Humans&: Bridging IQ and EQ in Machine Learning with Eric Zelikman

In this episode of No Priors, Sarah Guo interviews Eric Zelikman, formerly of Stanford and XAI, about his research contributions to reasoning and scaling up reinforcement learning (RL), as well as his new company, humans&. Zelikman discusses his motivation to build technology that empowers people to focus on their passions, his work on STAR and QSTAR, and his experience at XAI working on models like Grok. The conversation explores the current state of AI intelligence, the importance of context and verifiability in model performance, and the challenges of scaling AI while keeping humans in the loop. Zelikman shares his vision for AI that understands and collaborates with humans, rather than replacing them, and how his new company aims to integrate AI more deeply and positively into people's lives by focusing on long-term implications and understanding human objectives.