Descriptions:
Ethan Mollick, professor at the Wharton School and author of the Substack ‘One Useful Thing,’ joins Scott Galloway on Prof G Conversations to examine why corporate leadership is broadly mishandling AI adoption and what a more effective approach would look like. The conversation opens with Dario Amodei’s 38-page essay on AI risk: Mollick defends it as a sincere rather than performative document, while arguing that near-term harms — how AI reshapes work, education, and social behavior — deserve more attention than speculative existential timelines.
A highlight is Mollick’s analysis of Chinese open-weight AI models, including the Qwen family. He identifies an economic paradox: unlike open-source software, where free code drives paid services, AI companies releasing open-weight models receive no clear ancillary revenue benefit. He suggests this ‘AI dumping’ dynamic — releasing high-capability models at zero cost to undercut US competitors — may be state-influenced, while cautioning that Chinese frontier models remain roughly eight months behind US equivalents on a curve that is still accelerating.
Mollick and Galloway also discuss why enterprise AI adoption remains ‘woefully underpenetrated’ at the organizational level even as individual usage grows, the difference between using AI to empower employees versus simply reducing headcount, and the risk of enterprises locking into cheaper but less capable models when the capability curve has not yet plateaued. The episode references Section AI, DeepSeek, Anthropic, and the broader AI supply chain from chips through application layer.
📺 Source: The Prof G Pod – Scott Galloway · Published February 12, 2026
🏷️ Format: Interview
