The Essential Guide to Electronics in Shenzhen is a new book for non-Mandarin speakers going to Shenzhen trying to navigate the electronics world. It has technical terms, maps, and other useful hints for a newcomer.
The book is authored by Andrew "bunnie" Huang, an MIT-trained electrical engineer from Kalamazoo, Michigan. He and his business partner and friend, Sean "xobs" Cross, are the team behind the open source computing platform, Novena.
I reached out to ask bunnie and xobs a few questions regarding the book, Shenzhen, and the state of technology.
How did you get into the hacking field?
bunnie: Growing up, I always took apart electronics around the home. My natural curiosity to figure out how things worked eventually lead me down the path that some people call hacking.
xobs: I was very fortunate in that my parents had a Commodore 64 when I was in kindergarten. As a result, I was the 5-year-old helping the teachers type 'LOAD "*",8,1' into the classroom machines, because it was their job as teachers to teach and my job as a 5-year-old to play computer games. This set up a recurring theme in my mind that it's really interesting to be able to do things, even if nobody else thinks so.
The concept of hardware hacking became clear to me in high school when I built a "DAC" out of resistors. Nowadays, every device we own can play sound, but it used to be that sound cards were very expensive, assuming your computer even had an add-on port. The laptops I could scavenge had no expansion ports, so I built a "sound card" by creating a resistive ladder on the printer port. By tying together eight resistors of increasing size, I could make something that sounded somewhat decent.
xobs (left) and bunnie (right). Photo credit: Make Magazine via Crowd Supply.
Was hacking something you always knew you wanted to do or did someone or something push you in that direction?
bunnie: See above.
xobs: I was always interested in how things worked, and so it was more of a natural extension. I never knew how to answer the question "What do you want to be when you grow up?", but I'm very happy with what I'm doing now.
What areas of engineering do you focus on? Hardware? Software? Both?
bunnie: I focus primarily on hardware. xobs focuses primarily on software.
xobs: Generally, I focus on software, but it's important to not pigeonhole yourself. I don't like the phrase "That's not in my job description". On my desk, there are several machines for development: a microscope, a rework station, an oscilloscope, and a 3D printer. I may not be very good at designing 3D models yet, and I'm really good at accidentally making solder bridges, but I'll get there.
Why do you think Shenzhen has become the place to be for startups and tech companies?
bunnie: I'm not quite sure if I'd characterize it as "the place to be". It definitely has strengths for portions of startups that have a hardware component to them. The primary reason foreigners send engineers to Shenzhen is if they have something electronics hardware-related they need to build. There is a rich ecosystem of suppliers and talent for accomplishing most things related to electronics hardware.
The poor Internet connectivity, regulatory barriers, and language barriers make it a challenge to operate in, as well. Hence, it's not necessarily "the place to be" in a global context. It is, however, "the place to be" right now if you're Chinese looking to start a tech company in China.
xobs: The tautology is that Shenzhen is a very important place to be for hardware startups because that's where the hardware startups are.
Junkyards are fascinating places. They're fun to go to because there are all sorts of nifty little gadgets that have been thrown away, and they're yours for a very cheap price. You can visit a junkyard and pick out any number of motors, frames, wires, and lights, and end up building a really cool robot or car or coffee table. There used to be TV shows centered around this idea.
Shenzhen is like that, but with electronics and at industry-scale. When a product reaches its end-of-life, the surplus stock generally goes to the markets in Shenzhen. Sometimes the parts are sent there by the factory to try and recoup on failed payments. As a result, Shenzhen is a massive, city-wide junkyard full of electronics. Much of these electronics are still brand-new and available for cheap. And they can have your parts to you within the hour.
"Shenzhen is like a junkyard, but with electronics and at industry-scale." —xobs
Do you think that the growth in Shenzhen is sustainable? Why or why not?
bunnie: No growth curve is ever sustainable. You'll eventually hit some limiting factor—space, energy, manpower, financing, etc... The question is whether it will plateau or if it will crash when the limits start to kick in.
xobs: I do not think the growth is sustainable. The heart of the electronics market is Huaqiang Bei, which used to be a nondescript corner with anonymous buildings everywhere. Shops advertised the chips they'd sell, and there was an entire building called the "LED Building".
If you go there today, you will see quadcopters everywhere, hear loud music with announcers selling Hoverboards, and even find people selling fancy boxes of tea. It is increasingly becoming an area of pop-culture, with finished products available for sale. There are still markets that sell parts, but they are spreading out from the core area.
The worrying thing is that this trend has happened before in Asia. Today, Akihabara in Tokyo is famous for being an area of pop-culture, but it used to be a place you would go to get parts for your old radios or to find that one vintage SGI workstation part. Over time, they moved away from parts and towards finished goods, where margin and volume are higher and there are more buyers. I see Huaqiang Bei moving the same way.
The difference is that manufacturing still happens in China, and so the surplus parts will always be there. Many bulk-sales are now performed online on sites such as Taobao. So while I don't think the growth is sustainable in the city of Shenzhen, I think that overall growth will continue because those resources are still available— they're just decentralized. And they still have one-hour delivery.
Are there other areas in China or the rest of the world that you think might be able to be the next tech hub?
bunnie: There's a lot of other places in the world vying for that title, and many will get there. I actually work and live in Singapore, and commute to Shenzhen on a monthly basis. I think Singapore is a pretty good hub for our approach to tech.
There are spots all over Europe, Asia, and the US that are sprouting tech hubs in various specialities. Boston is a biotech hub; there's a resurgence of tech manufacturing in central EU; China is investing heavily in upgrading infrastructure and attracting investors into its central regions.
xobs: Shenzhen was a perfect storm of tax-free area, inexpensive land costs, a large amount of manufacturing taking place nearby, and easy access to a Westernized area. There are few other places that tick all of the boxes.
I see some places in Thailand that could fit the bill. And perhaps Myanmar, in a few decades. A tech hub may appear somewhere else in China, but I don't forsee a hardware hub forming in Europe or North America. They can't support the low-margin surplus markets, and it's usually cheaper to throw away unused goods than it is to resell them.
What was the motivation behind writing this book? Did someone suggest it as being something that would be useful to non-Mandarin speakers?
bunnie: Well, I'm terrible at Mandarin and I'm always struggling to specify parts in the market. Part of the challenge is that the Chinese can't even agree on translations for many technical terms. So, initially, I thought to write the book for myself. But I figured if I'm going to go through the effort, I might as well add some illustrations and a cover and make it available for others to use!
Image courtesy of Adafruit.
How long did it take you to make the transformation from a concept to a full book?
bunnie: I started writing the book in mid-August 2015. It was crowd-funded in March 2016, and delivered by June 2016.
To be fair, I had been collecting translations of various terms some time before I started writing the book proper.
How do you decide what your next product will be? Is it based on a need you see or do you get suggestions?
bunnie: xobs and I have a tradition of going for a long beer lunch every Friday. I'd say many of our product ideas come out of those sessions, discussing what we think might be interesting or not.
Some of my other product ideas come out of observing a need that simply isn't met in the market, and others stem from contracts others give us, so that's like getting a suggestion in a way.
xobs: The next product always scratches an itch. Whether it's wanting to build our own laptop, wanting to print a book, or wanting to learn how USB works, each product requires we learn a lot. And little by little, over time, we build up a repertoire of both technology and knowledge, as well as personal relationships with the factory owners.
The closer you get to the source of a thing, the easier it is to change it. If you're at home cooking and you want to try adding an exotic ingredient to a cake, it's easy to do so. If you're ordering a cake from a shop, it's more difficult to do so, but perhaps still possible. At a restaurant, you may be able to do so if you call ahead early. If you're flying in an airplane and they serve you cake, it's impossible to add any ingredients. Unless you can work with the company that produces the meals.
In Shenzhen, with the factories nearby, it's very easy to get access to the factories, and some of them will even entertain your strange requests. Entire product lines can be made by changing one small parameter in the factory. The trick is to find a new process that's interesting, make a change, and see if it is compelling enough to make a whole product around it.
What’s the next step you see for technology? Is there going to be a shift toward hardware or software?
bunnie: That's a hard question. I think both will continue to be very important. I think as hardware continues to commoditize, there will be a premium market for bespoke, purpose-built hardware. That's sort of what "IoT" has come to mean these days.
But in all cases, the most interesting hardware requires a lot of firmware to make it do anything interesting. A typical ratio of hardware to firmware engineers in a big corporation is about 1:20, or more. For startups, it tends to be closer to 1:1 but they are a special case, and often times that ratio imbalance is made up for by shipping hardware initially with shoddy firmware that continues to be improved and upgraded for years.
In the long term, though, I see biotech as being the thing to watch. I personally consider biotech and hardware to be closely related disciplines. So through that lens, maybe it's a shift toward hardware— but probably not as you intended in the original framing of the question.
"I see biotech as being the thing to watch." —bunnie
xobs: The big advantage of software is that it's easy to iterate. You can make changes and rebuild the program to validate those changes, and the entire process is very quick. In hardware, a "rebuild" is very expensive and takes weeks. This is why simulation and FPGAs are huge, because they knock the complexity of hardware and chip design down into the software domain, where a rebuild becomes much easier.
With the slowing of Moore's Law, the variety of hardware available will stop changing so rapidly. As a result, an ecosystem of readily-available parts will spring up, enabling hackers and companies alike to create interesting projects by simply wiring parts together. Already we see this in the Arduino community, which is built on technology from the last millennium. Many products coming out of startups began life as Arduino prototypes by stringing together modules, allowing for rapid development of hardware.
Having said that, software will always be easier simply because it's largely intangible. With the proliferation of computing devices, it's much easier to write "Hello, world!" than it is to do the hardware-equivalent—blinking an LED.
Hardware modules will make for interesting projects, but they will all be driven by software.
Huge thanks to them both for their time and thoughtful replies!