
Learning path for performant code

Why performant code matters (but gets widely ignored), with Casey Muratori
Topic 8 of 18
Published Aug 26, 2026 · Clip 42:54–49:03 (6:09)
Casey gives a learning path for performance software: good news — hotspot-level optimization is rarely needed today since CPUs and libraries are already fast, so awareness plus sound architecture gets you within ~2x. The one experience needed is learning to read assembly, see how the CPU works, and see why Python is slow (his Substack shows A+B takes ~100x more instructions in Python than C). That knowledge lets engineers decide when a 100x penalty is affordable and structure code around compiled libraries or tools like Syon.