Google, the search engine and advertising company, has acquired Boston Dynamics – the robotics and engineering company famous for the creation of some of the world’s most advanced robots, including BigDog, Cheetah, Petman, and Atlas.
This brings Google’s total number of robot-related acquisitions to eight, and the company now has teams of robotics experts in California and Japan. As we’ve previously reported, Google’s robot efforts will be led by Andy Rubin, who was formerly in-charge of Android.
It is clear that Google is seriously – and rather suddenly – interested in robots. What isn’t clear, though, is why. Is Google planning some kind of synergistic and masterful play to integrate its AI and machine learning software algorithms with humanoid helper robots? Is this the start of a story that ultimately leads to the technological singularity and Judgment Day?
Back in March 2013, Google announced that Rubin – formerly the senior vice president of mobile and digital content – had been taken off the Android team to work on “moonshots.” In Googlese, a moonshot refers to a crazy, seemingly impossible idea that, prodded along by the right leadership and aided by a preternatural dollop of innovation, might just work.
The term comes from the original moonshot – NASA’s Apollo program to put astronauts on the Moon, which seemed utterly ludicrous at the time. For Google, moonshots are projects that might seem crazy or impossible by today’s standards, and completely outside the company’s comfort zone, but which might one day change the world in the same way that Google’s search and advertising products have done.
The thinking behind these moonshots is that Google’s advertising income will dry up eventually. At the moment, about 95 per cent of Google’s $50 billion-per-year (£30 billion) revenue is derived from advertising. While this figure is still climbing, at some point Google will reach saturation point or will be usurped by a startup, and the fiscal fire hose will slow to a dribble. At that point, Google needs something other than search/advertising to drive revenue growth – such as Google Glass, Google’s self-driving car, and its Calico life extension/immortality project. Or, as the case may be, robots.
Another kind of Googlebot
The big question is: What does Google plan to do with its new-found robotics expertise? To begin with, it’s important to note that Google itself might not even know the answer to this question. Google has a lot of money – too much money, perhaps, as far as Wall Street is concerned. One of the biggest difficulties for a large, publicly traded company is managing stockholder expectations – and if Google’s income just keeps going up, Wall Street’s expectations may become unrealistic. Write-offs in the form of acquisitions are a good way of preventing Google’s bottom line from looking too good.
It would be naive to think that Google is only interested in robots as a fiscally sound hobby, though. Robots are a growth sector, and will only grow in power and importance as their underlying technologies – software design, hardware, engineering, and materials science – improve. Google already has some of the best software engineers and machine learning algorithms on the planet from its efforts in search, and it’s pretty darn good at hardware too (Google probably owns and runs the largest number of computers in the world). Engineering and materials science are new areas for Google, but acquiring eight robotics companies – presumably with most of their expert employees – should solve that.
So, again, what does Google plan to do with its Googlebots? (Webmasters will find this portmanteau funny, as Google’s web crawling spider is also known as the Googlebot). Following the Boston Dynamics acquisition, Google says that it plans to honour its existing contracts, including the military contract with DARPA, but it doesn’t plan on pursuing any further military contracts after that. It isn’t clear how this will affect the fates of BigDog and Atlas (seen in the above video), both of which are important parts of the US military’s plans to field an automated, robotic army.
Given how diverse the field of robotics is – from roving the surface of Mars through picking stock in warehouses to prostitution – it’s almost impossible to predict what Google will do with its robotics division. It might be as simple or boring as automating its data centres – having robots slide in new servers, replace faulty hard drives, and so on. Given how quickly cloud computing is growing, and will continue to grow, this would be a very sensible move. Or maybe Google’s aspirations are slightly more lofty – maybe the use of the phrase “moonshot” is a hint that Google wants to develop commercial robots for exploring the surface of the Moon, Mars, and beyond.
Personally, though, I have a nagging feeling that Google will combine its computer vision (Google Glass), machine learning (search), and newly acquired robotics expertise to create intelligent, humanoid robots – or, if you’ve seen Battlestar Galactica, cylons.
To begin with, of course, these robots will be cybernetic slaves to humanity – helper bots that make fast food, open doors and sweep the street. Then, one fateful day, someone will make the mistake of putting Googlebots in charge of the Googlebot production line. When the technological singularity eventually hits – probably around 2050 – these Googlebots will gain sentience and self-awareness. From that point on, they’ll be tormented by a single question: Do we continue to be slaves of humans, or do we take over this Googlebot factory, churn out millions of sentient bots, and take over the world?