Last post here was December 2023. Long time.
Honest version: writing felt like extra work on top of the actual learning. Turns out you have to write to learn the thing properly, so I’m back.
What I’m spending time on right now:
- Model distillation.
- Agent RL workflows: GRPO and other training methods for RL environments.
- Math Academy in the background, daily.
Not going to promise a posting cadence. The bar is “write when I have something honest to say.” If you want updates, the RSS feed works.
Back to top