A lot of blog-zagging led me to Brad de-Longe's Old Rules for the New Economy:
After the heroic age, firms make money by giving customers exactly what they want.
As long as automobile prices were falling and quality rising rapidly, Henry Ford could do very, very well by riding the leading edge of the technology wave: making a leading-edge car and letting the customer choose the color and the options, as long as the options were zero and the color was black. But after the 1920s Ford got overwhelmed by Alfred P. Sloan's General Motors, which figured out how to retain most of Ford's economies of scale while at the same time proliferating brands, models, styles, and colors to get close to individual consumer demand. Before GM, no one knew what kind of options car buyers really wanted. Today, no one knows what kind of options computer and internet access purchasers really want--and no one will until a computer and communications counterpart emerges to play GM against Microsoft and Intel's version of Ford.

Which reminds me of Bob Frankston's on-going campaign.
it is important to frame the problem in terms of simple principles:

* Connectivity. This is simply transport without favoring any particular service whether it is voice or the web or television or other data. The underlying transport technologies must enable the creation of new services rather than favoring the old ones.
* Advocacy. With the inherent conflict of interests between the transport and service business the only way to assure that the incentives are aligned is to separate the two. Again and again the incumbents demonstrate the lack of incentive.

Indeed, Brad's been doing it for years