The failure of Google Fiber to connect America’s cities to ultra-fast gigabit networks took a bizarre turn this week: San Francisco, the incubator of the apps economy, is seriously considering a billion dollar bond issue to build a whole new network. The project is driven by San Francisco Supervisor Mark Farrell, who appears to be taking advice from law professors Susan Crawford and Cathy Sandoval. (Note: San Francisco’s county and city limits are the same, so the town is run by county supervisors rather than city councilmembers.)

The rationale doesn’t appear to have much connection to the project. A 2013 study maintains that 12 percent of the city’s residents don’t subscribe to broadband at home,  despite universal availability of networks provided by Comcast, AT&T, Webpass, Sonic, and others. The study says cost is one of barriers to adoption, but Comcast offers subsidized services for $9.95:

To address the digital divide, Comcast’s Internet Essentials program offers Internet service at 10 Mbps for $9.95, low-cost computers and training to low-income families with children eligible for free and reduced lunches. In August of 2015, Comcast launched a pilot program to expand Internet Essentials to low-income seniors 62 years and older. Via the senior pilot, which was created in partnership with the City and County of San Francisco, Comcast is working with the Department of Aging and Adult Services and local non-profit organizations to connect seniors to home internet and discounted computers, and with SF Connected to provide computer training to seniors throughout the city.

So it’s not clear what the city hopes to accomplish.

San Francisco has a Housing Problem

By some measures, San Francisco is the most expensive city in the nation. Housing costs account for three quarters of the cost of living there, with median rent greater than $3,000/month. Broadband isn’t cheap: Comcast charges $64.95 for the 25 Mbps tier of broadband without a TV and phone bundle, and $49.95 for the basic 10 Mbps tier. These prices are obviously dwarfed by housing costs.

So it’s peculiar that the city is considering a billion dollar bond issue to rewire itself for FTTP. As preliminary estimates suggest that the operating cost to the city for each residential connection will be $43/month, the project only makes sense on the assumption that today’s 10, 25, 50, and 100 Mbps connections are preventing the city from prospering.

If poor people are discouraged by today’s $9.95 fee from Comcast for basic service, it’s hard to imagine they’ll have a change of heart if speeds increase. Rather than taking on a billion dollars in debt, it would be more sensible for the city to subsidize faster broadband for the poor on the current networks. They certainly have the capacity.

Gigabit Fiber is its Own Reward

Ultra-high-speed, all-fiber networks have sex appeal. In the early ’00s, conventional geek wisdom maintained that gigabit networks had world-changing potential. A number of us campaigned for such networks to be constructed around the nation without much regard for the applications they would spawn or even how they were financed and operated. This line of thinking was behind Verizon’s FiOS program that aimed replace the copper telephone network with fiber to every residence on a short timeline.

This wasn’t a bad idea. Copper telephone networks were nearing the end of the line, and fiber was the future. Fiber optics are essential to the future of networking, without question. As with many forecasts, however, the details matter. The difficult questions are when and where today’s copper wires will be replaced by fiber.

The vision of fiber to every home and office had to be modified as technology narrowed the gap between fiber and other technologies. Not only did mobile become the preferred access technology for most of us, the technologies that push bits through (and around) copper wires continued to improve.

Multi-Modal Networks Rule

Fiber remains the best way to push bits between fixed locations, but the other technologies are more than good enough to meet many – if not most – practical applications. Freight trains are ideal for moving cargo across long distances, but there’s a lot to be said for the flexibility of delivery vans. And in the broadband context, fiber is the railroad and copper and mobile are the vans, trucks, cars, and bicycles.

Copper networks are also paid for. Hence, the development of Full Duplex DOCSIS 3.1 – a cable based system that will operate at 10 Gbps with equal speed in both directions at the same time – is more significant than many realize. And 5G mobile networks that will ultimately achieve gigabit speeds are even more important than DOCSIS.

All of these technologies are complementary parts of the multi-modal networks of today and tomorrow. While the people we geeks call “fiber bigots” insist that one and only one technology will do, our experience with networking has always shown the value of diversity. We’ll get to higher speeds with more flexibility, greater reliability, and better security when we need these features more than we do today. Perhaps these things will come a little sooner than we want them to, but not much before we’re willing to pay for them.

Closing the Divide

All of this being said, the growing gap between the technically literate and the rest of us is concerning. There is a stark and growing divide between technology elites and the common person.  The advent of robotics, global trade, and highly capable AI systems has enabled a portion of the population to live in ways that are far beyond the lifestyles of our grandparents and great grandparents. “The future is already here, it’s just not evenly distributed”, as William Gibson said.

The technology gap is not simply a matter of residential broadband speeds. The networking gap is not even that narrow as we spend more time using our mobile devices than sitting at desks using wired broadband networks. The schools need good broadband connections, but they also need Wi-Fi, fast computers, insanely great smartphones, and highly-capable software.

And all of these things require users with the curiosity, imagination, drive, ambition, and the nerve to explore the digital world. Hence, if I were the Emperor of San Francisco I would subsidize broadband for the poor and spend most of my budget on building enthusiasm for digital life.

The city would be wise to take a much broader view the problem it seeks to solve. San Francisco is home to a significant tech sector that currently lives in an uneasy relationship with the rest of the city. Bridging the cultural divide is more productive than trying to build a fifth, sixth, or seventh broadband network built on contempt for the poor and illiterate.