A tenth engineer at Facebook recounts the brutal reality of early startup life — 120-hour weeks, sleeping with a phone to babysit News Feed scripts every four hours, and constant server and memory constraints. That scarcity forced extreme focus: every engineer knew their work was the single most important thing. Mark Zuckerberg institutionalized this by rejecting charitable initiatives, arguing the company's comparative advantage was building free products that nonprofits could leverage, not making donations directly.

Over time, that discipline eroded. With more money and employee pressure, leadership stopped saying no. Each digression — charitable programs, niche product features — appeared reasonable in isolation with positive local ROI. But collectively, hundreds of such decisions created drag: cluttered UI, bloated teams, slower products, and no clear path forward. The author notes the same dynamic in product development: a small feature for a subset seems harmless until repeated a hundred times.

The piece identifies a pernicious selection bias: customers acquired for niche features revolt when they're removed, and employees resist losing perks. Surviving requires willingness to disappoint. The author doesn't miss the sleep deprivation but misses the "profound sense of focus" — and argues the best time to kill a distraction is before it starts.