Google has been snatching up robotics companies left and right, but what is Google going to do with a bunch of robots? Here are some answers.
Google has been on a buying spree lately and robotics companies have been at the top of its shopping list. They have purchased eight robotics companies in a matter of months and haven't made a peep about why they're so interested in robotics.
Here is the list of companies Google has acquired:
- Schaft Inc.
- Industrial Perception, Inc
- Redwood Robotics
- Meka Robotics
- Bot & Dolly
- Boston Dynamics
- DeepMind Technologies
This list covers smart AI, disaster relief robots, and a four-legged robot that can run almost 30 mph. Google has diversified its product offerings over the last decade, but what is it going to do with a bunch of robots.
The short answer is simple, robotics are becoming increasingly data driven, operating based on sensors and algorithms. Google is a company that has made data and algorithms its business, so it's a logical step for a company that values innovation as highly as Google does. According to Anthony Mullen, a senior analyst at Forrester Research, it's getting them closer to consumers.
"Robots, like smartphones, are a platform for products and services," Mullen said. "Both require data and intelligence to operate well and Google is very good at data and algorithms. To ensure that they aren't disintermediated in the 'last mile' to the consumer (or employee) means getting involved in the physical world with hardware.
"Android OS and smartphones were about staying close to the consumer and capturing/leveraging data first hand. The robotics play is the same—control and influence in the last mile where data is gleaned from the physical world and activities are informed by intelligence in the cloud. It's a good fit and after Kurzweil joining them to head up AI. I'm not surprised it's happened so quickly over 2013. "
The long answer is a little more complicated, and one has to wonder if the consumer market is even ready for robots. Hod Lipson, associate professor and Director of Creative Machines Lab at Cornell University, said, "For the last couple of decades, most of the robotics we've seen have been in structured environments." The structured environments that Lipson is referring to are places like factories and warehouses where they work alongside humans to accomplish tasks.
Typically, robots in these "structured" environments are placed behind locked gates or operate behind a guard. The guards are not intended to protect the humans from the robots, but to protect robots from changing environments. Changing environments pose a threat to robots because they lack the ability to engage in the crux of biology—adaptation. Up to this point, robots have struggled with the ability to improvise and adapt.
Outside of military robotics and vacuum cleaner robots like the Roomba, Lipson said there was a lot of skepticism about robots existing outside of structured environments. But, he said the technology is getting to the point where it is moving outside of factories and into our homes. Tom Austin, a Gartner fellow, echoed this idea, citing advances in AI as a contributing factor.
"We believe we’ve finally turned the corner because of new algorithms (e.g., see deep neural nets), new hardware (massive improvements in parallelism, throughput and interconnect, all at continuously plummeting prices) and great seas of data to feed to the hardware and algorithms. The second half of this decade will see many of the early dreams of AI theorists finally coming true," Austin said.
This could mean that Google is making a land grab and storing up technology and talent for the inevitable breakthrough of serious robotics into the consumer market. More than likely though, what it means for now is that Google is just building a portfolio of companies to help drive innovation as one of their "moonshots."
Lipson believes that these robotics purchases are not a unique collection that will enable something that hasn't been done before, and they might not even be slated for specific product angle either. Austin sides with the moonshot theory, but said they're also making a long-term investment.
"Why take on these projects? Because the engineering challenge and the long term benefits for society can be enough to motivate them to make the investment. (I also believe they’ll try to make money out of these types of projects where they can but let’s not get carried away with that logic. They didn’t create their semi-autonomous car demonstrations to make money there.) They did it to fulfill the spirit of Google Moonshots," Austin said.
Just because Google categorizes something as a moonshot doesn't mean that a consumer application is out of the question. Schaft Inc. started out building disaster relief robots after the 2011 Fukushima nuclear disaster, and it would be impressive if they could be rolled out based on the information coming from Google Crisis Response.
Some have proposed that the robotics purchases are going to be funneled into Google's driverless car program. According to Lipson, the driverless car is one of Google's primary drives toward robotics, but this purchase probably won't help. Whether it comes from Google or not, Lipson said a driverless car will likely be the first fully-autonomous robot that people will interact with; but certain anomalies (rogue plastic bags covering sensors, reflections in puddles interfering with cameras, etc.) are holding back its development. These newly-acquired companies, while wildly innovative, don't seem to be building the technology to address these issues.
When asked what Google's goal was for the robotics purchases, a Google spokesperson declined to comment.
But, Lipson said one of the main things to remember is: "Robotics is different from software because it turns data into physical action, automatically."
Regardless of what Google is going to do with these companies, Lipson said that their recent purchases provide validation for robotics in the consumer space. He is hopeful that Google will help drive innovation through its own people and validations in the field to show that robotics is a safe investment.
- Google's next stage of internet pwnership: Fiber targeting 34 cities means they are a real ISP now
- Google buys 8 robotics companies in 6 months: Why? (CBS News)
- Back from the dead: Google breathes new life into Project Ara with a developer kit and conference
- The three factors keeping Google from a full-scale enterprise assault