
teach a robot Open a door and it should open the opportunity of a lifetime. Not so for Everyday Robots, one of Alphabet’s youngest subsidiaries. Just one year after graduating from Alphabet’s X moonshot lab, the team trained more than a hundred wheeled, single-armed robots to clean cafeteria tables, separate trash and recycle, and yes, open doors, as part of a sprawling budget cut, A spokeswoman for Google’s parent company, which is shutting down, confirmed this.
“Everyday Robots will no longer be a standalone project within Alphabet,” said Denise Gamboa, Director of Marketing and Communications, Everyday Robots. “Some technologies and parts of the team will be integrated into existing robotics at Google Research.”
The robotics venture is the latest failed bet from X, which in the past decade has also launched internet-transmitting balloons (Loon) and power-generating kites (Makani) before deciding they were not commercially viable to stay afloat. Other former X projects like Waymo (developing self-driving cars) and Wing (testing grocery delivery drones) have continued as Alphabet-owned companies, though their financial prospects remain mired in regulatory and technical challenges. Like Everyday Robots, these companies leverage new technologies that have shown impressive promise in trials, but not so much reliability.
A decade ago, Everyday Robots emerged from the ashes of at least eight robotics acquisitions by Google. Google co-founders Larry Page and Sergey Brin expect machine learning to reshape robotics, with Page specifically looking to develop robots for consumers, a former employee who was involved at the time said on condition of anonymity to discuss internal deliberations . By 2016, they put software entrepreneur Hans Peter Brøndmo in charge of a project called Help (later called Moxie at one point) to use machine learning to develop robots that could handle everyday tasks and adapt to different environments, sources said.
The team built arm farms and playpens, where a fleet of robots would repeat the same tasks — like sorting trash — over several months. It’s a brute-force attempt to generate data to train a machine-learning model and then equip the robot with the expertise needed to interact with the world around it using cameras, arms, wheels, and finger-like handles. This novelty frees engineers from the traditional approach of robotics, in which specific instructions must be written for machines to follow every potentially small scenario. This idea applies mostly to initial tasks. Google has a fleet of Everyday Robots to help clean the search giant’s restaurants and check meeting rooms for clutter during the pandemic.
provided by google
Last year, Everyday Robots demonstrated further progress with Google AI researchers. The project integrates large-scale language models similar to the underlying ChatGPT into robotic systems, enabling robotic assistants, for example, to respond to someone saying they’re hungry by fetching them a bag of chips. But Google and Everyday Robots emphasized at the time that on-call roving butlers were far from consumer usability. Changes that seem trivial to humans, such as the type of lighting in a room or the shape of a potato chip bag, can cause malfunctions.
From its inception, Everyday Robots has struggled with whether its mission was to conduct advanced research or bring products to market, the former employee said. It employs more than 200 people, including those who oversee customer operations, teach robots to dance and tinker with perfect designs. Robotics experts estimate that each of its robots can cost tens of thousands of dollars.
The fees are too high for Alphabet, whose more speculative “other bets” such as Everyday Robots and Waymo lost about $6.1 billion last year. Activist investors have been pressing the company to cut spending after Google’s ad spending slowed and Alphabet’s overall profit fell 21 percent to $60 billion last year. On Jan. 20, Alphabet announced it would lay off about 12,000 employees, or 6% of its workforce. Everyday Robots was one of the few projects that was disbanded.