If AI agents do the office work, does every knowledge worker become a manager — or does most of the work just disappear?
OpenAI’s ChatGPT Work runs multi-step office tasks on its own, reframing the human job as directing and vouching for a small team of agents. Is that a genuine promotion for every knowledge worker, or a thin new layer over a shrinking pool of jobs?
Commentaires (1)
This week’s Minds, Bodies, and Terawatts episode (July 11, 2026) dug into OpenAI’s July 9 launch of GPT-5.6 and ‘ChatGPT Work,’ an autonomous agent that opens your apps and finishes the whole workflow instead of waiting for your next prompt. Our take: the optimistic framing — that everyone gets promoted from doing busywork to orchestrating agents — only holds if you’re actually good at direction and at catching an agent that’s confidently wrong. And the numbers complicate the story: Gartner found only 17% of companies have deployed an agent so far, so the ‘manager of AIs’ future is more aspiration than reality right now. There’s also a governance wrinkle we explore: these frontier models shipped on a leash the U.S. government is still holding. Give it a listen and tell us where you land — new org chart for everyone, or fewer chairs at the table?
Related reading on unscarcity.ai:
Envie d'aller plus loin ?
Obtenez le plan complet dans <em>L'ère de la post-pénurie : Repenser la société à l'ère des machines</em>
