Four counterintuitive leadership lessons

90 minutes of unfiltered product advice from Snap and Discord’s product chief | Peter Sellis

Topic 9 of 14

Published Sep 20, 2026 · Clip 46:24–1:02:13 (15:49)

Peter walks through four counterintuitive lessons. He 'rides his best people into the ground,' handing strong performers more and more decisions until they start making bad ones, always defends his people in public, and admits he has no good mechanism for improving average performers so he lets them fail; he also warns that team principles no one can disagree with aren't good principles, which covers collaboration. Growth, he argues, comes from the core: network-effect businesses grow implicitly, DAU/MAU is a form of retention, and moving the ratio a few basis points beats acquiring new MAU — proven by Snap's Android performance push after the 2018 redesign and Discord's 2024 refocus on gaming friends, both of which produced growth renaissances; 'there's always money in the banana stand,' and gamers are the most demanding users you can serve. On when to bet on something new: a 2x2 of core strengths x other niches, watching users 'vote with their feet' in unexpected ways, or rare blank-sheet backcasting from the future unencumbered by the present. The host offers Airbnb as proof nothing beats the core; Peter notes the risk is a 10x black-swan disruption — the Thanksgiving turkey — and offers SpaceX as a company that guards against it.