Are humans still writing GPU kernels, or are self-improving loops taking over?

julsimon · x · 2026-10-14

Julien Simon asks whether anyone still writes kernels by hand instead of running self-improving loops. Quoting a SemiAnalysis podcast: is GLM 5.3 worth sounding the alarms — "as long as you have the budget, it will write faster kernels than humans" — plus sparse attention, the unpriced memory bottleneck, and why cheaper tokens may mean more HBM demand, not less.

Original post →

More from coding & agent

coding & agent channel →