Two recent events, which couldn't be less alike, have made me wonder whether we're possibly in the midst of a game-changing revolution (or two) the likes of which usually only comes along once in a lifetime, if that.
The first is the wave of uprisings that appear to be sweeping through Africa and the Middle East, seemingly feeding off one another in a way that's truly mind-blowing if you stop and consider their scope. Social media, a term that presumably wasn't even coined a decade ago, is now being partially credited with toppling dictatorships. Let me just repeat that: software intended to facilitate 'keeping in touch' or describing what your cat had for lunch is now contributing to regime changes! What armies, black op forces and political will couldn't accomplish, Twitter and Facebook are helping to make a reality! This is beyond unbelievable, really.
On the other side of the human experience, there's IBM's supercomputer Watson appearing on Jeopardy! and defeating the two greatest human contestants that show's ever seen. I'm not sure how many of the millions of people who watched that bit of entertainment have any idea just how significant that outcome really was. As the Nova episode that provided a lot of background on Watson made perfectly clear, this was a stunning accomplishment for the world of artificial intelligence! Because of the human-friendly, computer-unfriendly format of the game show's questions, success there required Watson to master what's called 'machine learning'. Essentially, it couldn't be pre-programmed with enough static information to handle the challenges it would face because the 'chatty' nature of the questions is designed for humans, using puns, plays on words, and double meanings to simultaneously conceal and reveal their meaning.
This is something that wetware like anyone reading this blog learned to do rather intuitively early in your normal mental development, but which is extremely difficult to quantify in the form of software. To get Watson over that hurdle, IBM introduced machine learning into the AI's bag of tricks: Watson would examine all correct answers ('his' own and others') and look for patterns in them to help 'him' learn on the fly. In my mind, that last statement is every bit as potentially monumental as the earlier one about social media. Getting a computer to learn on its own is a thought both wondrous and terrifying. It's tempting to downplay it and say, "It's still only doing what it was programmed to do." But I think that's shortsighted. We have no way of knowing where the vaulting of that particular obstacle may lead artificial intelligence in the next ten years, say. I'm not necessarily predicting SkyNet - though I'm not ruling it out, either! - but that's only because I don't think we can forecast the result yet. I just think it's going to be huge.
In the same way that some people think in terms of "pre-9/11" and "post-9/11" when they consider events from the past couple of decades, we may eventually talk about "before Watson on Jeopardy!" and after it. This could be that important.
Or, of course, I could be wrong on both fronts. But it certainly doesn't feel that way right now.
Subscribe to:
Post Comments (Atom)
1 comment:
Good post. I think you are right on the importance of Machine Learning...this does seem like the break-thru that the sci-fi writers have taken for granted.
As far as Social Media, now that it has everyone's attn, I think it is only a matter of time before Corporate USA figures out a way to control it and sell that to Governments...just like the free press was quickly controlled - I give free Social Media maybe 2 yrs before it can be controlled.
Post a Comment