brAIn drAIn

Photo by Orbon Alija via iStockPhoto / Getty Images

Brain drain is the “gentrification” issue of emerging markets. Real estate appreciation and improving neighborhood conditions build prosperity, at least to those who own local property. That increase in wealth though can have a negative effect on those excluded from the market, who suffer rising prices from rents to local services. Ultimately, they are often pushed out of the neighborhood entirely. One’s sympathies tend to align either with material wealth and progress (a richer, safer neighborhood is fundamentally a good development), or with the people forced to scatter.

Brain drain is the inverse, but instead of real estate, it’s about talented individuals. Countries educate and invest in their citizens with the hope of improving their national economies, but open global mobility means that those individuals have the right to migrate to the best opportunities available to them in the world. China and India, for instance, have well-trodden pathways to academic excellence, and those pathways often lead to excellent and well-paying jobs in the West and specifically Silicon Valley. While by no means an easy path, it is absolutely possible to grow up in the impoverished Chinese countryside and eventually become a multimillionaire with a mansion off University Avenue in Palo Alto.

Brain drain (as one can observe from the valence of the term) is the bête noire of development. Education is critical to push countries up the income ladder, with well-educated entrepreneurs launching the companies and running the agencies required for wealth formation that creates prosperity for everyone else. When these elites depart for better overseas opportunities, their home countries lose access to that engine of future riches, leaving them developmentally trapped. Once again, we see split sympathies, aligned either with the individual striving to make the best use of their limited time on Earth, or with the plight of the developing countries left behind in their wake.

Much like my opinion on gentrification, I find the whole concept of “brain drain” to be anathema to a pragmatic moral system: no one should be forced to develop an industry in their homeland when they can generate significantly more prosperity for themselves, their families and their communities in the broader global market (a subtext I explored a bit in “Brainwash Departures”). Indeed, immigration isn’t zero sum, but can often be a key way to bring capital home to developing nations. About a quarter of countries like the Philippines rely on remittances from overseas workers for about one-tenth of their GDPs, while more than a dozen countries receive one-fifth of their GDPs from overseas sources.

The sudden arrival of reasonably proficient AI models the past few years though is rapidly changing the debate around brain drain and the development traps that plague many emerging market economies. Dani Rodrik and Joseph E. Stiglitz, two superstar developmental economists, wrote a summary of their views in January on where the world stands.

In their history, heavy industrialization was the key route out of the poverty trap for many countries, including Taiwan, Singapore, South Korea and others. As industrial sectors became saturated with competition, India and countries like it who developed later sought a new strategy. Instead of industrials, India focused on services built on top of the newly-emerging internet in order to take advantage of its vast population in a cost-effective way. The country became a global hub for medium-skilled services like business process outsourcing through companies such as Infosys and Wipro.

Now, technologies like artificial intelligence threaten to remove that step — and others — entirely from the development ladder, forcing the world to search anew for pathways forward. Criticism of human brain drain that were pervasive in the 1990s and 2000s have given way to the new threat of … brAIn drAIn (yes, dammit, I am proud of that), of AI bots that usurp the vast labor that these economic sectors absorbed.

Take Cognition, the startup that has created a firestorm in the past few weeks with the release of its automated software engineer named Devin. Devin can code reasonably sophisticated software to specifications, and can either act as a copilot alongside a seasoned developer or work entirely autonomously. There are hundreds of millions of lines of code written for internal enterprise applications which aren’t novel yet still take prodigious time, resources and attention to detail to get right. Devin and other coding bots like it have the potential to make all that work disappear into the ether of the cloud.

As I noted in “Garrulous Guerrilla” a year ago, the challenge that AI poses is that it is precisely these low-to-medium-skilled creative jobs — even among well-paying professions like software engineers — that are most at risk for elimination. As Rodrik and Stiglitz highlight, these service-sector jobs are plentiful and absorb much of the excess labor capacity of the globe’s talent pool. Critically, they also offer a way to higher-skilled positions as workers gain experience.

For countries like India, which desperately needs hundreds of millions of middle-class jobs (youth unemployment stood at 44.5% while college graduate unemployment held steady at a whopping 40%), AI is an employment disaster in the making. Workforce analysts have identified outsourcing in service markets like call centers, data entry, business process automation and broader IT support as among the industries most likely to be replaced by AI in the next few years.

With employment in services expected to considerably narrow, maybe these countries should head back to the industrialization path blazed by the Asian tigers a few decades ago? What Rodrik and Stiglitz emphasize is that industrialization is not much of a benefit either anymore, given the highly-scaled nature of manufacturing today. In their words:

However, just as global economic integration, [export-oriented industrialization], and [global value chains] became the corner pieces of economic development strategy, their benefits were being undermined by a process of “premature deindustrialization” in developing countries. The primary culprit was skill- and capital-biased technological changes in manufacturing. These changes increased labor productivity substantially in the advanced economies where innovations originate. But they also undercut the comparative advantage of low-income economies in traditionally labor-intensive manufacturing. The quality and technological standards set by leading firms in [global value chains] rendered labor-intensive production in export-oriented sectors even less viable.

In other words, the world has moved away from affordable workers assembling parts into machines and toward automated factories where the critical skillset is the engineering and maintenance of robots. America’s industrial giants fell in the 1970s and 1980s because they were burdened with high-price labor costs against far cheaper global competitors. Today, our top manufacturers win through superior technology which can lower ultimate marginal costs and outcompete those previously cheap armies of human laborers.

Even worse for the developing world, that focus on automation and robotics is pushing factories toward more upfront capital costs in order to be more cost competitive in the long run. Access to plentiful and cheap capital is thus the crucial ingredient for firm success, an ingredient that’s widely available in developed nations but impossible to find in developing ones. As the U.S. Federal Reserve has raised interest rates, developing nations have been strangled by debt loads. As the World Bank wrote at the end of last year, “In the past three years alone, there have been 18 sovereign defaults in 10 developing countries—greater than the number recorded in all of the previous two decades. Today, about 60 percent of low-income countries are at high risk of debt distress or already in it.”

In short, the world needs an entirely new approach to development. Rodrik and Stiglitz emphasize that the future of industrial policy is focused on the green transition and what they describe as “productivity enhancement in labor-absorbing, mostly non-traded services” (think small businesses in the local community). Both are new and relatively unexplored avenues, much like how the industrialization and service-sector formulas of the past few decades were untrodden before implemented.

Unfortunately, the base challenge remains the same: how should poor nations upgrade the skills of more of their workers faster, equip them to be entrepreneurial and create their own opportunities, and train them to be rapidly adaptable to change? AI might be part of the story, offering more individual educational instruction to people who might not otherwise be able to have access to a high-quality school. Maybe. Yet, “brAIn drAIn” fundamentally benefits the developed over the developing just like human brain drain did — and the gap has accelerated. That portends extreme troubles ahead for the material needs of billions of people, and a whole new set of economic, social and national security challenges in the years ahead.

Lux Recommends

  • Our scientist-in-residence Sam Arbesman recommends Adam Rogers’s feature in Business Insider on the rise and fall of Smashmallow. “Maybe you've heard of Smashmallow; maybe you even bought some. In the couple of years before the pandemic, they were everywhere. Now? Pfft. The problem wasn't the marshmallows — they were, by all accounts, delicious. The problem was scale. Smashmallows were designed to look like an artisanal, boutique product, but that wasn't enough for [Jon Sebastiani]: He wanted to manufacture billions of them, to build a company that would bestride Candyland like a squishy colossus. That meant he had to grow fast and figure out the engineering on the fly — the classic entrepreneurial strategy of Silicon Valley. When it works, you get Tesla; when it doesn't, you get Theranos.”
  • Kathryn Schulz offers a panoramic look in The New Yorker at the threat of solar storms on our (very digital) way of life. “In space weather, every day is a sunny day. There is no interstellar rain, no interplanetary snow, no sleet spinning off the rings of Saturn; all the phenomena we call space weather originate on the sun. And so, to start, you must shed the idea—implicit in our meteorology and omnipresent in our metaphors—that the sun is a mild and beneficent force, a bestower of good moods and great tans. In reality, the sun is an enormous thermonuclear bomb that has been exploding continuously for four and a half billion years.”
  • Sam highlights Benj Edwards’s reminisces on what the world was like before the internet. “Before the Internet, giving your home address to others (even in a printed periodical) didn’t instantly mean an invasion of privacy like it does today. There were no consumer satellite maps like Google Earth to see exactly where you lived from the sky by address, and no Google Streetview to get a look at your street and house. If you wanted to check out where someone lived without there permission, you had to physically travel to that location—and that behavior was generally called ‘stalking.’”
  • Molly Templeton has a fun essay “On Letting Go of the Idea of ‘Keeping Up’,” in line with what I described way back in “Atomistic Literacy.” “I didn’t want to talk about what I’d been reading in the way this person wanted me to respond. I suddenly wanted to hold my cards, and my books, extremely close to the chest. Reading felt gamified, like a thing where you went down a list of titles and got points for which ones you’d read.”
  • Finally, Sam highlights Will Douglas Heaven’s interview with OpenAI’s first artist-in-residence, Alex Reben. “In a critical sense, again, going back to photography, the world is flooded with images and there are still people making great photography out there. And there are people who set themselves apart by doing something that is different.”

That’s it, folks. Have questions, comments, or ideas? This newsletter is sent from my email, so you can just click reply.