AI Reshapes Work Beyond 90/10 Productivity Gains as Managers Struggle to Judge Talent
Updated
Updated · The New Yorker · Oct 2
AI Reshapes Work Beyond 90/10 Productivity Gains as Managers Struggle to Judge Talent
3 articles · Updated · The New Yorker · Oct 2
Summary
Early evidence suggests AI is changing work less through mass job destruction than by making many employees faster and superficially better, while leaving managers less able to tell who is truly creative, trustworthy or ready for authority.
The analysis centers on a “90/10 production function”: AI can get workers to roughly 90% quality quickly, but often makes the last stretch to excellence harder, especially in “star tasks” where top performance matters.
That dynamic also threatens training pipelines, because junior staff often learn judgment, relationships and credibility through routine work that AI now automates, potentially weakening how firms develop experienced decision-makers.
Jobs built as “strong bundles” of linked tasks may still gain from AI through “robots above” and “robots below,” with humans handling exceptions or using AI to access expertise, while isolated, predictable roles remain more exposed.
The piece argues the near-term bottleneck is organizational adaptation: companies need workers who understand AI, day-to-day workflows and office politics, a scarce mix that may slow workplace transformation more than the technology itself.