Joelle Pineau is the chief AI officer at Cohere. Pineau joins Big Technology Podcast to discuss where the cutting edge of AI research is headed — and what it will take to move from impressive demos to reliable agents. Tune in to hear why memory, world models, and more efficient reasoning are emerging as the next big frontiers, plus what current approaches are missing. We also cover the “capability overhang” in enterprise AI, why consumer assistants still aren’t lighting the world on fire, what AI sovereignty actually means, and whether the major labs can ever pull away from each other. Hit play for a cool-headed, deeply practical look at what’s next for AI and how it gets deployed in the real world. Chapters: 0:00 - Intro: Where AI research is heading 2:02 - Current cutting edge of AI research 5:14 - Memory vs. continual learning in AI models 9:46 - Why memory is so difficult 14:23 - State of reasoning and hierarchical planning 17:19 - How LLMs think ahead while generating text 19:00 - World models and embedded physics understanding 21:32 - Physical vs. digital world models 24:13 - Do models need to understand gravity for AGI? 25:51 - The capability overhang in AI 28:22 - Why consumer AI assistants aren't taking off 30:42 - Companies vs. individuals adopting AI 31:44 - Why AI labs stay neck-and-neck competitively 33:41 - Commercial applications of AI in enterprise 38:11 - Impact on entry-level and mid-career employees 41:12 - AI coding agents and vibe coding 43:13 - Concentration of AI in big tech companies 46:23 - Social media entrepreneurs vs. research scientists 48:09 - Economics of AI advertising 51:02 - Can the pace of AI adoption keep up? --- Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice. Want a discount for Big Technology on Substack + Discord? Here’s 25% off for the first year: https://www.bigtechnology.com/subscribe?coupon=0843016b