Things I write, math I find beautiful, and links worth your time.
Coming soon
Why I put a language model inside an autopilot stack
A walkthrough of the reasoning behind the Semantic Observer project — why existing perception wasn't enough, and what it actually means to add semantic understanding to a real-time safety system.
Coming soon
The Clohessy–Wiltshire equations, explained from scratch
Linearised relative orbital dynamics are surprisingly elegant once you see where the equations come from. A visual, intuition-first derivation for anyone who's ever wondered how spacecraft formation control actually works.
2024
The Bitter Lesson — Richard Sutton ↗
The most important essay in AI research that almost nobody has fully internalized yet. Sutton argues that general methods that leverage computation consistently outperform methods that leverage human knowledge. Required reading for anyone building ML systems.
incompleteideas.net · Richard Sutton
2024
Proximal Policy Optimization Algorithms — Schulman et al. ↗
The paper behind PPO — still the most practical policy gradient algorithm for locomotion tasks. Remarkably readable for a foundational RL paper.
arxiv.org · OpenAI
2017
Visual Evolution Strategies — David Ha ↗
A beautifully visual walkthrough of evolution strategies as a family of optimisation algorithms — covering CMA-ES, OpenAI-ES, and beyond. If you've been working with MPPI and want to understand what else is out there, this is the natural next step. Highly recommended for anyone thinking seriously about black-box optimisation in robotics and control.
blog.otoro.net · David Ha