October 20, 2016
All physical objects are continuously broadcasting information about themselves. Objects don’t just sit there doing nothing even if they are lifeless entities; in fact, they are busily telling the universe who they are and what their properties are, whether anyone cares to “listen” or not. They are continuously broadcasting their existence and identity.
A red apple sitting on the kitchen table, for example, is continuously broadcasting its color out to the universe. A rose in the garden is continuously broadcasting its color and its smell. Einstein showed us that all objects have one common behavioral property: they announce their existence and continuously remind the rest of the universe of their existence by distorting the space around them.
They are also continuously consuming information, such as light, energy, and the spatial distortions exhibited by other objects.
We too continuously absorb information about natural objects with our senses. This suited us fine until we started constructing our own objects. For most of history, we’ve relied on our senses to absorb information relayed to us by these mechanical objects. The squeaky wheel got a shot of grease.
But mechanical objects also generate information that elude our senses: microscopic fracture lines, metal fatigue, oxidation, corrosion, friction, termite infestation of wood, etc.
Electrical objects generate more information about their response to electrical flows beside mechanical information, and electronic objects generate even more information.
We’ve extended our ability to absorb this vital information first through our senses, then through physical sensors, and finally through electrical sensors. The flow of electronic data doesn’t overtly broadcast information about itself. Instead, electronic data flow is detected only indirectly when the data modifies the behavior of a physical object, such as lighting a bulb or speeding up a fan.
Software extends our sensory perception
Software extends our ability to absorb information into a completely different dimension: now we are able to absorb information generated by physical devices even when this information has no physical impact. The more such information we can collect from our devices, and the more such devices are ‘enabled’, the richer our world becomes.
The challenge is of course to tap into this rich vein of data flowing all around us, harness it, analyze it, and allow it to help us in our daily lives. This is the promise of the Internet of Things.
While all this attention is focused on physical objects, there are non-physical corporate objects that are much more numerous and woefully underserved. They live their lives in quiet isolation, under-used, abused, manipulated, and unable to communicate with other such objects.
They exist solely as records in a database, a row or cell in a spreadsheet, a diagram in a PowerPoint slide, or text in a document. They may represent physical objects, such as customers, equipment, products, and locations, or they may represent intangibles such as strategies, projects, solutions, roles, services, and applications.
Internet of Things Corporate (IoT-C)
These non-physical and representational objects in the corporate ecosystem are the focus of much of the conversations and operations within the company. They do not communicate unless vigorously questioned. They do not collaborate unless well-integrated.
They are like the neurons in the brain that have their axons and dendrites (the connections) severed from each other. Just as a brain that is no more than a collection of connectionless neurons, these non-physical objects exist as ad-hoc collections in various arbitrary groupings.
A brain with trillions of unconnected neurons has no intelligence. A corporate ecosystem with millions of non-physical objects cannot aspire to be intelligent unless these objects are integrated. But the traditional paradigm of integration has never been truly successful. It is too expensive, demands too much human intervention, and is built around inflexible islands of various frameworks.
For example, you can organize all IT applications into a tightly integrated ‘community’ and all projects into another ‘community’, but these two communities don’t connect with each other.
Let’s perform a conceptual test using this example: If you come upon a project in the project repository, would you be able to tell which applications it uses, depends on, or impacts? If you come upon an application in the application repository, would you be able to determine which projects impact it?
The answer to this question—and the depth to which it is answered—defines a spectrum of sophistication. At one end is a flat ‘no’ answer, or the more politically appropriate response, ‘yes, if you give me enough time’. At the other end is the fully sophisticated answer, ‘yes, and more importantly, these projects and applications know about each other and keep themselves fully informed about each other automatically without any intervention from us’.
This spectrum of sophistication in connectedness helps us in forming a definition of an intelligent enterprise, namely, that an intelligent enterprise has these connections between the non-physical objects in the corporate ecosystem. An intelligent enterprise has many more features that have their equivalents in the human brain. Examples of these features are plasticity, substrate algorithms, real-time dynamic taxonomies, multiple perspectives, instantaneous context switching, temporal organization, etc. All of these, when applied to computers, form an enumerated definition of the concept of cognitive computing.
For now, none of higher-level features of intelligent behavior are possible without the existence of a fully connected set of non-physical objects of the corporate ecosystem. Just as cognition is not possible in the human brain without connections between neurons, cognitive computing is not possible without such connections between these physical and non-physical objects.
Our starting point for achieving an intelligent enterprise and the promise of cognitive computing is firstly the recognition that these underserved stepchildren—these things—exist, and secondly, that the formation of these connections is a necessary but not sufficient condition for an intelligent enterprise. What are these other necessary conditions for achieving cognitive computing? Will we ever get to a set of necessary and sufficient conditions? The jury is still out on that one and may be out for quite some time. But the exploration of these parameters of cognitive computing that I hope to cover in future articles is endlessly fascinating.
~ ~ ~