Key Takeaways:
- Boston Dynamics is leading the development of humanoid robots with its Atlas model, which is being tested for widespread use.
- Atlas is being trained using machine learning and supervised learning techniques to master complex tasks.
- The market for humanoids is predicted to reach $38 billion within the decade, with Boston Dynamics and other US robot makers competing against state-supported Chinese companies.
- Humanoid robots like Atlas are expected to change the nature of work, taking over repetitive and backbreaking labor, but also requiring management, training, and servicing.
- Experts believe that humanoid robots will not become sentient or rogue, but will instead augment human capabilities and provide superhuman abilities.
Introduction to Humanoid Robots
The development of humanoid robots is rapidly advancing, with Boston Dynamics at the forefront of this technology. With the support of South Korean carmaker Hyundai, which owns an 88% stake in Boston Dynamics, the company is testing a new generation of its humanoid robot, Atlas. As Scott Kuindersma, head of robotics research at Boston Dynamics, notes, "A lot of this has to do with how we’re going about programming these robots now, where it’s more about teaching, and demonstrations, and machine learning than manual programming." This shift in approach has enabled Atlas to learn complex movements that were previously thought to be impossible for a machine.
Training Atlas
The training process for Atlas involves several techniques, including supervised learning and motion capture. Machine learning scientist Kevin Bergamin demonstrated an example of supervised learning, where he took direct control of the humanoid and guided its hands and arms through each task until Atlas succeeded. As Kuindersma explained, "That generates data that we can use to train the robot’s AI models to then later do that task autonomously." Another technique involves using a motion capture body suit, which was worn by 60 Minutes correspondent Bill Whitaker while performing jumping jacks. The data collected from the suit was fed into Boston Dynamics’ machine learning process, allowing Atlas to learn the motion and perform it autonomously.
Limitations and Future Developments
While Atlas has made significant progress, there are still limitations to its capabilities. As Kuindersma noted, "There are no humanoids that do [routine tasks] nearly as well as a person." However, he believes that "we see a pathway to get there." The future of humanoids is exciting, with Boston Dynamics CEO Robert Playter stating, "There’s a lot of excitement in the industry right now about the potential of building robots that are smart enough to really become general purpose." The market for humanoids is predicted to reach $38 billion within the decade, with Boston Dynamics and other US robot makers competing against state-supported Chinese companies.
The Future of Work
As humanoid robots like Atlas become more advanced, they are expected to change the nature of work. Playter predicted that "the really repetitive, really backbreaking labor is really, is going to end up being done by robots." However, he also noted that these robots will not be autonomous and will require management, training, and servicing. This will create new job opportunities and require humans to work alongside robots. As Playter said, "We would like [robots] that could be stronger than us or tolerate more heat than us or definitely go into a dangerous place where we shouldn’t be going. So you really want superhuman capabilities."
Addressing Concerns
Some people may worry about the potential risks of humanoid robots, such as displacement of workers or the possibility of sentient robots. However, Playter believes that these concerns are unfounded. As he said, "[If you] saw how hard we have to work to get the robots to just do some of the straightforward tasks we want them to do, that would dispel that worry about sentience and rogue robots." Instead, he believes that humanoid robots will augment human capabilities and provide superhuman abilities, making them a valuable tool for various industries and applications. As Kuindersma noted, "I think the thing that’s really exciting now is we see a pathway to get there," and with continued advancements in machine learning and robotics, the future of humanoid robots looks promising.
https://www.cbsnews.com/news/boston-dynamics-training-ai-humanoids-to-perform-human-jobs-60-minutes/