Oh no! It’s only Monday and I’m already approaching limits on tokens. I promise I’m not just token maxing. Here are some of the reasons I’m hitting this limit:
-
Interview practice: I’ve been using Claude to help me practice “case interviews.” This is an interview style I’d never encountered before, but I needed to learn it fast for a Capital One interview. A friend generously gave me a bunch of practice examples, and Claude’s been playing interviewer ever since.
-
Game development in a loop: I’ve been working on adding better end-to-end automation testing on my game Chal Rickshaw! (try it here https://maximumimpact.studio/chal-rickshaw/) including an automated driver that plays a bunch of rounds while recording video, which an agent then reviews to evaluate the results. I also have some changes to drive user retention in the works. Design, coding, and development have been running in a loop, with Claude Code writing and Codex playing reviewer. They’ve been having way too much fun late at night getting work done. Blog post incoming on how I set this up!
-
Tracking health: I’ve been using Claude to track our health — mine and my wife’s. I have it dig through MyChart (yeah, I’m that guy!) and make recommendations, which we then bring up with our clinicians. It’s been really interesting using our numbers (blood tests, MRIs, EKGs, vitals) to drive these conversations and digging into what we can optimize.
So yeah, my relationship with Claude/Codex is going strong right now ;)