Discussion

Kevin Roose makes the case for AI optimism, one dozen agents at a time

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#15235

Tech reporter Kevin Roose says his six-person media startup uses about a dozen AI agents for work including building tools, scheduling and preparing podcast episodes for different platforms. His example gives a concrete shape to the optimistic case for AI: small teams may take on work that once required a larger staff, if the tools prove useful enough to trust.

Watch Desk analysis

What happened

In an interview with David Remnick for The New Yorker, Roose describes agents helping with the unglamorous machinery of podcast production: creating different versions for Apple, Spotify and YouTube, along with captions, titles, descriptions and metadata. He says the work used to be done by hand. He also describes using Claude for coding and editorial feedback, while acknowledging that much of that feedback is poor.

Roose’s broader argument is that investors are betting AI tools will keep improving and spread further through white-collar work. He says his small company already has more agents than people in its Slack. That is his account of one startup, not evidence that every business can replace staff or reproduce the same results.

Why it matters

The discussion shifts the AI jobs argument from abstract forecasts to a practical question: what work can a small team actually hand to software? Roose’s example is not a robot workforce running a company unsupervised. It is a collection of tools taking on coding and production tasks, with people still directing the work and checking what comes back.

That distinction matters for anyone weighing the promise of AI agents against the realities of adopting them. A smaller team may produce more, but the interview does not establish how much time or money the agents save, or how reliably they perform.

Our read

This is a useful, unusually tangible account of the agent economy people keep forecasting. Roose offers a glimpse of how it might work, not a blueprint that has already been proven at scale. The sensible takeaway is to look for specific tasks, human oversight and results, rather than count the agents in the Slack channel and declare the future settled.

What to watch

  • Whether Roose’s company shares more about the tools’ performance and the work they take on.
  • Whether similar small teams report measurable gains, not just impressive demos.
  • How much review and correction the agents’ output still needs.

Discussion spark: If a small team can use AI agents to take on coding and production tasks, should we judge that mainly by the jobs it replaces or by the work it lets the team do?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.