The Skill That Separates AI Power Users From Everyone Else (Why "Clear" Specs Produce Broken Output)

What's really happening with AI coding tools and how we work alongside them? The common story is that Claude Code and Codex are just competing products, but the reality is more complicated when the difference between a CNC machine and a skilled machinist defines two entirely different relationships with AI. In this video, I share the inside scoop on why the colleague versus tool distinction will define AI adoption across all knowledge work:

• Why Codex works like a CNC machine and Claude Code works like a machinist, and why that metaphor matters more than any benchmark comparison

• How senior engineers get compound leverage from autonomous agents precisely because they know what they know and are honest about when they don't

• What happens when you can't specify precise intent upfront and why that determines which tool you should actually reach for

• Why this same dynamic will shape all non-technical knowledge work as colleague-shaped AI moves beyond the codebase

For individuals and organizations navigating 2026, Cursor ran ChatGPT 5.2 for a week straight and produced three million lines of Rust code with no human touching the keyboard. The question isn't which AI is better. It's whether you're honest about which situation you're actually in.

Subscribe for daily AI strategy and news.

Hosted on Acast. See acast.com/privacy for more information.