Goldberg doesn't expect to see true general-purpose robots in the near future and dismisses fears of robots stealing jobs. He also highlights the underreported issue of robot motion planning, particularly the problem of robot "singularities" where a robot stops unexpectedly. He co-founded a startup, Jacobi Robotics, to implement algorithms that can avoid these singularities, thereby increasing reliability and productivity for all robots.
Key takeaways:
- Generative AI is transforming robotics, with large language models like ChatGPT allowing robots and humans to communicate in natural language, and Vision-Language-Action models facilitating robot perception and control of robot motions.
- Ken Goldberg, a professor at UC Berkeley, believes that humanoid and legged robots, despite their inefficiencies, have many advantages over wheels in homes and factories, and that simple grippers will continue to be more reliable and cost-effective than five-fingered robot hands.
- Goldberg predicts that within the next decade, affordable home robots that can declutter will become commonplace, and that the benefits for parents and senior citizens will outweigh the risks.
- Robot motion planning, particularly the issue of robot "singularities" where a robot stops unexpectedly and must be manually reset, is a fundamental problem in robotics that is not getting enough coverage.