Before I begin some background information.
One of the consequences of the development of utility services in computing is the growth of ecosystem models such as innovate, leverage and commoditise or ILC (see Graph 5). These models appear to allow a company to become linear (possibly super linear) for both customer responsiveness and the innovation of new activities, with respect to ecosystem size. Combined with economies of scale, the effect of this is that those companies which exploit such models will become harder to compete with as their ecosystems grow. Contrary to Porter's three strategies, these companies can become simultaneously more efficient, more innovative and more customer focused with growth.
The trick to this is data, the timeliness of the data and the ability to respond. Use of APIs inherently provides more timely and more voluminous data than market research on product usage but it's the combination of API provision to an ecosystem (provides access to raw information on consumption), metrics on usage and diffusion (algorithms), utility infrastructure (raw compute power) and distributed big data systems (processing) which makes exploitation of this more possible.
Ecosystem models such as ILC where innovation is encouraged elsewhere through reducing cost of failure, the ecosystem is then leveraged to spot diffusion of innovation and then those diffusing activities are commoditised to component services in order to grow the ecosystem - a process of eating the ecosystem in order to grow it - have only become possible in the last decade.
The practice should itself diffuse to dominate all industries touched by technology and those thinking that a bigger company (in terms of size) can beat a smaller company with a larger ecosystem (in terms of size) should prepare themselves for a shock. It's not companies but ecosystems that fight in this future world.
This won't be the last time we see billion dollar market cap companies employing thirteen people, quite the opposite. We should expect this, because whilst we're on the tail end of one industrial age which has covered the internet, social media and cloud computing, we're also at the start of a new age which will be built on many of the components that the last produced.
These component will include utility infrastructure (i.e. compute power and storage), big data processing and increasingly a widespread sensor network from smart phones to the software components that we interact with such as SIRI and EVI.
These components will lead to agent software, an environment where my abilities are augmented by the network around me. From dynamically determining my schedule, to working out what to buy my sister for her birthday to giving me a heads up on my bosses latest meeting. For some background on this see "Any Given Tuesday"
However, the ability of my agent network will depend upon access to information, algorithms, raw compute power and processing capabilities. Even if access to information and processing capabilities were open (as in open data and open source respectively), my ability to exploit this will depend upon the algorithms and volume of computer power I can afford.
In other words could my ability to exploit agents to my personal advantage relative to over others will depend upon how much I can afford?
So not only is education a resource to be bought, so too will the ability to create advantage through agents and to hence more effectively exploit the precious time we all have. Counter to the levelling that open source has brought, this change could lead to an increasingly stratified world between the have more and the have less.
As Tim O'Reilly has pointed out, we are part of this machine and we're not yet sure what we are building. In the excitement of all the possibilities of big data, there exists a dystopian nightmare of reduced social mobility.
Could my failure to afford the right sort of education and the right sort of agent condemn my son's future to answering endless Turk requests on "which would make a better birthday present?" in order to serve the software agents of others who spend their time on higher matters because their parents can afford to buy them into this strata? Does it matter? Will ability matter?
Certainly the movement of social mobility in the US / UK gives me pause for thought but who is looking into the social impacts of big data?