The long way is the shortcut: friction, mentorship, and experimentation in AI | Ethan Mollick

Nobody knows anything about AI. Not the labs, not the CEOs. Wharton professor and Co‑Intelligence author Ethan Mollick has a front‑row seat to AI’s most chaotic phase yet. In this episode of Strange Loop, he traces how we’ve raced from classic machine learning to chatty copilots and now to agents that can rewrite how organizations work, often before leaders realize what’s changed. Ethan and Sana’s Lauren Crichton explore his idea of the “jagged frontier” of AI abilities, the rise of dark‑factory-style workflows, and why IT can become a graveyard for ambitious AI projects. Along the way, they dig into the collapse of apprenticeships, what it really means to build an AI‑native company, and why our response to this technology may matter more than the technology itself. Topics covered: – The three eras of AI and what really changed with agents – The jagged frontier of AI abilities – Dark factories, token burn, and when to keep humans in the loop – Apprenticeship collapse and rethinking how people learn at work – Leadership, Lab, and Crowd as a simple frame for AI‑native orgs Timestamps (00:00) Nobody knows anything (03:40) From co‑intelligence to agents and long‑running work (06:55) The jagged frontier: where AI is brilliant and where it breaks (09:02) Wizards, opacity, and why AI decisions feel like magic (11:27) Recursive self‑improvement, takeoff risk, and anthropomorphizing AIs (13:56) Organizational context, implicit knowledge, and limits of agents (16:02) StrongDM’s “software dark factory” and full‑stack AI coding (18:19) Experts, apprenticeship collapse, and training the next generation (21:02) Agency, incentives, and adding the right kind of friction (22:46) AI in education: tutors, cheating, and real learning (24:48) Formal assessment at work and the trap of productivity‑only KPIs (30:47) IT departments, de‑weirding AI, and why playbooks fail (34:34) Tokens, costs, and new metrics for innovation (37:36) Interfaces as the bottleneck and designing beyond chat (41:29) Deep knowledge, wide knowledge, taste, and agency (43:06) Jobs, the O‑ring model, and living through chaotic transitions (46:26) Management as an AI superpower and the rise of builders (48:34) Rethinking org charts, cross‑functional teams, and agents (50:36) Parenting with AI and raising flexible kids (52:31) Unanswered questions, exponential curves, and what Ethan’s proud of About Strange Loop Strange Loop is a podcast about how artificial intelligence is reshaping the systems we live and work in. Each episode features deep, unscripted conversations with thinkers and builders reimagining intelligence leadership, and the architectures of progress. The goal is not just to follow AI’s trajectory, but to question the assumptions guiding it. Subscribe for more conversations at the edge of AI and human knowledge.