Ok day, did a good amount of work in no particular order.

• Found some great Bayesian statistics books and watched the first statistical rethinking lecture
• Spent 3h reading the rest of the ELK document
• Spent 1h reading the math chat on ElutherAI's discord. Heard about some crazy theorem from logic where a statement true in "large finite fields" is automatically true in $\mathbf{C}$ (apparently this is a consequence of Tychonoff's theorem, WTF)
• Spent 40m reading about differential forms in the PCM, learned a bit and put it in Anki but wasn't that productive
• Spent 1h reading this paper linked on eleuther.
• 45m Reading ELK prize results. I'm upset, I totally could have come up with some of the strategies that got a prize. I guess that's a lesson to being intimidated by prize money, don't let that stop you!
• 1h 30m Reading the PCM while eating, the section on inequalities is good, also read the section on the halting problem (learned the proof, and that Diophantine equations are undecidable in general), Fermat's last theorem (some ideas for how its approached, learned it was proved by computer for $n \le 4\cdot 10^6$ and how it's connected to elliptic curves, etc)
• 1h Random browsing about, asked the author of the paper I read some questions on Twitter.

There is a truly massive amount of books (math and non-math) I want to read. I've got to start reading 5-6h/day minimum if I want to finish them in any reasonable time.

Tomorrow I've got an interview, hopefully I do well.