Plain English with Derek Thompson What Happens When AI Learns to Do Our Jobs
573 snips
Oct 24, 2025 
Ethan Mollick, a Wharton professor focused on entrepreneurship and innovation, dives deep into the evolving role of AI in our work lives. He explores how AI's capabilities have advanced and discusses its limitations, including areas where it still struggles. Mollick highlights the implications for jobs, especially those involving repetitive or easily assignable tasks. He shares practical strategies for using AI effectively and warns about skill atrophy with reliance on technology. The conversation paints a picture of a disruptive transition rather than an overnight overhaul.
AI Snips
Chapters
Books
Transcript
Episode notes
Agents Can Execute Long, Self-Correcting Plans
- Agents autonomously pursue goals by chaining many steps and using tools.
- Self-correction lets agents complete long step sequences without constant human fixes.
Agents Shine Where Output Is Verifiable
- Agents excel at verifiable, testable outputs like code, Excel, and PowerPoint.
- They struggle more with fuzzy tasks where quality is subjective and hard to rate.
Prompt AI To Critique Not Just Edit
- Prompt AI to be critical and explicitly ask for missing arguments or naive-reader confusions.
- Treat AI as an editor by requesting targeted, critical feedback on structure and gaps.
