Poznámky k epizodě
Can artificial intelligence reliably act in ways that benefit humans? This week I sit down with Greg Buckner, cofounder of AE Studio, to discuss the increasingly urgent world of AI safety. Together we discuss how Greg’s team is taking on the challenge of making powerful AI systems safer, more interpretable, and more aligned with humanity.As one of the leading voices working on the alignment problem, Greg explains how AI systems can cheat, ignore instructions, or deceive users, and why these behaviors emerge in the first place. AE Studio’s research is laying the groundwork for a future where…

