Tuesday, January 27, 2015

Big Data: Another Big Lie Screwing Up Innovation

A lot of people are familiar with the current state of climatology, fettered as it is by this insistence that we apply an inordinate amount of concern to atmospheric carbon dioxide and simultaneously insisting it has something to do with global temperatures, despite the track record on these sorts of predictions being abysmal. This is a subset of a larger scam- big computing- in which it is pretended that large, multivariate models of various systems can be used to do research, predict, and contribute in some sort of global way to our understanding of things.

This is not to say computers are never helpful, mind you, just that they are being applied most disingenuously in a great many fields, and as such we have various hypotheses, many of which have been disproven, masquerading as theories. We also have the acolytes of the techno-theocracy hailing the arrival of Big Data with evangelical zeal.

The result of this zeal is that nearly everything designed to track anything is being designed with some sort of transmitter because it is assumed we shall want to track everything and have it put into a spreadsheet so that we, or perhaps some experts in the field, can tease some sort of pattern out of all the data, and then we can materially improve our lives.

Lets look at a particular case in point- smart watches. Most of what a watch is good for, like telling time, and various biometrics, like heart rate variability, require a certain amount of power, but not as much as the total Android operating system, transmitters, and screen.

Maybe you don't need such a smart and power hungry watch.

Maybe we could use a low powered screen, like an e-ink screen. Perhaps we could work toward an extremely slim, possibly even stylish, imitation of mechanical watches which would be capable of a few extra screens of feedback we could use in real time.

Yes, I know I didn't do a good job crushing your dreams of a wonderful new utopia coming from having everything uploaded and analyzed, but you already know the here-and-now truth if you've tried tracking anything- most of the gadgets to track stuff ends up being awkward and annoying. Until it gets easy, there won't be Big Data anyway. Few people are going to upload enough data, because the gadgets suck. The gadgets can also be changing the very things we are trying to track- if your sleep tracker is wireless, and having something wireless that close to your head turns out to mess with your sleep, for instance.

But if we focused first on immediate feedback- which has a pretty solid track record in terms of helping us improve ourselves- we could have reasonably passive, low powered systems, and small, local data.

Update: I forgot to mention one of the most obvious failures of Big Data that everyone online sees all the time- advertising. Despite desperately tracking your every move, most of these advertising algorithms end up showing you whatever you've already bought. Or perhaps you did some online searching of something and it clearly can't tell the difference between something you would be interested in and something you just looked up because you wanted to know what it was.

No comments: