I made this. One of the most fun things about using coding agents is making learning more interactive. A note I wrote to myself while back:

"I forgot this" keeps turning into "oh, I never actually understood this."

Using coding agents to work through those gaps has become one of my favorite parts of research. Mildly humiliating, but fun.

This is built on Mycelium, the open-source theme my blog runs on. I combined it with my RL study repo, which started on 31 December 2016 as a list of talks and courses, and turned it into this. The same idea could work with any set of lectures.

It might look gimmicky, but this is what I'd imagine interactive doodles from listening to a lecture on reinforcement learning would look like.

The lectures are drafts written with coding agents, and each one is marked unreviewed until a person has checked it. Checking each figure against a separate Python implementation turned up mistakes in four of the lectures, which are fixed now.

One of my goals is to get people interested in learning, or relearning, in general. And it's something I have fun with.

- Code: https://github.com/0bserver07/rl-study-site and - Original Study Notes: https://github.com/0bserver07/Study-Reinforcement-Learning