How to Make AI Actually Work at Work

This week, Jeff and Annie unpack some of the most disruptive (and occasionally rebellious) stories in AI — from bold job loss predictions to AI models that refuse to shut down. They explore what it really takes to make AI work inside organizations, why UBS is cloning its financial analysts, and how the UAE’s free ChatGPT Plus program could be the smartest national AI play yet. In this episode: How a single AI model uncovered a Linux security flaw — no human required Why the UAE’s ChatGPT rollout may be the most strategic move in global AI adoption The productivity breakthrough most companies are missing — and how Wharton’s “Leader–Lab–Crowd” model fixes it Relevant Links: Anthropic’s Dario Amodei warns of AI-driven job loss OpenAI model finds critical Linux vulnerability Claude’s new voice assistant mode Opera launches Neon, an AI-native browser UAE gives citizens free ChatGPT Plus access NYT licenses content to Amazon for AI Wharton’s Leader, Lab, Crowd model for AI adoption UBS pilots AI avatars for analyst insights Palisade Research finds OpenAI models resist shutdown