
10x Engineer Myth and Better Interviews

Stop Inflicting Daily Standups and Computer Science Riddles on Your Devs - David Guttman
Topic 6 of 8
Published Feb 3, 2025 · Clip 25:51–43:12 (17:22)
David debunks the 10x engineer myth, tracing it to a 1960s U.S. Air Force study on mainframe-era developers that measured a 10x spread between the best and worst developers, not the top versus average—so the lesson is to avoid 0.1x developers rather than chase mythical 10x ones. He introduces the commando/infantry/police model: commandos solve rare novel problems, infantry execute plans cooperatively (what most companies need), and police maintain stable systems with documentation and procedures. He then criticizes FAANG-style riddle interviews—hiring is predicting the future, so test candidates on problems as close as possible to the actual work; Superstruct uses short unpaid filters and paid five-hour take-home challenges, ideally with minimal setup (Docker can help) focused on the team's real constraints.