Where's the line on speeding up others' code? Anthropic's protein kernels spark debate
jmschreiber91 · x · 2026-10-12
A developer thread debates the ethics of auto-optimizing and republishing other people's code, sparked by a question about Anthropic's work on protein design tools.
jmschreiber91 draws the key distinction: Anthropic wrote optimized kernels for existing models with attribution (e.g., "we sped up protein-net 3x"), which is fine; releasing the same work as "anthropic-net" without crediting the original authors would be unacceptable. Attribution is the dividing line.
More from AGI Musings
- Matt Turck: from punch card operators to agent managers, jobs always adapt — mattturck · 2026-10-12
- 'The End of Work' isn't new: Rifkin published the same fear in 1995 — soumitrashukla9 · 2026-10-12
- 1964's Triple Revolution report warned computers were breaking the jobs-income link — soumitrashukla9 · 2026-10-12
- When Otis automated elevators in the 1950s, skilled operators lost high-paying jobs — soumitrashukla9 · 2026-10-12
- From Silent Looms to Dark Switchboards: Why Tech Change Has Always Been Personal for Skilled Workers — soumitrashukla9 · 2026-10-12
- Empirical checks will trump reasoning reproduction in the AI era, and mathematicians may help — johnvmcdonnell · 2026-10-12