JPMorgan's Tech100 just gave us a preview of how fast AI coding agents will hollow out white-collar headcount.

The Summary

  • Jack Dorsey told JPMorgan's Tech100 conference he's using the AI coding agent Goose for "a few hours every morning" and now plans to cut nearly half his company's workforce
  • Databricks CEO Ali Ghodsi confirmed similar habits and pressure on his team to match that productivity curve
  • This wasn't a panel about AI. This was executives casually discussing layoffs like a morning routine.

The Signal

The Iran war got five seconds. Space data centers got a keynote. But the real news happened when Dorsey admitted he's already restructuring Block around what one AI agent can do before breakfast.

Goose, for context, is an open-source coding agent. Not some proprietary enterprise system. Not a custom-trained model that took millions to build. A tool anyone can download. And it's productive enough that a CEO looked at his org chart and saw 50% redundancy.

Ghodsi's comment is the tell. When the CEO of Databricks, a company built on helping others wrangle AI infrastructure, says his own team is feeling pressure to keep up with what agents can do, that's not speculation about 2030. That's happening now. The race isn't "will agents replace developers" anymore. It's "how fast can you reorganize before your competitor does."

Jeff Bezos talking orbital real estate is interesting. Lauren Bezos wanting less social media is fine. But Dorsey treating a 50% workforce reduction as a casual byproduct of his new morning habit is the most honest thing said at that conference. No hand-wringing about reskilling. No platitudes about human creativity. Just math. If an agent can do the work in three hours that used to take a team three weeks, the team shrinks or disappears.

The Implication

If you're a knowledge worker whose job is turning instructions into code, content, or analysis, your runway just got shorter. Not because AI will replace you someday. Because your CEO's CEO just heard Dorsey say he's doing it now, and they're wondering why they're not. The companies moving fastest won't be the ones with the best AI strategy. They'll be the ones willing to cut deepest while calling it optimization. Watch your leadership's conference schedules. If they're at events like this, they're getting ideas.

Source: The Information