Physical teleoperating robot could become an important job in the future, according to AI holy place, based in Vancouver, Canada. The company also believes this could provide a way to train robots how to perform tasks that are currently beyond their (machines) reach, and imbue machines with a physical sense of the world that some people might not be able to. The given person is needed to unlock the human level. artificial intelligence.
Industrial robots are powerful, precise, and mostly stupid. They can’t apply the kind of precision and responsiveness needed to perform sophisticated manipulative tasks. That’s part of the reason why the use of robots in factories is still relatively limited, and it still takes an army of workers to assemble all the confusing parts into the guts of an iPhone.
But when such work is nothing to a human, why not ignore the complexity of trying to design an algorithm to do the job?
This is one of Sanctuary’s robots—the upper half of a humanoid robot—that performs a series of complex maneuvers. Beyond the screen, a person wearing a VR headset and gloves with sensors is remotely controlling the robot.
near holy place ran what it calls the first “real world” test of one of its robots, by having a humanoid like this work in a store not far from the startup’s headquarters. The company believes that being able to do physical work remotely could help solve the labor shortage many companies are facing today.
Some robots have received some remote help from humans when they got stuck, like I wrote about. The limits of AI mean that robots working in restaurants, offices, and on the streets as delivery people are bewildered by unusual situations. For example, the difficulty of implementing fully autonomous driving means that some companies are working to get the remote pilot truck on the road.
Sanctuary’s founders, Geordie Rose and Suzanne Gilbert, run Kindred, another robot-based telemedicine company that was acquired in 2020 by Ocado, a UK supermarket company using extensive automation. In this video, the couple talks about company history and plans for the future.
The ultimate aim is to use data from humans controlling the robot remotely to teach algorithms to do more tasks automatically. Gilbert, Sanctuary’s CTO, believes that achieving human-like intelligence in machines will require them to interact with and learn from the physical world. (Sorry, ChatGPT.)
openAI, the company behind ChatgGPT, is also interested in remotely controlled humanoid figures. It is leading a $23.5 million investment in 1X, a startup developing human-like robots. “The OpenAI Startup Fund believes in the approach and impact 1X can have on the future of work,” said Brad Lightcap, COO of OpenAI and manager of the OpenAI Startup Fund.
Remote operating system ALOHA.Courtesy of Tony Zhao/UC Berkeley
In order for humans to be able to help robots remotely, AI may also need to be developed to ease cooperation between humans and machines. Chelsea Finnan assistant professor at UC Berkeley, recently shared details about a fascinating research project involves the use of machine learning to enable low-cost, remote-controlled robotic arms to operate smoothly and accurately. This technology could make it easier for humans to remotely control robots in more situations.
I don’t think I’d enjoy controlling a robot remotely all day—especially if I knew that one day the robot would come back and kick me out the door. But it could make working from home more accessible and also make certain types of work more accessible. Alternatively, we may have just glimpsed a backward future for the workplace.
