Learning to code without mutation allows you to build on equational reasoning from grade school (+ inductive reasoning from discrete math) for quite a while. Adding mutation means deleting equations and adding all sorts of complex reasoning about state. Should be introduced carefully later imo.

I like the "misinterpreter" approach a lot (working w/ Peter-Michael Osera in various directions along the same lines w/types!). I also like the SMoL Tutor approach of showing that the same ideas show up in multiple languages. I think throwing mutation in immediately is a mistake though!

The best part of going to a conference is the things you learn; come cool talks I saw include @kcsrk.info about formal verification of GCs, others about @neurocy.bsky.social's Hazel, @sadiq.toao.com on OCaml and AI, and how I got Shriramed by... @shriram.bsky.social anil.recoil.org/notes/icfp25...

@neurocy.bsky.social defines the core problem as bad defaults in scientific code tools like Jupyter that make it hard to be FAIR without lots of work #icfpsplash25

@neurocy.bsky.social arguing for the digital commons as a massive single live program that’s collaboratively edited by thousands of contributors and live updated! A Wikipedia-meets-GitHub-meets-Jupyter #icfpsplash25

@negabinary.bsky.social is on it!