was nearly full, I remarked to them that I couldn’t remember the last time being in such a place where I was so far outside of the distribution with respect to demographics. If I had to guess, the average age was probably 25-…
Will AI revive the Labor theory of value
I teach history of economic thought at Baylor every year.
My strategy for talks on agents for research
My preference for teaching new concepts in statistics and causal inference to people is like this.
How to improve understanding with Claude Code
Paul Goldsmith-Pinkham posted on his substack feed the other day a link to this interesting blog post by Geoffrey Litt called “Understanding is the new bottleneck”.
Demographic differences in online dating
I recently found some interesting surveys over at the Pew Center about online dating in America.
Episode 16: Landing the Plane
After 16 episodes of working with Claude Code to study the effect of abortion clinic closures on marriage rates using a natural experiment, so to speak, in Texas called "House Bill 2" and continuous difference-in-differences with a treatment of "change in distance", Caitlin and I start our descent.
Hidden Curriculum (August 5th and 6th)
Greetings again!
Prose is not constraints and other things I learned from Paul Goldsmith-Pinkham's NBER talk on AI in empirical research
By mid-2026, I think many researchers understand that coding agents like Claude Code and Codex have dramatically reduced the marginal cost of producing empirical research.
Sex recessions, rise in casual sex, online dating, declining trust, screening and other changes in US sex patterns
I always wanted to be the Kinsey of economics who specialized in sex research.
Which research mistakes matter for human capital?
And does it matter?
Books are here!
Well well well, what have we here.
Countdown to the Remix
Don't call it a comeback
Stale code, deprecation, and the canon
I think I first noticed I needed a structured workflow with using Claude Code when during a long project the analytical sample went from having N observations to N+k observations.
Hidden Curriculum Workshop (August 5-6th)
Greetings!
Am I'm reading too much?
Some days I think I have adopted a new way of working with Claude Code that has revolutionized my entire life.
Covariates, diff in diff and Lalonde test
Well can you believe it?
Episode 15: Exploring Confounders That Could Impact Our Parallel Trends Assumption
Listen now (66 mins) | Caitlin Myers and myself have been studying the effect of House Bill 2 in Texas, which closed half the state's abortion clinics, on marriage rates using "continuous difference-in-differences" estimation for 14 episodes straight.