Webpage reboot — agentic coding and research

I'll be writing about what I spend most of my time on: agents, orchestration, and model training — pre-training, post-training, RL, and scaling test-time compute. And — can we get models to improve themselves at pre- and post-training?

The way we build things and do research is a fast-moving landscape — and it's never evolved this fast. More posts to come soon.