Found this article thanks to a Reddit comment. Every day or so there's someone asking in math reddits what "dy/dx" means or why there's a "dx" in the integral notation, and then an army of people come out of the woodwork to push the orthodoxy of limits like some stockholm syndrome prisoners... And a small number of people point out that nonstandard analysis is actually a real thing and works... Anyway, back before I was born Sullivan did an actual experiment on teaching https://web.archive.org/web/20210506151356/http://academic.brcc.edu/crow/Projects/Calculus%20with%20Infinitesimals/Files/Sullivan%20(1976).pdf