Jaron Lanier, who wrote this well-received book Who Owns The Future (WOTF), is associated with Microsoft Research and has been active in connected things and big data for a very long time. He loves the world of distributed instruments and big data. Yet his talk is one of caution and concern, and perhaps a way out:
‘I have a profound concern that we’ve taken a wrong turn and that the business models we’re pursuing are unsustainable.’
Using an analogy from Quantum Physics, Maxwell’s demon, Lanier talks about how “there is no free lunch. Whenever there is the appearance of easy or free money, there are serious side effects. The first stock market flash-crash was already in 1987. If you had superior computation capabilities, you could almost create money out of thin air. But if for example healthcare companies start excluding their most costly clients, the model will break. If in a hot climate everybody uses air-conditioning, it will actually get hotter and hotter.
The same applies in the economy: the margin is being squeezed out of everything. It’s hard to make money out of devices, hardware. If you can get the data, that’s everything! But what is the business model?
The financial sector became addicted to big data early on. The scheme of long-term capital management. I believe the whole world will pay for that. Any slow changing variable can be forecast by statistics to some degree, but how much is uncertain. The Scheme doesn’t actually represent modeling of the world, just a limited view into the future. But that reality is hard to accept because at first it works! You first have profits out of nothing, and then it fails. This is one of the problems.
What changed was the financial crisis of five years ago (bundled mortgages). Now statistics started to incorporate individual data, scraping the net in order to sell people stupid mortgages in a Big Data sales effort. If you look at the market of personal data, the biggest customers are insurers and banks. Because it works. I believe it will fall of a cliff.
Problems with this model
If you’re going to be selling data, it’s very hard to be in control of the data. Google for example, imported same model as financial data and uses the personal data to place advertisements. But now ad-block blocks the ads and Google needs to pay ad-block to let ads through. Almost with no value, ad-block got in between Google and their business model.
There is a limit to how useful Big Data is: Big Data certainly is not a substitute for wisdom or understanding. If you’ve used the obvious examples of finding where to build a hospital or mall, you start running out of examples quickly. There is a limit on how much useful information is in data. So soon that space becomes very crowded. And if there are a few power players who apply ‘sorting’ like Maxwell’s demon, you’ll have a power-law outcome with just a few very successful startups, YouTube videos or universities. Yes, it becomes inexpensive for new entrants to enter, but the tail is very long and flat.
If Big Data is only used as a routing or sorting mechanism, it can only be used to sell ‘something else’. But at the same time, all that data is being used to replace the ‘something else’ – this process of replacement is true in every industry. The people doing well are closest to the Big Data computers, until the whole thing collapses. There are more and more examples of industries that are being hollowed out, which is unsustainable.
What to do about it?
I’m not sure, but an answer I’ve been exploring is an old model, discussed in the sixties by Ted Nelson (Harvard). In this model, you keep track of where bits originated in an online system. Then you pay people in micropayments based on what data is valuable. It creates a more balanced system that is sustainable. It could create a big data driven society with a high degrees of automation and still a high employment.
Would the micro-payment paradigm result only in negligible amounts? Probably not:
- Mathematics: if you look at distribution of outcome in hub-spoke network, the sources of information follow a power-low. But in graph network (like Facebook) there is wide number of sources and it follows a bell-curve. If explored well, it would create a middle class.
- It’s hard to calculate what data is worth, but you can look at the differential between what you would pay for something with or without sharing all your data. Or look at what third parties will pay for your data. In research I hypothesized that even for the most boring person the data is already worth hundreds or even thousands of dollars. If that number hits the poverty line, it may be new model. Today we have basically two models to solve inequality: redistributive systems or a laissez-faire market. This could be a third model, that is not political and much more comprehensive and it still is laissez faire, and perhaps much more valuable.
Will we be in time for IoT? It remains to be seen: Even basic functions such as the phone are still pretty glitchy. All companies have had data breaches. It is my major concern for IoT: do we even have the systems in place to make the ‘Things’ run well?