What's really happening inside OpenAI's Codex revamp when they shipped a desktop agent that can drive any Mac app in the background while you do other work?
The common story is that this is a coding tool update — but the reality is that Codex shifted categories entirely, and the gap to Claude's computer use is wider than I expected after running them side by side for a week.
In this episode, I share the inside scoop on what OpenAI is really building and why it looks so different from Anthropic:
• Why Codex finishes in two minutes what takes Claude five or six with fumbles and retries
• How the Workflow-to-Shortcuts-to-Sky team made background agents actually usable
• What Chronicle tells you about training signal for computer use
• Where Conway fits into Anthropic's bet that the ecosystem will cooperate
Leaders who keep waiting for vendors to ship agent-ready interfaces are missing that Codex doesn't need the software industry to build for agents — the body just uses whatever's already there.
Subscribe for daily AI strategy and news.
For deeper playbooks and analysis: https://natesnewsletter.substack.com/
Full Story w/ Prompt Kit: https://natesnewsletter.substack.com/p/grab-the-workflow-audit-that-tells
","author_name":"Nate B. Jones"}