Tuesday, July 28, 2015

Early Indications July 2015: Crossover Points

I recently read an enjoyable study of the airport as cultural icon (Alastair Gordon’s Naked Airport; hat-tip to @fmbutt) and got to thinking about how fast new technologies displace older ones. Based on a small sample, it appears that truly transformative technologies achieve a kind of momentum with regard to adoption: big changes happen rapidly, across multiple domains. After looking at a few examples, we can speculate about what technologies might be the next to be surpassed.

Gordon makes uncited references to air travel: in 1955, more Americans were traveling by air than by rail, while in 1956, more Americans crossed the ocean by plane than by ship. (I tried to find the crossover point for automobile inter-city passenger-miles overtaking those of railroads, but can only infer that it happened some time in the 1920s.) This transition from rail to air was exceptionally rapid, given that only 10 years before, rail was at its all-time peak and air travel was severely restricted by the war.

Moving into another domain, I was surprised to learn that in 1983, LP album sales were surpassed not by the CD but by . . . cassette tapes; CDs did not surpass cassettes for another 10 years. In the digital age, the album is no longer the main unit of measurement, nor is purchasing the only way to obtain songs. This shift in bundle size is also occurring in news media as we speak: someone asked me the other day what newspaper(s) I read, and it struck me as odd: I can’t remember when I last had a physical paper land on my porch. That’s the other thing about these crossover points: they usually happen quietly and are not well remembered.

The smartphone is taking over multiple categories. Once again, we see a new unit of measurement: in the film camera age, people developed rolls of film, then perhaps ordered reprints for sharing. (That quiet transition again: can you remember the last time you took film to the drugstore or camera shop?) Now the unit of measurement is the individual image. Interestingly, digital still cameras surpassed film cameras in 2004, but not until 2007 were there more prints made from digital than from film. After 2007, digital prints have steadily declined. Furthermore, digital cameras themselves have been replaced by cameraphones: only 80 million point-and-shoot digital cameras shipped in 2013 and that number is dropping to well under 50 million this year, while smartphone sales are on target for about 1.5 billion units this year.

Standalone GPS units, MP3 players, and video camcorders (with GoPro being a notable exception, albeit in relatively tiny numbers) are other casualties of the smartphone boom. Landline-only houses were surpassed by cellular-only in 2009. Smartphones surpassed PC sales back in 2011.

The implications for employment are tremendous: Kodak employed 145,000 people in 1988; Facebook, a major player in personal image-sharing, has a headcount of about 9,000, most obviously not working on photos. Snapchat has 200 employees at a service that shares 8800 images EVERY SECOND, a number Kodak could not have conceived of. When these technology shifts occur, jobs are lost at a greater rate than they are gained. Railroads employed more than 1.5 million Americans in 1947; it’s now about a sixth of that. U.S. airlines, meanwhile, employed a peak of about 600,000 workers in the tech boom of 2000, well less than half that of the railroads, in a more populous country with more people traveling.

Let’s look at the smartphone. Given globalization, what used to be U.S. telecom numbers no longer equate. AT&T employed around a million people at its peak; right now AT&T plus Verizon (which counts cable TV and other operations) employ roughly 425,000 people. Apple’s 2015 headcount of 63,000 includes about 35,000 retail employees and about 3,000 temps or contractors. Samsung is a major player in world telco matters, but figuring out how many of its 275,000 employees can count toward a comparison vs AT&T is impossible. All told, more people have more phones than they did in 1985 but employment in the phone industry looks to be lower, and lower-paying, given how many retail employees now enter the equation.

Coming soon, we will see major changes to ad-supported industries. Already newspaper revenues are in serious decline. Digital ad revenue is already higher than newspaper, magazine, and billboard combined. “Cable cutting” is a very big deal, with clear demographic delineations: a 70-year-old is likely to read a paper newspaper and watch the big-4 network evening news; a 20-year-old is highly unlikely to do either. Comcast announced in May that it has more Internet-only subscribers than cable-TV subscribers, and the unbundling of cable networks into smartphone/tablet apps such as HBO-Go will likely accelerate.

In personal transportation, there could be two major upheavals to the 125-year-old internal combustion regime: electric cars and self-driving vehicles. Obviously Tesla is already in production in regard to the former transition, but the smartphone example, along with such factors as Moore’s law, cloud computing, and an aging Western-world demographic could fuel rapid growth in autonomous vehicles. In regard to cloud computing, for example, every Google car is as “smart” as the smartest one as of tonight’s software upgrade. Given the company’s demonstrated expertise in A/B testing, there’s no reason not to expect that competing models, algorithms, and human tweaks will be tested in real-world competitions and pushed out to the fleet upon demonstrated proof of superior fitness. 

There are many moving parts here: miniaturization, demographics, the rise of service industries relative to manufacturing (including cloud computing), growing returns to capital rather than labor, and so on. The history or technology substitutions and related innovations does have some clear lessons however: predicting future job growth is perilous (in 1999, the US Bureau of Labor Statistics was bullish on . . . desktop publishers); infrastructure takes decades while some of these cycles (Android OS releases) run in months; and the opportunities in such areas as robotics, AI, and health care are enormous. The glass may be half-full rather than half-empty, but in more and more cases, people are looking at entirely different scenarios: Kodak vs Snapchat, as it were. Whoever the next US president turns out to be will, I believe, face the reality of this split, perhaps in dramatic fashion.