Amazon Memo Calls Token-Usage AI Adoption Metrics 'Deeply Flawed,' Pitches 'Renaissance Developer'

Amazon circulated an internal memo in September — and posted it publicly on Oct 27 — calling token usage and lines-of-code-generated 'deeply flawed' as proxies for AI adoption because they push engineers toward tokenmaxxing. The company already shut down an internal leaderboard in May after finding staff were gaming the meter. Guidance tells engineers to use AI selectively and judge value by long-term customer and business outcomes; the memo frames the target as a 'Renaissance Developer' who pairs AI with judgment. Amazon says it has not capped per-engineer token spend.