Yesterday, I had the unique opportuinty to spend an hour with the Watson/Ai Video Project.
You all have probably heard about Watson, IBM's remarkable foray into the world of Artificial Intelligence. More than likely, your informaiotn is largely driven by Watson's victory in the TV game show Jeopardy. Mine was.
Watson, and AI are far more, of course.
Almost every field is going to be effected by AI, from law to accounting to medicine to engineering.... and, I suppose, just about everyting else.
I was invited to Watson's NY HQ on Astor Place, and got a chance to take a look at some of the rather interesting things AI is able to do with video now... and had a conversation about where it all might conceivably go.
At the moment, AI's best focus on video is the ability to recognize information in videos - locations, story lines, characters, dates, face recognition and so on. In a world in which people are uploading 500 hours of video to YouTube alone every minue, the world is awash in video, with almost no way to categorize it or even find it, except screening it in real time.
That means that just on YouTube alone, we are all uploading 720,000 hours of video every day.
That means it would take you 246 years to screen one day's video.
Clearly, something has to be done.
The more interesting discussion (not that this one was not interesting), was the idea of creating a software that could create, if not finished video products, then at least rough cuts.
As you know from watching the shooting lessons on TheVJ, we have been pretty good at parsing the grammar of shooting down into the 5 shots. These make up about 90% of any video to start with.
Then, if you add to that the kind of structural grammar of storytelling (much of it is indeed formulaic - just look at House or Houes Hunters), it is indeed possible, (though we are not there yet by a long shot) to create algorythms that could create general approximations of video cuts.
But what do I know.
It is sobering to talk to such smart people. It reminds me of when the Flintstones met the Jetsons. Made me feel like Fred.