Will AI agents replace knowledge workers, and what should people do about it?

AI agents are not expected to replace knowledge workers entirely, but they will fundamentally "refactor" their roles. Creators generally agree that while specific tasks and lower-context roles may disappear, the overall demand for expert services will likely increase as AI makes them cheaper and faster to produce.

The Refactoring of Knowledge Work

Instead of mass unemployment, many creators predict a shift from "doing" work to "managing" it.
* The Jevons Paradox: Y Combinator explains that just as containerization exploded global trade by making shipping cheaper, AI will lower the cost of services like legal drafting and engineering, which will likely reveal a massive "pent-up demand" for those experts rather than eliminating them Y Combinator — What Everyone Is Getting Wrong About AI And Jobs @ 03:02.
* From Doer to Manager: The consensus is that knowledge workers will move from manual execution to supervising "swarms" of agents. Alexandr Wang notes that this role has a specific name in the current workforce: manager Y Combinator — Alexandr Wang: Building Scale AI, Transforming Work With Agents & Competing With China @ 18:23.
* Empowered Solo Workers: Julie Zhuo highlights a trend toward "flattening" organizations where individuals become more empowered. A single person can now use AI companions to handle design, coding, and data analysis simultaneously, tasks that previously required a full team Lenny's Podcast — From managing people to managing AI: The leadership skills everyone needs now | Julie Zhuo @ 12:10.

What Knowledge Workers Should Do

To adapt, creators suggest shifting focus from technical execution to high-level strategy and "soft" human skills.
* Develop "Agent Management" Skills: The most valuable skill in the near future will be the ability to define success criteria crystal-clearly for agents Greg Isenberg — Most Valuable Skill of 2026: Managing AI Agents @ 01:00. This involves writing "evals" and objective criteria so you know without question whether an agent has succeeded Lenny's Podcast — From managing people to managing AI: The leadership skills everyone needs now | Julie Zhuo @ 12:10.
* Lean into Curiosity and "Agency": Dan Shipper advises workers to find their "moment of joy" with AI tools by experimenting with them rather than fearing them Lenny's Podcast — The AI paradox: More automation, more humans, more work | Dan Shipper @ 40:20. The "only universal skill" moving forward is being able to roll with changes and exercise human agency to decide which problems are worth solving Lenny's Podcast — An AI state of the union: We’ve passed the inflection point & dark factories are coming @ 30:30.
* Identify Boring Bottlenecks: Greg Isenberg suggests looking for "mundane, repetitive tasks" like data reconciliation or manual updates—these are the first areas where agents will provide massive leverage Greg Isenberg — The best AI startup opportunities will be automating boring mundane tasks of knowledge workers @ 00:00.

Where they disagree:
While most are optimistic, some offer starker warnings. Paul Buchheit predicts that "Zoom-based workers"—those whose work is entirely digital—could be "transparently replaced" by AI deepfakes within 10 years Y Combinator — Gmail Creator Paul Buchheit On AGI, Open Source Models, Freedom @ 33:38. Similarly, Greg Isenberg frames the transition as "Darwinism at its finest," warning that those who fail to change their baseline skill level will simply become less valuable to the market Greg Isenberg — Nicolas Cole explains why it is a misconception that AI is taking people’s jobs away @ 00:00.

— Sources: 12 videos across 5 creators