Field Notes

There is at least one certainty where Cambridge Analytica is concerned. If forty thousand people scattered across Michigan, Wisconsin and Pennsylvania had changed their minds about Donald Trump before 8 November 2016, and cast their votes instead for Hillary Clinton, this small London-based political consultancy would not now be the subject of breathless headlines and Downing Street statements.

More than two-thirds of Americans get their news from social media, and the vast majority of that group uses Facebook.

After a massive study from MIT found that fake news traveled faster than accurate news on Twitter, the Dartmouth political scientist Brendan Nyhan told me: “We can study Twitter all day, but only about 12 percent of Americans are on it. It’s important for journalists and academics, but it’s not how most people get their news.”

Almost unique among American companies, Facebook is the outgrowth of one man’s sensibility.

The tech community maintains an entrenched belief that the law will never understand what it does, and certainly the higher-level court decisions made in Oracle v. Google do nothing to dispel that notion. Yet Alsup’s very existence is a challenge to this belief: a 72-year-old former engineering student who has been quietly and delightedly tinkering away in BASIC for decades, playing with his radios and his cameras, teaching his clerks and random journalists alike the things that he knows. In the same patient and meticulous manner that Alsup explained the ionosphere to me, he explained software to the Federal Circuit in his Oracle opinion. And because it was so precise and so specific, he forced the Federal Circuit to engage with software development more than it otherwise might have. Regardless of the Federal Circuit’s ultimate decision, a great deal of Alsup’s understanding of software remains preserved in appellate caselaw.

Ideas, not technology, have driven the biggest historical changes. When technology changes people, it is often not in the ways one might expect: Mobile technology, for example, did not augur the “death of distance,” but actually strengthened the power of urbanism.

People in the innovation-obsessed present tend to overstate the impact of technology not only in the future, but also the present. We tend to imagine we are living in a world that could scarcely have been imagined a few decades ago. It is not uncommon to read assertions like: “Someone would have been unable at the beginning of the 20th century to even dream of what transportation would look like a half a century later.”6 And yet zeppelins were flying in 1900; a year before, in New York City, the first pedestrian had already been killed by an automobile. Was the notion of air travel, or the thought that the car was going to change life on the street, really so beyond envisioning—or is it merely the chauvinism of the present, peering with faint condescension at our hopelessly primitive predecessors?

They called it the “end of history illusion”; people believed they had reached some “watershed moment” in which they had become their authentic self.2

Coal, notes the historian David Edgerton in his book The Shock of the Old, was a bigger source of power at the dawn of the 21st century than in sooty 1900; steam was more significant in 1900 than 1800.

Time capsules, after all, exude a kind of pathos: They show us that the future was not quite as advanced as we thought it would be, nor did it come as quickly. The past, meanwhile, turns out to not be as radically distinct as we thought.

In 2015, the top 100 firms had a combined market capitalization (the total value of a company’s outstanding shares) that was 7,000 times that of the bottom 2,000 firms. Twenty years ago, that multiple was just 31.

“The dominant pattern here is that Amazon is more likely to target products that are performing well on Amazon.com,”

To understand how the dynamic is playing out on Amazon, Zhu and a colleague identified 164,000 products sold exclusively by third parties on the company’s platform during June 2013—that is, items Amazon itself did not sell. Ten months later the researchers examined those product categories and found that Amazon had begun directly selling items in 3% of them, for a total of 5,000 items. Although 3% may not sound like much, Zhu says that it’s a significant amount over such a short interval and that it suggests Amazon may move into competition with many of its third-party sellers over the long term. The analysis also showed that products sold directly by Amazon often became the default option in search results.

“Their view is that the platform owners essentially run the platform as a lab, letting people innovate and compete against one another, and then cherry-pick the best products for themselves and capture the value,”

In 1987 Ernst Dickmanns and his team at the Bundeswehr University in Munich had their autonomous van drive at 90 kilometers per hour (56mph) for 20 kilometers (12 miles) on a public freeway. In July 1995 the first no hands on steering wheel, no feet on pedals, minivan from CMU’s team lead by Chuck Thorpe and Takeo Kanade drove coast to coast across the United States on public roads. Google/Waymo has been working on self driving cars for eight years and there is still no path identified for large scale deployment. It could well be four or five or six decades from 1987 before we have real deployment of self driving cars.

A lot of AI researchers and pundits imagine that the world is already digital, and that simply introducing new AI systems will immediately trickle down to operational changes in the field, in the supply chain, on the factory floor, in the design of products. Nothing could be further from the truth. The impedance to reconfiguration in automation is shockingly mind-blowingly impervious to flexibility.

When you want to change information flow, or control flow, in most factories around the world it takes weeks of consultants figuring out what is there, designing new reconfigurations, and then teams of tradespeople rewiring and reconfiguring hardware. One of the major manufacturers of this equipment recently told me that they aim for three software upgrades every twenty years.