
Survivorship bias and hacky internal tools

High growth startups: Uber and CloudKitchens with Charles-Axel Dein
Topic 15 of 17
Published Sep 24, 2025 · Clip 1:27:28–1:32:46 (5:18)
Charles agrees the chaos of hypergrowth likely contributed to Uber's success — speed in shipping has trade-offs in quality and maintainability, but building a successful product comes first. He illustrates with Uber's 'ping' mechanism (an inefficient pull design that worked for years until it was refactored to push) and Uber's user/trip tags and JSON attributes columns, which were hacky but bootstrapped many products. He argues internal tools should leave users in control — like spreadsheets, which support mass change and WYSIWYG — rather than being over-constrained, and that startups should optimize for speed and user control, with guard rails, since engineering time is the limiting factor.