Death of theories?
From Wired Mag,
pointed via Duckwall...
"All models are wrong, but some are useful."
So proclaimed statistician George Box 30 years ago, and he was right. But what choice did we have? Only models, from cosmological equations to theories of human behavior, seemed to be able to consistently, if imperfectly, explain the world around us. Until now. Today companies like Google, which have grown up in an era of massively abundant data, don't have to settle for wrong models. Indeed, they don't have to settle for models at all.
At the petabyte scale, information is not a matter of simple three- and four-dimensional taxonomy and order but of dimensionally agnostic statistics. It calls for an entirely different approach, one that requires us to lose the tether of data as something that can be visualized in its totality. It forces us to view data mathematically first and establish a context for it later. For instance, Google conquered the advertising world with nothing more than applied mathematics. It didn't pretend to know anything about the culture and conventions of advertising — it just assumed that better data, with better analytical tools, would win the day. And Google was right.
Google's founding philosophy is that we don't know why this page is better than that one: If the statistics of incoming links say it is, that's good enough. No semantic or causal analysis is required.
= = = = =
Point of the article: we don't need hypotheses any more. We don't need to operate on the basis of judgment and ever-improving conclusions. We can just boil down the data to a conclusion if we have enough data and a good stove.
Google's method certainly works for advertising, where your goal is simply to reach the maximum number of eyeballs. This is basically the way ants operate. Send out millions of feelers, follow the path where more of the returning feelers bear food in their jaws. Google sends out ads to a variety of sites, counts the ads that return with clicks in their jaws, follows the smell toward sites with similar wording. Because placing each ad is nearly costless, the ant-like process is harmless. Nobody dies when an ad goes unread.
= = = = =
I'm torn between two responses.
On one side, the prevailing sin of this era is our tendency to follow fantastically stupid and genocidal theories like Global Warming to their inevitable conclusions without even looking
at the damned data. And the effects of such stupidity are far from costless. Millions of actual human beings are dying from the effects of Green theories: prohibition of DDT, breaching dams, fear of genetically-modified crops, switching crops from food to ethanol, keeping malarial wetlands in place, doubling the cost of energy, etc.
If blind mechanical analysis of the actual data could kill bad theories and eliminate reliance on purely imaginary models and simulations, it would be a godsend.
On the other side, total reliance on numerical analysis kills the most important basis of Western civilization: each human being is tremendously valuable and deserving of dignity, unless and until a human proves by his actions that he is unstoppably destructive.
Modern totalitarians from Hitler to Mao have used numbers and science quite competently to destroy the basis of civilization, to break the undefinable bonds of family and neighborhood.
Numbers don't care how they're used. If the goal is to maximize a politician's power, blind numerical democracy will get you there. Place enough news stories and polls in the media to drive the voters toward the correct choice, then wait for the ballots to return with power in their jaws. Karl Rove and Dick Morris have perfected the method.
However! Good empirical thinking requires more than theories, more than raw data. It requires judgment and discrimination, and it requires freedom to use and discuss your judgment and discrimination. You have to ask yourself and ask others whether a proposed action will advance the cause of civilization, and you must be able to state the goals and qualities of civilization clearly and openly. Our Communist masters have made this method suspicious in some areas of life and absolutely illegal in other areas.
So we are pinned between the ant-like juggernaut of lethal a-priori theories that cannot be questioned or invalidated by mere numbers, and the ant-like Google tricks of the electoral engineers. The middle course, the human course of examined judgment is gone, probably irretrievable.