Wednesday, November 19, 2014

“It Can’t be Done”

Some of the most-dangerous statements an experienced and knowledgeable executive ever can make is that something “cannot be done,” or that a new way of doing something is underpowered, under-featured and essentially a non-serious approach to solving a problem.

If confronted with a requirement to support huge amounts of bandwidth, hundreds of times to perhaps 1,000 times greater than anything yet seen, it might seem obvious that only fixed networks will be able to handle the load.

That is why Marcus Weldon, Bell Labs President and Alcatel-Lucent CTO believes sophisticated core and fixed networks are essential, and that explorations of Internet access networks using unmanned aerial vehicles or balloons are unsophisticated approaches little better than “toys,” compared to the best of today’s telecom networks.

The phrase "toy networks" as applied to new Internet access platforms such as balloons or unmanned aerial vehicles reflects a perhaps-understandable reaction to new networks that lack the sophistication of the existing and future networks envisioned by the telecom industry.

But it is profoundly dangerous to underestimate the threat posed by such underpowered or feature-deficient new approaches. You might recall that the same sort of sentiment was uttered about voice over Internet Protocol.

Disruptive innovation, a term coined by Harvard Business School Professor Clayton Christensen, describes a process by which a product or service takes root initially in simple applications at the bottom of a market and then relentlessly moves up market, eventually displacing established competitors.

Such innovations might reasonably be derided by existing suppliers as “not very good” products with limited feature sets, unstable quality and some restrictions on ease of use. Skype initially could only be used by people communicating using personal computers, for example.

Microwave Communications Corp. (MCI) originally competed with AT&T for long-distance voice calls using a microwave network that likewise was deemed less reliable than AT&T’s own network.

Wi-Fi hotspots originally were hard to find, sometimes difficult to log on to, and obviously did not have the ubiquity of mobile Internet access or the speed of an at-home Internet access service.

Netflix originally required mailing of DVDs to view content (it was not “on demand”), and could not be viewed on demand, on a variety of devices.

What happens, over time, is that disruptive attacks gradually “move up the stack” in terms of features and quality of service, eventually competing head to head with the incumbents.

If you live long enough, you might see many examples of such derision.

I can remember being at a meeting at the headquarters of the National Cable Television Association, in the earlier days of high definition television discussions, where it was proposed that a full HDTV signal could be squeezed from about 45 Mbps of raw bandwidth to the 6-MHz channelization used by the North American television industry.

The room essentially exploded, as the attendees, mostly vice presidents of engineering from the largest cable TV and broadcast firms, disagreed with the sheer physics of the proposal. Later, the executive who suggested HDTV in 6 MHz was indeed possible talked with his firm’s engineering vice president, about the the science, to reaffirm that such a thing actually could be done. “Are you sure about this?” was the question, given the magnitude of opposition.

To make a longer story short, it did prove feasible to compress a full HDTV signal into just 6 MHz of bandwidth, making for a much-easier financial transition to full HDTV broadcasting, as well as an ability for cable TV operators to support the new format.

Similarly, when the U.S. cable TV industry began to ask for analog optical transmission systems capable of carrying 20 channels of standard definition video without complicated channel-by-channel coding and decoding, a distinguished engineer from Bell Laboratories privately assured me that such a thing was in fact not possible, and that people who claimed it was possible were simply wrong.

To make a longer story short, it did indeed prove possible to take a full complement of analog video signals (40 channels, as it turned out), convert the full set of broadband signals to analog optical format, and deliver them over distances useful for cable TV purposes.

On another occasion, the vice president of one of the world’s biggest suppliers of equipment said privately that “digital subscriber line does not work” as a platform for high speed Internet access, even at relatively low speeds. Ultimately, that also proved incorrect. Over time, DSL performance was not only proven to be commercially viable, but also delivered much-faster speeds, over longer distances, as experience was gained.

The point is that when a smart, experienced, thoroughly-knowledgeable executive says that something “cannot be done,” one has to translate. What the statement means is only that, at a given point in time, before the application of effort and ingenuity, a given entity has not been able to do something.

That does not actually mean something literally “cannot be done.” Quite often, formerly impossible things actually are made possible, after dedicated investigation and development.

That sort of thing happens often enough that statements deriding novel approaches to solving problems should not be lightly dismissed. New platforms and approaches often do appear to be “toys” at first. But that is not where developments remain for all time.

Executives generally truly believe disruptive new platforms and approaches are unsatisfactory substitutes for higher-performance solutions. That often is quite true, at first. But substitute products often do not remain fixed at such levels. They often improve to the point that, eventually, the new approach is a workable solution for a wider range of applications and customer use cases.
 
Having lived long enough to see the “smart guys” proven quite wrong, I am careful never to argue something really cannot be done. Sometimes, somebody, or another company, is able to do so, even when a reasonable, smart, experienced practitioner “knows” it cannot be done.

No comments:

Many Winners and Losers from Generative AI

Perhaps there is no contradiction between low historical total factor annual productivity gains and high expected generative artificial inte...