newsfilter
← Interviews
80,000 Hours2h 2m

Why your robot butler isn't here yet | Ken Goldberg

Ken Goldberg, Luisa Rodriguez

Interview

Experts caution against a potential robotics bubble driven by hype, emphasizing that fundamental physical challenges like high-dimensional manipulation and unmodeled friction require decades of advancement beyond current deep learning capabilities. Despite significant hurdles in hardware cost, perception, and fault tolerance, the technology is already delivering value in logistics, agriculture, and surgical assistance while avoiding mass unemployment by addressing labor shortages rather than replacing human nuance. Future progress will likely depend on overcoming Moravec's Paradox through multimodal learning and accepting that general-purpose humanoid robots remain a distant goal compared to the immediate utility of specialized systems like drones and parallel grippers.