The Few can now Watch the Many: 1984 vs 2014 and Exponential Growth in Surveillance Tech

My friends know lots of things about topics like Machine Learning and Big Data. They can detect phrases in speech, objects and faces in billions of pictures, or sort trillions of numbers in a few minutes.

And not unsurprisingly, these same people could do so much less just 10 years ago. Machine Learning and Big Data have been on such an amazing curve that these whole fields have been reinvented very recently. The technology is reaching a totally new scale and range of capabilities.

You think for a minute about that Orwell book,1984. It's suddenly a bestseller again.

Half of us, Spies?


What would it have taken to build the surveillance state in our 1984? Well, you would have had to hire half the population just to spy on the other half. Without technology to help, you would need massive quantities of human attention to do surveillance. It would be an immense cost and massive operation. That's one of several reasons it never happened.

But today, 30 years later, it's conceivable to filter all voice communication through very smart algorithms, to run face detection at airports, to log every license plate that drives by many thousands of locations. Easy stuff now.

If the NSA is paying $600 million to Amazon or $860 million to build a new datacenter, those numbers are pretty small, even tiny, compared to hiring just 10,000 people to do the same work manually.

Let me say that again. Thirty years ago, it would take 100 million people to keep track of what everyone was saying and doing. And this year, it's 10,000 times cheaper. (I even looked at the license-plate reading cameras and thought of buying one for my house. It's under $1000 for a complete system.)

So maybe, ten years from now, what if you could just buy today's NSA cloud computing infrastructure for maybe a million dollars?

Ultimately, it doesn't matter how much you cut the NSA budget. Technology is getting cheaper really fast. The algorithms are getting better, while the storage is getting lots cheaper.

And I've started to think that this exponential growth is the story that isn't getting written. We know from the tech boom that technology centralizes control. A few people can write code that makes billions of dollars in just a few years. That's unprecedented in history, kind of like this "new" surveillance issue.

Even knowing that, there are still two kinds of people in the world: the 0.1% who are watching this exponential growth in machine technologies and understand the implications of the scale and centralization, and everyone else, who thinks maybe the NSA/Snowden thing is just some questionable policy and nothing new in the world other than perhaps an abuse of power.

People are completely and utterly unprepared for this particular exponential growth curve. Surveillance is going to be millions of times cheaper than it was ever before.

But most of the world's population has no way to reason about this, or to understand it. There is no precedent for this in history. 

No comments:

Post a Comment