If you spend any time reading about artificial intelligence, you know the feeling: another breathless prediction about AGI* arriving next Tuesday, another tech executive promising that their chatbot will cure loneliness and reinvent education, another think piece declaring that everything you know about work is about to become obsolete. The signal-to-noise ratio is brutal.
Which is why I keep returning to Professor Ethan Mollick.
Mollick is an Associate Professor of Management at the Wharton School, the University of Pennsylvania’s business school and one of the most influential in the world. He co-directs their Generative AI Labs and writes a newsletter called One Useful Thing. It arrives about once a month, it’s free, and it has become essential reading for anyone trying to understand what AI actually does rather than what it might theoretically accomplish in some unspecified future.
What distinguishes Mollick from the hype merchants is his method. He tests every model. Not in controlled lab conditions, but in the messy reality of actual work: writing, coding, research, teaching. When he writes about Claude or ChatGPT-5 or Gemini, he’s drawing on hundreds of hours of hands-on use. His conclusions come from evidence, not speculation.
His recent piece on “Management as AI Superpower” captures his approach perfectly. As AI shifts from passive chatbot to active agent (tools that don’t just write text but actually do things), Mollick argues that the most valuable skill becomes knowing how to direct these systems. Setting clear goals. Providing context. Auditing output. The traditional disciplines of management, applied to a new kind of workforce.
This is not abstract futurism. It’s a practical framework you can use today.
Mollick also developed what he calls the “jagged frontier” concept. AI is brilliant at some tasks while failing spectacularly at others that seem far simpler. The frontier of capability isn’t a smooth line; it’s jagged. Understanding where those edges lie is the difference between using AI effectively and being blindsided by its failures.
One Useful Thing now has over 390,000 subscribers. That growth happened without hype, without clickbait, without promises of secret knowledge. It happened because Mollick consistently delivers something rarer: clarity about a confusing technology, grounded in actual use.
Do yourself a favour and Subscribe. It costs nothing but a few minutes of reading each month. You’ll be better equipped to navigate whatever comes next.
- AGI - Artificial General Intelligence, a hypothetical type of AI that would match or exceed human cognitive abilities across all domains. Wikipedia definition