Does anyone else think this company missed out on a terrific opportunity to name their AI “Deep Thought”? Anyone? 42? Is this thing on [tap, tap, tap]?
Beating out Facebook with a $500 Million dollar bid, Google is now the proud owner of DeepMind, a company focused on creating algorithms which will help computers learn in a way which resembles human experience. Artificial intelligence. Now, most people are linking this to Google’s recent acquisition of Boston Dynamics as if we’re all on the cusp of a Skynet inspired apocalypse. While I’m sure learning algorithms would be a tremendous boon to the robotics community after watching excerpts from the DARPA Robotics Challenge, it was very clear that we’re a LONG way off from Skynet or Asimov’s robots, there are a number of other projects at Google HQ which could also benefit from some good old fashioned learnin’
Improving the responses of automated, self-driving cars for instance. Making those systems more flexible and adaptable to changing road and traffic conditions.
Google is first and foremost an information broker.
Search will continue to be an ever increasing issue as we dump more and more info into this giant bucket we call the internet. Google is now facing incredible competition from companies like IBM with their Watson project. A computer system which isn’t self-aware, but is able to adapt, add information, and make sense of casual instructions and commands to deliver meaningful results. If Google doesn’t start working on a similar program, they’ll find themselves disasterously behind. Given a choice between Siri, Google Now, and Watson, I know I’d prefer to send my search query to the computer who spanked the crap out of the top two most winning Jeopardy contestants in that show’s history.
We instantly assume that working on aspects of Artificial Intelligence means that we’ll flip a switch someday and all our machines will have distinct personalities. What will most likely happen first is more likely to be painfully dull for the general populace. Purpose built computing systems which are subtly, but noticeably better at their individual jobs. Think of a WHOLE animal brain, so many different pieces to regulate and control various aspects of an organism’s life. We’re a LONG way off from poorly replicating the human frontal lobe. We’re a LONG way off from building a system which could rival a house cat’s ability to problem solve and learn tricks.
Let’s not forget that Google’ main competition for acquiring DeepMind was Facebook, a company that failed to put out a branded phone. If there were any more reassuring fact that we wont see this AI in some near-future skeletal doomsday robot warrior, it’s the fact that Facebook is most likely looking for better systems to sift through user data, not to operate machinery.
So when Google flips the switch and we all get slightly better turn by turn recommendations, when predictive search results and ads are a little better targeted at our needs, we can all shrug a sigh of relief. However, if I’m wrong, allow me to be the first to welcome our future robot overlords. I’d make a fantastic liaison officer in your new world paradigm where humans are kept like cattle for some unexplainable but dramatic reason.