I bought this t-shirt from Threadless some time ago — it was a tongue-in-cheek reflection on all the fanciful things that we saw in movies and cartoons as a child growing up that hadn’t quite come to fruition yet. It seems, though, that the second item on the list — that of the robotic companion — may be pretty close at hand.
I’m not one to make predictions, typically (even if it is new year’s day) but a couple of things that have recently come across my radar have got me thinking that the age of the robotic companion is coming — probably in about 5 years time we’ll see the first commercial versions.
What are the developments? Apple’s Siri is probably the most mainstream. This is natural language recognition and response in a consumer-grade mobile phone. Yes, internet assisted (much of the processing is done by much more powerful computing hardware than the phone). Yes, it’s early days yet (anyone that’s used Siri will be well aware of it’s limitations).
But what’s so special about Siri? Haven’t we had voice recognition on phones for some time? Siri is more than just recognition of pre-defined commands — it incorporates natural language processing, which is a real leap forward in my view.
In 5 years’ time, talking to our phone will be as natural as touch gestures are today, I suspect. This will in part be aided by the continuation of Moore’s Law, as Mark Pesce recently reminded us:
By 2020, some of us will be walking around with a teraflop in our pocket, interpreting our speech, watching our gestures, and effortlessly handling sophisticated social transactions – invisibly, continuously and tirelessly.
According to Gizmodo, Boston Dynamics expects to deliver PETMAN to the US Military as early as this year (2012). Given that, I suspect that the video that has been released is probably of an older variation of the technology, so is probably even further advanced in the lab.
The third is the advancement of facial expressions in robotics. See, for example, Actroid-F:
While the latter two examples are still in the research and development phase, and are clearly going to be extremely expensive, as is often the case with these things, these early developments will no doubt trickle down into consumer-level products shortly after. I’m putting a stake in the ground and saying 5 years (though it might end up being 10).
But in either case, at least one of those “damned scientist” wishes is just around the corner…