Susan Zhang Mocks 1T Lines of Slop Code Goal
The AI engineer sarcastically critiques using code volume to measure math progress.
TLDR
Susan Zhang, a research engineer at Google DeepMind with prior roles at Meta AI and OpenAI on systems including OPT-175B, posted the quoted remark. She describes a mathematician entering an AI systems setting to demand credit for the goal of producing one trillion lines of slop code, based on extrapolation from two points. The post contrasts this with traditional software and states that lines of code are not a strong metric of progress or significance in mathematics.
Combined views
580
1 Source, first seen 25d ago