Writing
My Writings
Things I've thought about and decided to write down.
The Second Law Is the Only Law
Why entropy runs through everything.
Apr 2026
7–9 mins
Symmetry First
The physics principle that makes ML work better.
Feb 2026
6–8 mins
Latent Space Is Not Understanding
What AI gets right and what it misses.
Jan 2026
5–7 mins
Why Most AI-for-Science Papers Won't Reproduce
The gap between benchmarks and experiments.
Nov 2025
8–10 mins
The Model Is Confident. The Model Is Wrong.
On overconfidence in scientific ML.
Oct 2025
5–6 mins
The Three-Body Problem Is Not a Metaphor
Three different reasons a system resists prediction.
Jul 2025
5–6 mins
What Feynman Meant
On knowing versus being able to produce.
Mar 2025
4–5 mins
Everything Interesting Happens at an Interface
Why failures happen at the boundary, not the bulk.
Mar 2025
6–7 mins
The Landscape Problem
Chemistry and ML share the same hard problem.
Jan 2025
5–6 mins
Approximation All the Way Down
Every simulation trades accuracy for tractability.
Nov 2024
4–5 mins