Hacking Priomoridal Soups

Why are some regions of the world just brimming with entrepreneurial talent?

Editor’s Note: This is a special column by Yudhanjaya Wijeratne, a Sri Lankan-based science fiction writer and activist, known for his novel The Salvage Crew and for his non-profit fact-checking organization Watchdog. This is his first column for the Riskgaming newsletter.

A curious tale is unfolding in consumer electronics. Hifiman has been slowly winning the hearts of audiophiles with their sophisticated planar magnetic headphones, competing not just in budget segments, but even in the rarefied spectrum of thousand-plus-dollar audio gear against more established rivals like Sennheiser, Audio-Technica and Beyerdynamic.

This is not a tale unique to audio. In action cameras, brands like Insta360 and DJI are eating away GoPro’s throne. A so-called “little giant,” Insta360 now claims over 40% of the world’s market share for action cameras. And speaking of DJI, the company that began inside a dorm room now produces cutting-edge consumer drones that are as equally useful at weddings as on battlefields. YouTube phone reviewers regularly compare former budget brands and point out the obvious: what began as frankly cheap shit is now legitimately quite good. The age of cheaper, cloned versions of Western products is over: much of this technology is innovative, even groundbreaking.

That the three brands mentioned are all Chinese shouldn’t be surprising to anyone familiar with the country’s scrappy electronics operations, particularly in Shenzhen. Home to the world’s largest electronics market in Huaqiangbei and three leading universities, it marries cheap hardware and open markets at nearly unprecedented scale. Industrial policy — both in nurturing universities and attracting large, multinational product companies — have attracted a dense pool of engineers who can actually make things.

And make things they do. Talented tinkerers have ready access to every possible component imaginable, while early and visionary adopters abound. This makes for a primordial soup of cheap, high-speed iteration. Budget hobbyist markets incubate higher-value domestic and overseas customer bases. Global ecommerce platforms like AliExpress and Linsoul can accelerate the reach of nascent technology companies.

My primordial soup thesis is simple: regions like Shenzhen and others I’ll talk about shortly combine a deep bench of engineering talent with dense markets for these engineers to buy cheap components and then create, test and cook their ideas. Intriguing products spawned in the soup can accelerate through ecommerce platforms and video review sites, helping some ideas to crawl out of the water and start eating everyone’s lunch.

Shenzhen is perhaps the most well-known primordial soup, but hardly the only one globally. Yet, the one notable country that goes without one of these regions is America, and that is a critical gap for a country seeking to reindustrialize.

Taking stock of the globe’s primordial soups

It's fascinating to see how fast a competitive market can evolve. In 2019, for instance, Dan Nosowitz wrote a fantastic article in The Verge about the world of Chinese hi-fi audio. What Dan observes is that initial primordial soup; a haze of unknown brands with all manner of origin stories:

Some start as original equipment manufacturers, or OEMs, meaning they actually make that name-brand stuff for Beats or Shure or whomever. “Some of them are just trading companies, some are engineers that left another factory, it’s every possible variation,” says Mike Klasco, an audio engineering consultant who has been scouting factories throughout Asia for 35 years. “If you have a van and a bottle of glue,” Klasco says, “you can be in the business.”

This is not hyperbole. Shenzhen is a place where you can glue together your own smartphone if you want to, and in fact, people have. Combine that with young, enthusiastic engineering talent and it seems like you have one hell of a potent recipe. In 2019, Hifiman was under the radar; today, on Reddit, enthusiasts eagerly await their Hifiman HE6v2 and debate the Susvara Unveiled ($8,000).

It isn’t just Shenzhen. Taiwan needs no explanation for its electronics prowess; TSMC is there, as is almost every major name in computer hardware like AMD, Nvidia, MSI, ASUS, MediaTek, Foxconn and Wistron. While Supermicro, which manufactures server motherboards, may be headquartered in the States, it is also a product of Taiwan. Much of the world’s compute capability lies on one island, and hot on compute’s heels is an emerging generation of startups from biosciences to materials. What’s encouraging tinkerers is Guang Hua market in Taipei, a sprawling, massively outsized electronics bazaar close to National Taiwan University, the country’s leading engineering school.

Let’s turn to an even smaller island: Singapore. Sim Lim, Singapore’s infamous electronics hub, was largely known to the outside world for its scams; and outside of Razer, Singapore’s technology ecosystem has largely been dominated by U.S. firms setting up outposts.

But over time, Sim Lim has evolved into a micro-Shenzhen of sorts. Lots of young engineers have been trained by Nanyang Technological University and National University of Singapore in response to rising demand, and the list of Singaporean companies making waves in the tech space has grown with them.

Silicon Box went hard on chiplets, building a $2 billion semiconductor foundry. Skye Renewables is spreading rapidly through South Asia’s energy markets. Flex is a juggernaut with 170,000+ employees, doing everything from bespoke product design to manufacturing: it became the world’s third-largest original design manufacturer just behind Pegatron. And in the humbler audio space, people who used to repair earbuds like FatFreq are now manufacturing their own.

Finally, I’ve seen an emerging primordial soup in Bangalore. On the occasions that I’ve ended up there, I usually find myself haunting MG Road; that’s where the best secondhand bookshops are. However, Bangalore has evolved far past my anecdotes, becoming a South Asian hub of startup activity, electronics engineering and design. Makerspaces have been popping up there since at least 2014 (when I first realized what a makerspace is); the government of India managed to lure companies ranging from TI to Oracle there, and molded technological universities like IITs into behemoths. A robust supply of components comes from the absolute madness of SP Road, which is where you want to be for wholesale electronics.

In the shadows of the WiPros, Infosys’s, and Accentures is a vibrant ecosystem of product-first engineering companies. Is it a coincidence that India’s space agency ISRO set up shop in Bangalore and managed to find talented engineers capable of putting a lander on the moon for just $74 million? Or — to stretch further afield — that India is increasingly building its own thriving product ecosystems from military hardware to mobile phones?

These four regions almost exactly mimic the early hacker days of Silicon Valley. Two guys in a garage building a computer in the shadow of a large corporation; someone with a van and a bottle of glue, launching Davids at Goliaths; it’s all the same spirit. America had its primordial soup days — RadioShack, Fry’s Electronics — and the electronics ecosystem around them spawned a generation of successes, Apple being the most notable. The Homebrew Computer Club still exists, if not in the U.S.

Soup for the Engineering Soul

The governments in all these regions did three things right. First, they went hard on STEM education. Engineers aren’t born, after all; they’re made. India has its IITs and other institutions; China has its universities; and from cradle to grave there is an overwhelming push towards science, math and engineering.

Second, these governments also played a key part in nurturing serious hardware ecosystems, from wooing foreign manufacturing operations to nurturing local startups. Taiwan’s TTA is a good example, as is Singapore’s Temasek; the Shenzhen government’s fingers are in almost every pie, while the Startup India program is early but useful. Their actions help to stitch dense networks of companies and opportunities, further attracting ambitious, creative and entrepreneurial people.

Third, they acquiesced to or actively developed markets. And not just any markets: fast, dense markets that are proper primordial soups where small businesses (the multicellular heroes of this story) play a major role and can dump their excess from almost any part of the supply chain. Hallam Stevens, analyzing Shenzhen’s markets, noted that:

The electronics economy depends critically not just on ‘makers’ but on all kinds of other labour. In particular, it depends on lower middle-class and low-class work—devices made by small factories and shops, sold by small enterprises and designed for the less wealthy, especially in developing countries. The human networks that connect these individuals are critical to the size, speed and density of the markets, allowing devices to be built and shipped rapidly, for parts and customers to be available.

Much like my biological metaphor, these local ecosystems swirl first as slow, informal affairs then accelerate as the density of life in the soup thickens. Informal entrepreneurs selling handmade parts to each other transform into startups searching for global product-market fit and delivering precision goods on wickedly-tight schedules. The rest of the evolution is well-trodden ground from the triple helix model and Michael Porter’s detailed observations on clusters in economies; although reality is a lot messier, more complicated and full of all sorts of horizontal and vertical interactions rather than neat, black boxes that feed into each other. Hence the primordial soup metaphor.

How important is market vitality?

Now, anecdotal evidence is only evidence of an anecdote, so to test the idea, I drew up a list of the world’s top 20 technology hubs:

I then did a back-of-the-napkin sort using all the open-source evidence I could find. What kind of DIY hardware markets do these places have? In true anime fashion, S-tier is the crème de la crème. “Decent” is “anything can be found, with enough patience.” “Limited” is “there’s no one place we can get everything, we’ll have to look around.”

There is an interesting pattern here: the fastest hardware product velocity is correlated to the best markets: the richest primordial soups.

Admittedly, this is not scientific evidence. My lists are biased by my perceptions and knowledge and what information I can find. There is also the question of the direction of causality. Did large companies create markets to serve them, or did markets create the behemoths sitting next to them? The historical evidence here though is very clear: the markets — at least early forms of them — came first.

In fact, there’s a set of well-explored tangents that explain this pattern more thoroughly. In Platforms, Markets and Innovation, edited by Annabelle Gower, several leading scholars explore research across economics and organization management to point out a central theme: platforms, especially when stretched over supply chains, form “building blocks that act as engines of innovation and redefine industrial architecture.”

Their very first exploration was of Microsoft Windows as such a building block, allowing developers to come together for organizations across the supply chain. Eventually, once network effects kicked in for both supply and demand, it became an unstoppable juggernaut of progress.

What are these electronics markets if not platforms of a more physical kind? To return to our analogy: a soup, within which products evolve and grow at blistering speeds?

Where’s the American primordial soup?

Overwhelmingly today, these kinds of primordial soups seem to be newly swirling in South Asia and Southeast Asia. The build-ship-iterate cycle is spinning up, and in some places it’s in full swing.

They may, if anything, be spinning down in America. Andrej Karpathy, one of the most influential figures in the field of AI, alluded to this problem in a tweet with his own biological metaphor: “I wished for tech to look like ‘a thriving coral reef’ ecosystem but sometimes it feels more like mostly plankton, a few clown fish, two tunas, and five killer whales circling above.”

Shenzhen, Taiwan and Bangalore, in particular, have done exceptionally well at nurturing that thriving coral reef. Singapore is joining the fray. In addition to market-leading companies, this situation also leads to phenomenal outcomes for the local labor pool; you end up with a bunch of scrappers that can build at lightning speeds, often on very tight but profitable budgets. It isn’t just a lost artifact of how things were simpler in the 80’s. It’s happening today, and it’s happening right now.

There are a few obvious counterexamples. In South Korea, the consumer electronics juggernaut of our times, innovation is largely driven by large corporations known as chaebol. Yongsan, arguably Seoul’s most famous electronics market, is very much on the decline, but that hasn’t stopped LG’s micro-OLED production or Samsung’s foundries. Japan is similarly dominated by large conglomerates, while a similar story is happening in Vietnam, several steps behind, where Vingroup produces everything from electric vehicles to smartphones.

Why the difference? It’s all about networks of small businesses. The best primordial soups egg on the kind of structures that can build fast, iterate fast, ship fast, and, above all, are extremely robust to the kinds of shocks that large capital, great expectations and market collapses bring. They get that a thriving coral reef is composed of millions of diverse organisms. Compare that to Korea, Japan and Vietnam where the state helped a handful of companies dominate. Or to the United States, where permissive antitrust laws allowed a Magnificent Seven to control most of the profits of the most profitable industry in the country.

If you do want a more vibrant, thriving ecosystem of small companies with outsized impacts and if you want a labor pool that isn’t dominated by a handful of large entities, then there’s a reasonable chain of logic here: fund STEM education. Find universities that do this well. Nurture manufacturing next to them, and then nurture cheap markets where curious young engineers can find parts. Give them the ability to get their hands on the stuff they need to build prototypes without breaking the bank. Let them cook.

Put these things together, and you have a recipe. You get people who can learn to out-engineer massively overfunded and sclerotic incumbents, launch moon missions, and quietly build the most advanced foundries in the world. The secret of primordial soups is their evolutionary potential: they can reach an exponential curve, and that’s a pattern that every innovation expert understands. Yesterday’s glued earpods will become tomorrow’s $8,000 Hifiman headsets.

Podcast: Industrial literacy and the fate of American progress

America was once the most progressive nation in the world — so what happened? Many analysts focus on policies that can accelerate the adoption of frontier tech and increase productivity, but Jason Crawford takes a more expansive and longer view of the challenge. As founder and leader of the Roots of Progress Institute and through his on-going publication of The Techno-Humanist Manifesto, Jason emphasizes that we have lost something important: our industrial literacy. America’s leaders no longer understand how prosperity was delivered from the Industrial Revolution onwards, and we’ve lost the ability to rebuild and expand wealth in its broadest conception for the next generation.

I talk with Jason about his manifesto and its focus on humanism, and then we walk through some of the major ideas he’s hoping Americans pick up. These range from more progress studies in high schools and colleges as well as a greater understanding about the value that technology delivers for quality of life to the importance of gratitude for our ancestors who delivered this prosperity to us and why technocrats and reactionaries can both be wrong about managing technological change.

🔊 Listen to “Industrial literacy and the fate of American progress”

The Orthogonal Bet: The Quest to Build the Fruitful Web

In this episode, Lux’s scientist-in-residence Sam Arbesman speaks with Laurel Schwulst. Laurel operates within many roles: designer, artist, educator, and technologist in which she explores — among other things — the intersection of the human, the computational, and the wonderful.

Sam wanted to talk to Laurel because of this intersection and particularly because of how Laurel thinks about the internet. As part of this, she helps to run HTML Day and its celebrations, promotes what is referred to as “HTML Energy,” and is even thinking deeply about what it would mean to create a PBS of the Internet. In other words, the Internet and the web are delightful and special for Laurel, and she wants more of that energy to exist in the world.

🔊 Listen to “HTML Energy and Creating the PBS of the Internet”

Lux Recommends

• Yudha’s column today isn’t the only thing he’s been working on. Through Watchdog and his counterpart Nimesha Periyapperuma, he’s built a digital twin mod for Cities: Skylines simulating Colombo, Sri Lanka. “Ultimately, what we realized was that there had to be some visual way to bridge the gap between professional expertise (often confined to academic papers and reports) and public understanding.” Yudha has also published a series of 32 maps of Sri Lanka that combine multiple datasets using machine learning to offer comprehensive visualizations for researchers and the public.
• Sam enjoyed Erik Ofgang’s look in The New York Times on how AI has accelerated the pace of unlocking the meanings of ancient texts. “Led by Enrique Jiménez, a professor at the Institute of Assyriology of the Ludwig Maximilian University of Munich, the Fragmentarium team uses machine learning to piece together digitized tablet fragments at a much faster pace than a human Assyriologist can. So far, A.I. has helped researchers discover new segments of Gilgamesh as well as hundreds of missing words and lines from other works.”
• In the Riskgaming tradition, I highlight Walter Pincus’s opinion in The Cipher Brief on how cyberattacks threaten America’s water supply. “One reason for the delay is the number of separate water and wastewater systems that exist in the U.S. — over 153,000 public drinking water systems and 16,500 public wastewater systems — and each is governed by multiple federal, state, and local authorities responsible for public health, environmental protection, and security measures.”
• Continuing Sam’s theme of ancient texts and AI, Ariel Sabar at The Atlantic looks at how a multi-disciplinary team of researchers is unlocking the Voynich Manuscript. “Researchers found that the manuscript’s roughly 38,000 words—and 9,000-word vocabulary—had many of the statistical hallmarks of actual language. The Voynich’s most common word, whatever it meant, appeared roughly twice as often as the second-most-common word and three times as often as the third-commonest, and so on—a touchstone of natural language known as Zipf’s law.”
• Via Michael Magnani, our favorite espionage reporter Zach Dorfman is back with another magisterial feature on the moral and psychological challenges facing NOCs — non-official cover — CIA operatives who often work for years under extremely stressful conditions. “Some had already begun to notice a certain drift in his bearing. The man with the uncanny ability to detach from his ‘real’ life, to fully inhabit another, no longer seemed quite so sure what his real life even was anymore. CIA officials knew that spending too much time so deep undercover could sometimes sever a person from himself. And that, generally, was when it was time to bring a NOC in from the cold.”
• Finally, Sam highlights Matt Clancy’s analysis of “The Decline in Writing About Progress.” “Is the accelerated drop in the frequency of the word ‘progress’ across English, French, German, Italian, and Japanese texts, all in roughly the same decade, evidence of a global vibe shift? It seems at least possible - if I was going to think of an event that might lead to a global reassessment of technological progress, the 1962 Cuban Missile Crisis would probably make the short list.”

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