What does the maturing of IT plus the rise of the so-called adaptive enterprise imply for IT people?
Aim as high in the value chain as you can. If IT is what you love, look at the world of autonomous agents, of pattern recognition, of globally distributed decision systems, where the next killer apps can come from. The economic value of the day-to-day IT jobs is going to decline and perhaps go overseas.
You cite companies such as John Deere, IBM and General Electric as using genetic algorithms (GA), in which rules in software evolve spontaneously, breeding better and better solutions. How important is that concept?
It's the leading end of a wedge. GAs are one of a number of techniques under the category of nondeterministic programming, where the objective is not to create code that never has a bug and gives you the same answer every time, but to solve problems that, in some cases, may be impossible for us to understand because they are nonlinear and nondeterministic.
In some cases, if you run it twice, you won't get the same answer, but you might get equally good answers. The application of GAs is a pretty scarce skill.
What about agent-based simulations, another non-deterministic computing technique that your book says is used by Southwest Airlines, the US Marines, Walt Disney and a few other organizations?
It's so far used by a handful of people in very customized ways. It will have a profound impact on the social sciences. It's the right way to model systems of people, of consumer behaviour, for example. The technique is pioneering today, but it's good enough that it's worth investigating, because it's a small expenditure for something that may yield new insights.
Will we see these techniques appear in commercial supply chain packages?
Yes, I know that (i2 Technologies and SAP) have been working on some of these ideas. One of the major competitive battle fronts will be making these packages more adaptive to real-time operations conditions.
In supply chains, things take a number of days to move from one place to another, but during that time, it may be that the highest and best use of something has changed, and so its destination should change. So, how do you build into your management of resources this real-time, sense-and-respond capability? It's real-time supply chain optimization, as opposed to optimization within a previously fixed set of rules.
You mention real-time pricing as one specific application of these non-deterministic techniques.
Yes, the ability to say, for example, "This car ought to be repriced and this model ought to be deeply discounted based on what's on the lot" is an example of a capability that's going to go into the big packages. In order to support that kind of thing, IT people will have to look at information as a data stream about the world, not just as transactions.
You say that much of this data stream will come from sensors, which will become cheap and ubiquitous as a result of the revolution in molecular sciences.
That means there's going to be a lot of data available that didn't used to be. So an IT group that's really trying to help will go to its management and say, "What information could you use to create value?" and, "We'll think about how to get that information in real time." Customer behaviour is a key one.
What challenges does this pose for IT?
If you take (the principles of) sense and respond, and learn and adapt seriously, you're going to have to use technologies that handle vast data streams. For many organizations, that will be a new challenge.
You say companies and their IT shops should "operate at the edge of chaos". How can that be a good thing? If you believe competitive advantage lies in the ability to sense change in the environment and respond to it faster than anyone else, and thereby keep your opponents off balance even though you feel off balance because you are operating as fast as you can, then IT can create competitive advantage by being able to go through the orient-observe-decide-act cycle faster.
Might such rapid change jeopardize IT's core mission?
The IT group can't view its mission as tending to the existing IT infrastructure any more, keeping it at 99.99 percent available and all that. Outsource the commoditized IT stuff, and be in the business of having the fastest change capability in the industry. The stuff that stays in-house will be somewhat chaotic, and it should be.