Google aims to sell off its high-profile robotics acquisition Boston Dynamics. Does this mean it’s given up on robotics?
Not market ready
As Bloomberg and MIT Technology Review recently reported: Google is pulling out of Boston Dynamics because it doesn’t see a marketable and profitable product emerging from the company any time soon.
According to Boston Dynamics’ Marc Raibert, the kinds of problem the company is trying to solve are proving very hard, needing “a lot of work”.
Despite impressive results showing the robot ATLAS tackling difficult terrain, the company is still very far from being able to create robots that can work in a human environment and learn and improve while functioning correctly, 100% of the time.
The humanoid robotics housekeeper remains a futuristic dream.
See a video of ATLAS tackling difficult terrain below
Software is easier than hardware
What Google is learning across a multitude of projects is that creating and developing hardware solutions is much more difficult than creating good software.
For one, software is only supposed to work inside a computer, to run a set of tasks and maybe improve its ways, like the Go-playing champion AI computer AlphaGo which recently triumphed over human Lee Se-dol (4-1).
To build a robot that can think independently, that is, a robot that can show intentionality, at the same time as it moves and performs complex physical tasks in a real environment is a far more complicated matter.
According to Bloomberg:
“To develop robots, you have two options: You can either simulate an environment and robot with software and hope the results are accurate enough that you can load it into a machine and watch it walk. Or you can skip the simulation and tinker directly on a robot, hoping you can learn things from the real world– but that’s awfully slow.”
Google is learning this the hard way with its self-driving car project, where it’s having to run innumerable tests both on a physical road and in simulation.
The long hard road
What Boston Dynamics have been trying to do for years, and continues to try to do, is to perform the much-needed hard work of matching physically capable robots with the smart software that beats us at Go.
It is a continuous research-project they share with many of the world’s leading universities, and their combined effort will perhaps bear fruits in the next decade or so. Or probably longer.
And this is the reason Google has pulled out.
A project that won’t make money for ten years isn’t the kind of thing Alphabet Inc., Google’s parent company, would allow. At least not when it has a project like Google car running alongside.
Possible takers
According to Bloomberg, possible takers for Boston Dynamics include Toyota Research Institute, a division of Toyota Motor Corp., and Amazon.com Inc.
The fact that these Google competitors are potentially willing to invest the needed time and money into the next generation intelligent humanoid robots, says a lot about them, and Google.
For all the talk about Google being a research-led enterprise, the company is very focused on driving share-holder profits, letting other companies develop technology they can later acquire. It’s a business model that’s currently working very well.
The question is whether Google should shoulder some more of the economic burden of actually inventing the future – not just profiting from it gets ready for market.
Want to find out more about how robots are changing our world? Make sure you sign up to the Robots at Work event in March and join your peers for a lively debate on the matter.