14 More lessons from 14 years at Google
Addy Osmani's second set of lessons from Google, focused on organizational dynamics rather than individual technical skills. The standout insights: meetings should start with "approve, choose, unblock, or inform"; reliability is a product feature (treat it like one); cross-team pain comes from unclear boundaries, not insufficient communication; and recurring heroism is a failure mode, not a badge of honor.
Key Quotes
"The fastest engineering team I ever worked with wasn't the one with the best programmers. It was the one where decisions happened in hours instead of weeks."
"If one person saving the day is a recurring pattern, that's a failure mode rather than a badge of honor."
"You wouldn't ship a feature without product review. Don't ship a system without reliability discussion."
"The work is about making it easier for normal people to do extraordinary things on a normal day."
Key Themes
The "approve, choose, unblock, or inform" framework for meetings is immediately actionable. Most meetings fail because nobody agrees on what the meeting is for. Starting with the ask eliminates ambiguity.
The anti-hero-culture argument is important and counterintuitive in tech. "Sustainable teams design normal operations that don't require exceptional effort" -- if you need heroes, your systems are broken.
The trust-as-latency-optimization metaphor is elegant: high-trust environments accelerate decisions because you don't need approval chains for every choice.
Critical Analysis
This is the organizational companion to How I've Run Major Projects. Where Kuhn focuses on the DRI's toolkit, Osmani focuses on the system the DRI operates within. The "team interfaces" lesson (unclear boundaries cause more pain than unclear communication) is the deepest insight. Adding more Slack channels and sync meetings is the organizational equivalent of adding more logging to a broken system -- it feels productive but doesn't address the root cause.
The AI curation lesson ("as code generation is commodified, taste and selection distinguish valuable engineers") connects to the entire AI-enhanced coding cluster in this batch. It's the management perspective on what AI Zealotry, The Next Two Years of Software Engineering, and Addy Osmani's Workflow are all saying from the developer perspective: judgment matters more than production capacity.
The "small PRs" advocacy connects to AI PR Reviewer and Learn from PRs Skill -- both tools work better with focused, reviewable changes than with monolithic diffs.
Sources: 14 More lessons from 14 years at Google Last updated: 2026-05-14