NYTimes | Artificial intelligence is booming. But why now?
Move over, social media and mobility: Silicon Valley has a next big thing, John Markoff writes , and it’s A.I. and robots. It is useful to think of them as part of the same thing, since many robots are autonomous machines programmed for decision making based on A.I.
The movement can be thought of as a spread of computing intelligence everywhere, on wheels and wings, in your pocket and all through your house. That’s a big enough idea to fund scores of companies, and quite possibly set up the next Silicon Valley boom. And bubble.
Yet it’s worth asking how much of this is reality and how much is wishful thinking. Why is A.I. growing the way it didn’t over the last several decades, despite promises that it would?
The answer to that lies in the precursors to this A.I. moment, which more than anything has to do with the Google-led search boom 10 years ago.
In 2006, Google and Yahoo released new methods of analyzing the quirky real-world data they were picking up from doing search. Data from browsers can be thought of as a proxy for human behavior, as people wander the web. It’s typically called “unstructured” data, as opposed to the more regular information of things like banking and airline schedules that filled most of the world’s databases.
That new way of seeing the real world helped make search profitable and also enabled companies like Facebook to look into even stranger social behavior. The success also gave these companies plenty of money to plow into the problem.
To money, and the first ever caches of natural behavior in digital form, add cheaper computing. In 2006, Amazon also introduced its cloud-computing business. Over the last decade, retail cloud computing has become an inexpensive way for lots of people to work on data analysis and pattern finding, the heart of A.I.
Only one more thing was needed, and in 2007 Apple came out with the iPhone. Let that stand for browser-type natural data collection moving off desktops and blowing through the natural world. Along with other cheap sensors tied to the cloud, it has given us huge amounts of data about all sorts of things.
That created many more places where computers could do what they’ve always done, which is to seek efficiency. That has created a cycle:more observation, more machine learning of patterns, more value capture funding more observation.
It’s enough to make you believe in this boom.
What could make you believe it’s also a bubble? Start with the name, artificial intelligence. So far there is zero evidence we will be about to build machines that possess intelligence or will really think on their own. If big money goes into that stuff, run for the hills.
0 comments:
Post a Comment