Hire for Slope

Attributed to Paul Buchheit by Sam Altman in Garry Tan’s fireside interview at Y Combinator’s AI Startup School, June 16, 2025 (video; excerpted by the Startup Archive on July 31, 2025), where Altman recalls the line and calls it “unbelievably great advice.” A primary source from Buchheit in his own voice (a blog post, a tweet, a recorded talk) has not been located; the phrase reaches readers through Altman’s recollection.

The metaphor predates the hiring application. John Ousterhout, a Stanford computer science professor, used the same framing in a Spring 2012 “Thought for the Weekend” for his CS 142 class, titled “A Little Bit of Slope Makes Up For a Lot of Y-intercept” (class page), arguing that how fast you learn matters more than how much you start with. Buchheit’s reframe applies the same math to who to hire: weight the rate of change over the current snapshot, because two candidates with different slopes cross at some future point and the higher-slope hire is ahead for the rest of a multi-year tenure.