[throwawaygh]

I made this exact jump. The answer, for me, was a four year bachelor's degree.

I was a self-taught software engineer, not a self-taught computer scientist. I was writing production code for fairly normal stuff, not proving theorems or implementing research prototypes of new ideas. So YMMV.

If you can make it through TAPL or PFPL on your own, then you probably know enough to be useful to some random grad student somewhere. Latch onto an implementation effort, get authorship on a popl/splash/pldi paper or two, and that's probably enough to convince some faculty somewhere to take a chance on you.

But most good graduate programs still require a regionally accredited degree. Getting an exception to that rule will likely require a publication record. (Getting into good graduate programs often requires a publication record in any case. I recommend against attending not-good graduate programs.)

[pretzell]

I'm almost halfway through this: https://plfa.github.io/

I was auditing a grad class at UT that used this book, then corona hit. Still plan to finish it eventually though

I'm not even really sure where I would get started working on an implementation or working on a paper. Are these projects open source?