Over the course of the last two decades, Silicon Valley has committed itself to two interlocking beliefs.

One: Getting the maximum number of people online at the fastest rate possible will lead to good outcomes.

Two: You can give any user on earth the same product, and they’ll use it in the same way. Therefore, geography and culture don’t really matter.

Together, these two precepts form the nucleus of what we have come to call “scale,” the doctrine that businesses should put rapid adoption of their product above all other considerations, even profits. One year was enough for Facebook to acquire its first million active users, six years less than Apple took to sell its first million computers, eleven years less than Ford took to sell its first million cars, at least six centuries less than ancient Rome took to reach one million living inhabitants.

Of course, Rome, Ford, and Apple had all scaled up as quickly as they could. What changed was the mobile internet, which eliminated any ceiling on the rate at which a company could grow.

When investors today talk about “scale,” they mean Facebook-style exponential growth. On a graph, this looks like a “hockey stick” — a flat period of investment and development followed by an almost instantaneous spike. Originally coined to describe the recent rise in global temperature, the term “hockey stick” has since been appropriated by venture capitalists looking for quick returns on the low-interest money that has kept the global economy afloat since the 2008 financial crisis. “Scale,” both a noun and a verb, is how you get a hockey stick.

Scale is more than a way of doing business; it is a lens for viewing the world. The greater a country’s population, internet penetration, and income, the more it matters. Other countries are lumped together in quarterly financial reports as “Rest of World,” a term with echoes of terra incognita on an old map. Indeed, so-called “Rest of World” countries are rendered on user growth charts as blank space, virgin territory for the hockey stick to scale up into.

Consider Myanmar. Between 2012 and 2013, the cost of one SIM card plunged from upward of $1,000 to under $2, as Myanmar’s military-backed authoritarian government, which was beginning to liberalize, opened up the telecommunications market to competition. The consensus in Washington and Silicon Valley was that good things would come from allowing Myanmar’s people to access platforms that had been founded and designed by Americans, whether or not those coders could locate Myanmar on a map.

The view from the ground, however, was different. That same year, Aela Callan, a journalist working in Myanmar, flew to Facebook’s California headquarters, where she met face-to-face with one of the company’s senior executives and delivered a warning. Messages like “Kill all Muslims” were proliferating across Facebook’s platform, combining the malice of village gossip with the internet’s speed and reach. Already, there had been cases where mobs of radical Buddhists, stirred by false rumors, had risen up to burn Muslim schools and mosques. Callan was among the first to draw what is now a widely accepted link between the killings (then fewer than 500) and Myanmar’s adoption of the mobile internet (then around 10%). Her warning was explicit. “I often wonder, in Myanmar, if I’m seeing the beginnings of the world’s first genocide spurred on by the virtual world,” Callan said, in a public lecture given at Stanford several months later, in 2014. “The good news is, I don’t think it’s too late.”

It wouldn’t be correct to say that Facebook did nothing. Some staff members joined a closed Facebook group, where Myanmar-based activists could flag hateful posts to be taken down. “Local journalists who are paid literally $150 a month by their employers were working for free as Facebook’s early warning system,” Callan recalls. At that time, Facebook employed only one Burmese speaker worldwide in Dublin, while its policy operations for Myanmar were run out of a regional office in Australia.

Staffing up in a market as small as Myanmar or tailoring Facebook’s architecture to take local conditions into account were not viable options. Such measures would have violated the notion that technology is neutral with respect to geography, because all users are the same. Silicon Valley, in other words, had convinced itself that it was delivering products with the universal homogeneity of Coca-Cola but even more profitability, because they didn’t require any in-country trucks, syrup factories, aquifers, or bottling plants. But they were wrong. Internet platforms are not at all like Coca-Cola. In fact, they provide an entirely new way of engaging with the world, a system for finding and sharing refracted bits of consciousness that mutate upon contact with each market. And despite the neocolonial assumption that high-speed internet access could be only a civilizing force, in many cases it proved to be an accelerant for the worst impulses coursing through a society.

Before Facebook, Callan recalls, radical Buddhist monks had to copy hate-filled speeches onto DVDs and transport them by bus around the country. Now, they could share them instantly. Callan remembers visiting a monastery and seeing groups of monks, each holding several cell phones with which to game Facebook’s algorithm by artificially inflating likes and shares of hateful content. “This was years before Cambridge Analytica,” she said. “They thought they were rock stars.”

A protest against the killing of Muslims in Myanmar in Kuala Lumpur on 25 March 2013


Original caption:

Myanmar Muslims living in Malaysia show the playcard against the killing of Muslims in Myanmar during a protest in Kuala Lumpur March 25, 2013. Protested sectarian clashes between Buddhists and Muslims in Meikhtila Myanmar heartland villages, undermining government efforts to quash the violence that has killed dozens of people and displaced thousands of Burmese. Photo by Samsul Said.
Samsul Said via Getty Images

By late 2017, more than 10,000 Rohingya Muslims had been killed. More than 700,000 had fled the country after losing their homes. Perhaps more significantly for Silicon Valley, the same kinds of disinformation and hate speech that had appeared in Myanmar were now showing up in places where CEOs lived, threatening to undermine the integrity of elections and the broader democratic order in Europe and the United States.

Since then, Facebook has been largely successful at reframing its failure in Myanmar — a failure to fully address direct and explicit warnings, from Callan and several others — as merely a tragic lack of foresight. But when it comes to the bottom line, ignorance of local markets isn’t a bug, it’s a feature. Today, young founders are told that their success depends on skipping over the scut work of due diligence, cultural anthropology, and political risk analysis. “If you’re not embarrassed by your first product release, you’ve released it too late,” said Reid Hoffman, the LinkedIn co-founder, in an interview with Mark Zuckerberg for his “Masters of Scale” podcast. Of course, this kind of mentorship in what Hoffman calls “blitzscaling” is almost entirely exclusive to the capital-rich West. The so-called Rest of World is seen as fuel for the growth chart, not a place to look for the next big idea.

There is no shortage of companies that have asked governments to get out of the way so that they could save the world, only to wind up aggravating the problem that they were scaling up to solve. Uber claimed that the gig economy would improve drivers’ lives; many wound up sleeping in their cars to survive on a fraction of a living wage. Airbnb advertised its regulation-flouting rollout with the claim that it was a friend to lower-income tenants. Instead, it attracted a new class of absentee landlords that drove up rents and priced many tenants out of their homes. Servers made by Blue Coat, a cybersecurity company from California, were used to surveil and identify Syrian dissidents by the government of Bashar Assad, which went on to murder thousands. Later, in 2016, Blue Coat was acquired by Symantec for $4.6 billion.

Beyond these direct effects of companies and their products, the demand that scale places on raw material is transforming the physical world. Exponential growth requires more data centers to feed our smartphones, more Amazon fulfillment centers to deliver the latest gadgets, and more smelters to salvage e-waste scrap after Apple and Samsung deliberately slow products down, forcing us to throw them away and buy new ones. Then there are the content moderators who spend their days inundated with the internet’s most violent, hateful, and sexually exploitative content. The number of people whose job it is to do this awful work is also increasing at scale.

Scale is not all bad. One can’t imagine how product recalls, disaster relief, or vaccine distribution would work if society lacked the capacity to scale up quickly. What’s harder to understand is what would have been lost if Facebook, Uber, and Airbnb had grown at a slightly slower rate, one that would have given them time to anticipate potential problems and actually learn something about the people and countries who would be affected by their products before pushing them out the door.

For the so-called Rest of World countries, the risk that comes with scale is ignorance of the local, which brings with it the potential for catastrophe. In larger and more developed countries, an unhealthy obsession with scale can cause a company to trade away its reputation for the sake of access to a market that’s too big to pass up. Consider Google’s long flirtation with China. In 2018 and 2019, the company was reportedly working on a search product called Dragonfly that would deliver censored results and compile a database of users’ locations. Google’s Dragonfly project was eventually terminated after leaks and employee protests. Google employees have also protested the company’s work on Project Maven, an AI system to automate lethal drone targeting for the Pentagon, and a cloud computing system for U.S. Customs and Border Protection, which carries out the Trump administration’s immigration policies. The temptation to take bigger risks — financial, social, and moral — increases the further up you go on the growth curve. The question over the next two decades is whether we’ll have the collective wisdom to flatten it.

Arguments against scale are easy to make from where we sit, in the first days of a publication that doesn’t answer to shareholders. Walking the walk will be harder. Over the coming days, we’ll be closely watching our own hockey stick, with page views and clicks on the y-axis, and the illusion of infinite acreage to be seized in the northeast. The “Rest of World.” The next billion, and the billion after that. The future. The temptation is always to run there as fast as one can. Experience would seem to indicate that it is often better to walk. Or even to stop for a moment, and listen.