Robotics

I started getting into robotics in early 2026, and it's been one of the most humbling things I've worked on. Software is forgiving: you can iterate quickly, errors are reversible, and the feedback loop is fast. Hardware is the opposite. A robot doesn't care about your logic if the motors can't execute it.

The early focus has been on fundamentals: understanding control systems, sensors, and what it actually takes to make a machine navigate an environment reliably. I wrote about starting this journey in my March 2026 newsletter. The short version is that robotics sits at the intersection of everything I find interesting: AI decision-making, physical systems, and real-world constraints that force you to think differently than you do when writing pure software.

The longer-term goal is to understand embodied AI: systems that don't just process data but interact with the physical world. The gap between a language model and a robot that can reliably complete a task in a messy environment is enormous, and I want to understand why.

This is early-stage work. More to come.

---

Outcome: Began building and programming autonomous robots, focusing on control systems, sensors, and the fundamentals of making machines navigate physical environments reliably.

What I took from it:

• Hardware is humbling in a way software isn't. Errors aren't reversible, the feedback loop is slow, and the machine doesn't care how elegant your logic is if the motors can't execute it.
• The gap between an AI making a decision and a robot carrying it out in the real world is much larger than it looks from the outside. That gap is where the interesting work is.
• Physical constraints force a different mode of thinking. In software you can abstract almost anything. In robotics, the physics always win.