 So welcome back everybody. After lunch, I hope you all had bites to eat. We followed the lighting sessions that were just on before. So I'm Torsten, one of the organizers together with Niana. And we're going to run the session for the next, what is it? Almost 50 minutes. And I'll introduce our speakers just now. Just a few remarks. We'll have some people coming in and out. I'm going to make it reduce the number of interruptions. But if you could tweet about the session and you do have access, so you tweet regularly, just annotate at FastAsia, just so we get some publicity. And I think we've been trending the last few days, which is great for this event. Okay, so this session that's going to kick off now is Shinsa in a case study alternative to Western style innovation. I won't read too much about the kind of topic overview. Bani will take you through that. Bani will also have to explain what happened to his leg yesterday. So Bani has been around the conference for many years now, participating in FastAsia. Really a background kind of known for his hacking the Microsoft Xbox and kind of part of the open source hardware community really through his innovation of Novina, which was a DIY laptop. Bani also received his PhD in electrical engineering from MIT already in 2002. And so he lives in Singapore and runs a private product design studio, Kosagi. And so he's actively involved in a number of startups and student initiatives with the MIT media lab. So Bani, I think you're all mic'd up. Yep. Okay. And over to him. Thanks. All the best. Hello, testing. Thanks very much for coming. I probably can guarantee you I'm the only speaker on drugs at this conference. They gave me some really good opioid painkillers and the coffees to help counter the side effects of it, which make me a little bit dizzy and drowsy. So I'm going to talk a bit about Shenzhen as an alternative to the American way of innovation. I say the American way because there's a sort of notion that there's a lot of innovation that comes from America and because of the existence of that of the result, the process must be the process. And I think there are other alternative processes that I hope to sort of introduce you to some through the lens of Shenzhen in particular. So, basically start with a case study of sort of how Shenzhen came about. It started as a fishing village about 30 years ago. Now it's a tech powerhouse and the question is how did that come about? How many people here have actually been to Shenzhen? Okay, so the next few slides I'm going to try and sort of go and look at Shenzhen through several different lenses. One is Shenzhen as a place, so we'll talk about what you might see there if you haven't been there yet. I'm going to talk a bit about Shenzhen over time, like the time evolution of Shenzhen and the history of it. And then I'm going to put a couple of lenses on one is the flow of capital within Shenzhen and the ecosystem and then the flow of IP, which becomes a segue into a broader discussion about IP practices and how emergent innovation may happen in the internet age. So, for people who haven't been to Shenzhen, I'm going to do a little exercise where I call zooming in on the place. This is like if you've seen those videos, like the powers of 10 where they start at the galaxy and they go to like planets and they go to like Earth and then hands up. It's a bit like that, but starting at the sort of the territory level of Shenzhen. So, when I talk about Shenzhen, and a lot of people when they say Shenzhen, they're actually referring to the Pearl River Delta. Shenzhen itself is a specific city of an area. It's like when people say Silicon Valley, Silicon Valley is what? It's like San Jose or San Francisco, is it Fremont? It's this huge area. And so, Shenzhen has sort of been used as a generic term for the overall Pearl River Delta. And in 1980, it had a population of about 300,000 people. And in 2011, it's had 10 million people. Currently seems to be about 40 million people and half a trillion dollars GDP in 2007. That's a decade old, those numbers. So, it's like, you know, it's a big area. It's very productive. It typically includes Shenzhen. People will mean they land in Hong Kong. They say they're going to Shenzhen. They land in Guanzhou. They say they're in Shenzhen. They go to Dongguan there in Shenzhen. So, that's a little clarification of the overall area. So, at one extent, these are the stereotypes that basically everyone knows. There's a company called Foxconn there. It has millions of employees. And hundreds of billions of dollar revenue produces hundreds of millions of products per month. They make iPhones in Shenzhen and there's a communism in China, right? These are sort of like the things that everyone knows about it. And then you look at sort of the stuff that comes out of it and both of these devices are made in Shenzhen. So, the one on the right, you're probably familiar with. That's an iPhone. The one on the left looks like an iPhone, except it's much smaller and it runs Android and it costs about 40 bucks. And both of these things are coming out of that same ecosystem, right? And this sort of becomes like a little thread. You can start pulling it. Like, how does this come about? Where does it happen? I mean, it's a very easy to sort of state the platitude of, oh, they just copied a bunch of IP and they put it into this box. Except that, like, this is not really a direct copy of IP. You don't just copy something and print it and then just turn the resolution smaller and get a smaller phone. It's actually, you know, fundamentally a different piece of thing that just happens to look the same on the outside. And it's more than just iPhones and clones. So, these are just like some examples of the types of things you might see there. And that top picture, everything in that case will make a call. So all those little model cars there have phones on the inside of them. This looks like an Apple Watch and this is from a long, long time ago. And the guy who was showing this off to me was like, unlike the Apple Watch, this watch actually has a full GSM phone and can make a call. It doesn't have to be tethered to your phone to make a call. It actually has the phone in the watch. And a lot of people in Shenzhen were like, why didn't Apple put a phone in the watch? This is, like, you know, they're so behind the times, even, you know, when they're first launched. And then they have like this weird stuff like, oh, I wanted some Marlboro branded phones, so they just built it. So then if you zoom in a bit on the actual area where most of the electronics trading happens, sort of electronics market area, there's a district. This is a picture from about maybe five or ten years ago. And every building you see there is related to the trade of electronics. This is not a financial district. This is not an oil district. This is not a residential district. All of those things are sort of top to bottom devoted to the trade of electronics. About a square mile in size. If you were to walk inside one of those buildings and take a look, this is what a typical view might look like. It's a little bit like, if you imagine a wet market, so if you went to the Hawker Center in Singapore or you went to like the wet markets, instead of them selling like kilo-two noodles or like, you know, you know, hind and chicken, whatever it is, they'll specialize in resistors or they'll specialize in diodes or capacitors or like whatever it is. They all have little stalls that specialize in their own little thing. And if you were to, and then if you're, and there's probably like, you know, hundreds of these shops inside each one of those buildings. So you already get the idea of like the immense amount of diversity and the immense amount of traffic that has. And then to go into one of the shops and take a look on the inside you would, you can take a look at the amount of inventory they have. So if people aren't familiar on the image on the left they have those round things, those are called taping reels. Each one of those has about 10,000 capacitors on them. Okay. And this shop, each of those boxes has about 5 to 10 of those taping reels on the inside. So this shop has literally millions of capacitors and resistors on the inside of it. And so this is a place where you can just walk in with cash and be like, hi I have a production run of 100,000 units a day, I need to buy my inventory for it and you just give them cash over the counter, you walk out with the inventory to do that sort of production. So it's not like a thing, I have to go to Digikey or I have to do the manufacturer, I have to place PO's for a long time, you just walk in and you just get the parts and go to production, like mass scale production. And another interesting bit about it is that like a lot of these shops, if you look at the chip inventory on the inside, I was like walking around and just sort of counting the chips and you went to go to like for example Digikey or Element 14, they'll have more inventory in the shelves of those shops than all of North America in these little shops for certain chips, right. And so it's pretty remarkable the amount of stuff they have. And it turns out that like a lot of these shops are really just sort of family run businesses. Like these are people who came from, a lot of them came from one particular part of China, they all speak the same dialect and instead of you know being a fisherman or trading in ducks or live animals, they just became experts in trading resistors. And they weren't, they don't have degrees in electronics, they don't have like actual knowledge of how to use it, but they're extremely good at trading it. They know where to buy them from, they know how to sell it, they know how to negotiate the price these people, you see these people that have like maybe a thousand products available and you can just be like you know I want this one, how much for a thousand, they'll just write out the price, they don't look it up in the book, this is all they long want to do is trade these components, they're very good at it. And then you know at the end of the day you can just bulk it up and hire a guy to push it around for you in the market and go to your factory and start production. One of the other interesting things is that when you go outside the market, I call this the market zoom, you go into the streets, you'll see things like this going on. It's easy to sort of like not even look at them because you're so overwhelmed by the pretty lights and like flashing things and whatever it is. But you'll notice like, you'll notice guys sitting with these little signs and a big pile of like junk in front of them. Or you'll see people like shrouded around mobile phones laying on the sidewalk. Or you'll see a scene like this where this is right next to it, you can see there's the little pink stalls, that's kind of a hawker stall kind of thing and right next to it is like just spilling over the cornucopia of like mobile phone cases. And people are walking up there and like taking parts out of the mobile phone cases doing this sort of thing. So you see this happening a lot in the district. Because the district has become a little more gentrified, you don't see it quite as much today. But this is really an interesting phenomenon. And what they're doing is they're basically doing a lot of recycling of the parts. And if you think about why these people are so excited to do this, just think a little about the economic incentive. So the value of one of these parts is new might be ten bucks. But if you can pull it off of a board and refurbish it and make it look like new, it's a perfectly good part. Silicon doesn't really go bad. You can sell for five dollars. And the other thing is a lot of times if you go into the market, you can actually, they'll actually tell you no these are used ones, these are new ones. So they're not dodgy about at that level. If they're using you, it's usually when you get the fake parts and you're supply chain, it's someone who bought it from them and then is blending it into their supply chain size. The recyclers are pretty much up front about what they're pulling out or not. Their cost of goods is virtually zero. And the minimum wage, this is a little bit out of date, but about a thousand RMB per month in the factory, maybe now about two thousand because of inflation, it's only about three or four hundred bucks per month. But the value of the parts in that little dish there alone is like a couple thousand bucks. So it's like twelve and a half X over a minimum wage if it took them a month to pull those out of the parts. And so if you were to compute the wage parity of what you would get, like your cost of living assuming that your minimum wage of eight dollars would be the equivalent of having a 200,000 dollar a year job basically pulling junk out and upcycling it. And one of the really important things that we know is that the skill used to do the recycling is actually a skill that's trained at factories. So like I was at a factory once that was doing video cards, a bunch of envidia video cards, and there was a person there who had a pile of cards and she was pulling off DDR memory, pulling the balls off, putting new balls on and preparing them for production again because those were defective cards, right, and they're not just going to throw them away. They're going to recover all the memory and they're going to reuse it in production. So this girl was being paid probably a minimum wage to recycle parts at enormous value to the company. And what happens is these people are like, wait, why am I being paid a minimum wage to do this? Why don't I acquire those parts myself and then go into business selling parts like new because this is what they do in the company's old time. So that's why you see a lot of recycling happening in the ecosystem. And to put a little bit more of a fine point on what the gray market and fakes are, a lot of people when they think of faking they think of the thing on the left, right, is like, oh, we're going to make a fake iPhone. The thing is it's super hard to do that and it's really trivial to tell it's a fake, right? It's actually the wrong equation. You put all this effort in and a lot of people are like, I'm not going to pay as much as a real iPhone for this because it's a smaller, crappier iPhone. They expect it to be really cheap. Actually the more common type of fake is what happens on the right-hand side. This is an example of a part that I actually had bought. It's a genuine design like Spartan 6, but if you look at the chip, you see a little white rectangle that I kind of highlighted in the arrow there. What happened was that the vendor had taken a laser cutter and sort of blasted off a label on this that said ES, engineering sample, right? And so they're perfectly valid silicon from Xilinx, right? In fact, the engineering samples for this particular chip had only one small Radom versus the main product, but when Xilinx retired the engineering samples, the shubas were all supposed to grind them up and throw them away because you're not allowed to sell them into the market. Instead what they did is this particular problem, they actually were very clever about it. They went, they laser marked off the ES, so when someone checks particularly the untrained people to say, okay, we have to make sure all the numbers are there is nothing extra. And then this white rectangle, they're like, okay, whatever, right? There's just a white rectangle on it, right? And then so it gets into the factory and they only blended it in at a rate about 3% into the supply chain, right? So if you had any problems you might think that it was just a problem with your design or some defectivity or whatever it is. But the average part vendor when they're in the supply chain only makes a few percent margin. So by blending it 3% they double their take essentially out of the deal on a very, very expensive part and it's very hard to detect. This is very easy to do, but very hard to detect. This is actually the much more typical type of faking that you have in the market. So the most notorious fakes are usually derivatives of genuine goods. So it's not like someone's actually gone and copied the good, but they're actually, for example, a ghost shift. So this is when you have a factory line that produces a particular product. It's fully approved. All the workers are trained. They have all the test procedures. What they do is they run a fourth shift in the middle of the night when the manager are out and they keep the product themselves and all the margin they sell directly at the market. That's a ghost shift. But those are very, very hard to detect because they're actually, for all intents and purposes, genuine. It's just that they may not care as much to throw away the rejects. They'll blend them into the production. Refurbishing, like they talk about, remarking of parts. Warranty fraud is a very typical thing. So this is where people will say, okay you've got a phone that's got a problem. I'm going to go ahead and do something to it so I can hand it back to Apple and get a brand new phone. So essentially they can take a phone that someone would throw away and hack it a little bit and they'll trigger a code from Apple that causes an automatic, like there's these codes they put in the screaming and go in the genius bar and plug it in. They're like, oh, this is ID code. Okay, well, you're getting a new phone. So essentially it's a form of fraud that they do to try and upcycle phones that would otherwise be thrown away and then also improperly destroy production defects. So anyone who does hardware knows that you don't get 100% yield. It's stuff that doesn't work. You're supposed to destroy it and throw it away. They don't get destroyed and throw it away instead of remarked or labeled and sold on the market. So that's more typically the types of fakes that you see. So that sort of shens in as a place. I'm going to talk about shens in over time. So the snapshot you saw was probably accurate as of about like two or three years ago. The place is very dynamic, constantly changing. But in the 1980s, China had a billion people and they didn't have a shenzhen. And what happened was around then some very large manufacturers, the Foxconn and the Huawei, started setting up huge mega factories there because of cheaper labor and had almost no infrastructure. They just brought in sort of assembly laborers to build PC motherboards back in the day. It was the big thing. Hand stuffing components, whatever it is, toys, whatever. And so out of a billion people, they maybe took a percent of them and they got 10 million assembly laborers trained up to do factory assembly. And then in the 90s, they had sort of like a diaspora of medium to small factories. What happens when you have one large factory is you have other small factories that crop up around it to service a large factory. So a Foxconn facility might be assembling a PC case, but then you need someone who supplies the metal for the case. You need someone who supplies the cables. You need someone who supplies test equipment, someone who services the assembly robots. And those all become small factories near the area because it's more efficient and more competitive. And so you have a diaspora of small to medium size factories filled with hundreds of thousands of technicians and engineers. These are people who over the last decade had maybe been assembling stuff and figured out how to repair things or figure out what's going wrong. They've learned a little bit on the job. And that led to around the turn of the millennium the rise of the Shanzai. The Shanzai are the people who are sort of the copycat barons of China. And essentially the perhaps the anecdote that I heard about where they come is they're a bunch of people who are like tax engineers in companies like Motorola and Nokia, BlackBerry Rim, that sort of thing. And they were just like, we could build this so much cheaper if you guys didn't have all this stupid process. They would have all these meetings and reviews and throw away things and like guys, we can build this phone for half the price if we just did it this way. And they would get pissed off and they would leave with all the IP and they'd be like, we're just going to build phones at half the price because we know how to do this. And they would build phones at half the price and sell them and they would change one letter in the name from Nokia to a Nikia or whatever it is. And then you had these sort of phones arising. And they were pretty much direct copies back at the time. But out of those, you know, sort of Shanzai, you end up getting thousands of sort of managers and designers of shops that now have routinely produced a difficult consumer product and they know a lot about it and they're pondering, okay, well, we have all this capability. We need to differentiate ourselves. We had to do something different. Now they really are at a position where they can be like, okay, what innovation should we do? And out of those from the Shanzai around 2010, we started to see the rise of companies like Xiaomi and Tencent, Haba, Alibaba, these sorts of things. They sort of emerged out of this as the people, the group of people who were able to actually come up with an idea, innovate, create something different. And now you have your sort of dozens of major tech corps in China sort of rising up out of it. But the really interesting thing is that like at the point at which the innovation happened, it wasn't sort of top down. Someone didn't walk in and say, we're going to build like a huge factory and do it. The point at which someone wanted to innovate, they had the whole supply chain behind them to do it. They could execute very quickly and very rapidly. And so a lot, one question I oftentimes get from US pundits is like, you know, how do we bring this to Silicon Valley? Why aren't we doing this in Silicon Valley? How much money would it take? Surely you just have to take a few key players, transplant them and give them some money and things will happen. And I'm just like, you don't understand, you need to have hands that know how to do these things to build these products or not just, you know, it's not like GitHub we just clone and make, right? So another lens to look at Chenzhen through is the flow of capital. And this is another piece of the puzzle that sort of explains the phenomenon. Decades ago, someone invested million dollars in a pick and place line. And this pick and place line probably built into a Blackberry phone, right? Or whatever, your little motor flip phone back in the day. And they're doing 0402 components that cost millions of dollars. That machine ran for three years and the company decided they had to upgrade so they got rid of the equipment. But they don't throw it away. They sell it at 10% of the price, $100,000 as scrap, right? Another factory picks it up and says, okay, I'm going to use these, I'm going to build, I don't know, DVD players using this old mobile phone line. They'll use that for 10 years. Robots don't go back, right? They have essentially an infinite lifetime as long as you replace the parts that wear out. That guy says, okay, I'm done with this business. I'm going to sell this line for $10,000, right? And it gets picked up by some random ass factory building weird things, right? You wouldn't think that they would need to have a robot that can do BGA in 0402 and all this sort of stuff. But for $10,000, someone's getting that million dollar line that was producing these phones a couple of decades ago. And so what ends up happening is you have a huge amount of excess capacity that reflects the innovation and investment in the ecosystem for very cheap at the very bottom of the pyramid. And so this is a picture that I saw, I snapped, that's a pick and place line on the back of a truck. And you can see how very carefully they've packed it for transportation because obviously it was a very expensive purchase and they really cared to get there in one piece. But I mean, you know, all that being said when we got to this side, they had technicians who knew how to like tweak and tune it. It wouldn't be an issue. They're like, save the money in the box, we'll just calibrate it when it comes. That's kind of the level that they are at that ecosystem. So if you look at sort of then also the flow of parts another lens on the flow of capital, when you have a build of a million plus units, right, there's two paths that the builds can go. One is to users, which is the lower path and then the users eventually throw it away. It becomes e-waste, there's zero cogs to it, goes in recycling and as I told you about it, you can resell it for 50% original price on the gray market. The other direction it can go is that like if you have a million plus unit build, you have what I call the hot dog and buns problem except on a massive barbecue sale. If you have a barbecue you have buns, pack of ten, hot dogs, pack of eight. And so at the end of the day you're going to have two buns left over. When you do a million unit build you're going to have some hot dogs and buns left over, except that at that scale it's like 10,000 is your tailings after the run. And of course the factory doesn't just throw them away. The factory will scrap and liquidate them for about 10% of the original purchase price. And one of the other key things to keep in mind is that factories often have an aging schedule. They can actually keep parts on the line for longer than two years because they have oxidation where problems. So they have mandatory trash deadlines for the inventory if it doesn't get used up in time. So there's this constant flow like trickle down flow of scrap and excess that's coming into that gray market. People will aggregate, repackage and test that scrap and sell on the gray market. So this is another source of the components in gray market. So when people say you go into the gray market and key trusted components, well you really have to know where it's coming from. If it came from factory excess and it's a bunch of resistors it's going to be fine. Just take it. It's not a big problem. If it's like recycled memory chips, okay, you might want to be a little bit careful because splash memory does wear out over time. So those are the three lenses I showed in Shenzhen just now. But now there's a question of IP. Now that we know that they can get parts and where the parts come from, how do they go from like a tray of like parts into an actual phone. It doesn't just happen overnight. It has to be some IP added. So I went through and I tried to sort of trace the flow of IP as it happens in the ecosystem. And this is an example of like a drawing I managed to pull off the internet inside China of like one of the older phones being built. And sort of the punchline is a system I call gongkai. It's not actually the word for open source. There's actually a proper Chinese word for open source. But as everyone here knows, open source implies a certain license. You have the right to license it under a BSD license. Or you have the right to license it on a GPL license. And gongkai, you don't have the right to license it. You have it, right? And so the gongkai sort of means public. It's a difference between being like, hey, I'm going to do a porn shoot and like put my picture out on the internet because I want to. Or I happen to be like carrying it from my window and someone took a picture of me and they put pictures out on the internet. They're both on the internet, right? But one of them was intentional, the other one was not. And so, you know, the gongkai system sort of evolved in the absence of Western influence largely insulated. It has a similar impact from the standpoint of interlink access, but a very different cultural legal construct and it enabled a rhythmic spurn of Harvard that we see. But the question is how, right? How does this come about? And the reason why it's a conundrum is if you review the Western IP theory, there's a social bargain that struck between innovators and the public, right? If you were to read the U.S. Constitution it actually says that Congress shall have the power to promote the progress of science and useful arts by securing for limited times to authors and inventors the exclusive right to the respective writings and discoveries, right? And so the idea was that you wanted to reward risk-taking behavior with an exclusive right to any inventions you may come up with, right? So this is the social bargain because, you know, obviously you want to disseminate the information so other people can innovate, but you also want to make sure the person has some time to go ahead and monetize it. There's some problems with the bargain today is that limited times is now 20 years for patents. 20 years in the tech industry is an eternity. That was like from the 8088 to the Pentium III, so we went from a 5 megahertz machine to 800 megahertz machine. Might as well be forever, right? Maybe it's in perpetuity. And copyrights are in fact an eternity. Anything that you saw when you were a child growing up you will never be able to rip, mix, and burn and use on your own. You can't own your childhood anymore. It is owned by Disney and other people who will sell your childhood back to you again in the form of sequels. So the interesting thing about open source is that the open source bargain recognizes that money wealth is not the sole driver of innovation. So this one strikes a slightly different bargain. One is that maybe I want public recognition. I want attribution. Maybe I want to have a legacy. Maybe I'm just curious. Maybe I have some altruism. And so what they did is they said, okay, instead of saying like I'm going to have a limited amount of time to sort of milk the public for this, you're going to promise not to sue me and you're going to go ahead and attribute me and then you're going to share my idea with the rest of the crowd. So that's what a lot of the basic licenses look like. The problem is that Western style openness still bears a heavy legal burden. And the example I decided this is that the two open source initiatives, open source initiative and open source hardware initiative had a lawsuit over their logo because they looked too similar. And the reason why they have this lawsuit isn't because they're a bunch of pricks. It's because there's actually like legal theory that if you don't try to enforce your rights, then you lose your rights. There's actually sort of an encouragement in the law to try and actively pursue, particularly around trademarks, these types of things. And so like who in this room has money and time to just go ahead and like go into courts and like fight over stuff like this. Unless you're a lawyer in which both of these institutions had like lawyers on their board so it's not a problem for them to do that. And then there's a whole bunch of other stuff just like you have to really comply with the licenses. There's a whole like, am I GPL clean? There's a whole bunch of like if you license something, GPL will never end up in a Microsoft repo for example, but if you do BSD, Microsoft might pick it up and use it. There's a whole bunch of overhead that goes with it. So what's the impact of that legal burden? Okay, so this is a perhaps slightly skewed example, but like I like to say this is an example of what Gomkai can produce on the left and what open source produces on the right. And so these two are like a snapshot from around the same time. On the right-hand side is the Arduino UNO. A lot of people are familiar with it. It's a typical piece of open hardware as practiced by typical hackers at the time. It's a 16-bit, 8-bit 16MHz CPU with 2.5K of RAM. Has a serial interface and a voltage regulator cost about $29.1K. Around the same time if you went into the markets and you look at what the Shanzai were producing, they were building quad-band GSM boards, phone boards with the 260MHz CPU, Bluetooth OLED display, MP3 lithium-polyneum battery, $12.1K. So with these people on the left-hand side had no burden at all to acquire the IP to build that phone. And so they were able to go ahead and build that phone. Whereas on the right-hand side, the people who built the Arduino UNO had no burden to acquire the IP to build the Arduino UNO. But that's about as far as you can get. You're good luck trying to get the IP to build a phone in the Western ecosystem. You try to write through the front door. They'll literally be like, you have to pay us a quarter million dollars to even see the datasheet for a lot of these phones. So based upon this, the question is how does Gunkai survive? According to Western IP theory, the basic bargain is broken. There's no motivation to take a risk. You take a risk and it's going to get stolen from you. But the evidence is that business is booming. You have Xiaomi, you have MediaTek, all this sort of stuff. They're coming up. So clearly, either they're broken or the idea of the bargain is broken. And the evidence seems to show the opposite. So the explanation I have for this is a picture of the Galapagos. China is like a Galapagos. It's an insulated ecosystem that was able to create its own sort of norms and rules in a stable fashion that was sustainable. And so, if you were to look at sort of the changing face of innovation at internet age, on the left hand side is a picture that I pulled out from whatever the internet of what, as a classic view of innovation, this is like, I think it's called Alchemist's Stone or something like this. And it's literally a gray-bearded white guy, right, kneeling in front of the moment of entry. It's sort of the Tony Stark idea that like there's a person who discovers something big and they are the big mind, the big man who deserves the reward for doing that. And there's some like, you know, faceless assistance in the back, whatever it is. But, you know, they're incidental to the plot, right. And then, you know, but on the other hand, if you were to look at sort of the internet age, you know, people, if you were to try to do, look at what the graph of innovation looks like, you get something like this. This is a picture of like, you know, the IP address of the internet, right. You just can't put, I mean, you might be able to say, you know, Linus Torvalds did Linux, but the list of contributors that go into it is just enormous that goes into the offering system. So the difference between sort of old-style and new-style innovation is that back in the day, attribution is very important. We can pull up the patent for the zipper, number 1,219,881. And I can tell you that G Sundback is the guy who has the patent on the zipper. And this is from like, I don't know, 1917, right. This is amazing. We can attribute this guy to the dimension of the zipper. On the other hand, on the right-hand side, we have cat memes. And, you know, I've been asked in the past when I give presentations to signed forms that say that I have the right to present everything that I'm presenting, you know, I've asked the license holder and I was like, okay, I want to show some cat memes in my presentation. Who took this picture of this cat meme photo? Like, you can't find out who took the photograph. And these are all technically copyright violations. The person who takes the photo has an inherent copyright in that photograph. You're not allowed to copy that and remix it or do whatever you want. But what really matters is virality, right? You know, that person is not going to be like, oh, you can't use my photo for cute cat things, right? You know, because it's mine, I want to charge you some money for it. I think they're more excited to see their cat now, you know, trending on Reddit and being like, oh my God, my cat is trending on Reddit. And then that's it for them, right? So virality is sort of like the new currency, not so much attribution. And so the concept is that like, you can have different forms of IP protection, right? So these, like whales and bats are both mammals. They're actually both they both get birthed like young, but all sorts of stuff like we are. We're also mammals too. But they're very, they look very, very different. So it's not that you don't have IP protection in China, it's just that they're very different classes of organisms. So IP in China, the best way I can sort of summarize it, it's factory culture meets the internet, right? And so China has sort of like this tendency to sort of promote factory culture. Like this is from propaganda posters from the Communist Party celebrating the working man, right? You know, going into the factory and building stuff and working the factory is glorious, right? And then the thing that happens is that when factories are incredibly cheap and the parts for factories are easy to acquire, you end up in a situation where everyone is owning a factory, right? In this picture here, every person here is an owner of a factory. This particular facility is that old guy. This is a factory where you can go ahead and sort of do a heat press lamination of plastic bits together. So if you want to do like a business card holder or something like this, he just bought a bunch of equipment and he has a sitting around and it's basically a garage and this is his factory, right? It's easy to go ahead and start a factory. The result is that you have what I call capability over inventory. So inventory is the picture on the left. If you want a USB cable that's 1.8 meters long and you were to go to Challenger or Best Buy or whatever your retail outlet is and you go to the nice guy and you say, hey, I'd like to buy a USB cable that's 1.8 meters long. He'd be like, I'm sorry, we only have 1.5 or 2 meter cables. How about you buy the 2 meter cable and just coil the bit that you don't use and you're like, no, no, no, I need a 1.8 meter long cable. Please make one for me. The guy looks like you're crazy, right? You're like, no, I have these cables. Why don't you be reasonable and buy one? That's inventory. They try to sell you on something that you don't need because they have it in their supply chain, right? Capability, on the other hand, is the ability to produce cables. In China, there's a guy who has a machine that takes cables and USB ends and the bottom is the injection molding machine that goes in and puts the head on it and there's a number he can dial in as 1.8, enter, cut to length and he come up with a cable that's 1.8 meters long. This guy you have a conversation with and you say, hey, I want a cable that's 1.8 meters long. He's like, sure, yeah. How many you want? You want 1, you want 10,000. They'll be available tomorrow. That's capability, not inventory. When you have a factory culture you have a lot of capability. When you lose your factories, what you have is a lot of inventory at the end of the day. You have a huge massive ecosystem of factories inside China. There are factories that feed factories that go into the gray market that eventually produce the big factory at the end that you hear of the Foxconn that eventually goes to consumers. If you were to go to a maker fair in Shenzhen, it looks more like a trade show unless like a show and tell is one of the criticism people had from the West. This is like, oh, this is just a startup incubator. This is not a maker fair. This isn't real. These aren't real makers. No, no, no. This is making in China. This is the base level, entry level because everyone has a factory. The internet in China is different than the internet outside. Inside China content is very openly shared. This is a screenshot of a website, duikt8.net where they're just, you can just download whole blu-ray videos, finding Nemo, Born, Ultimatum 5, wherever it is. You just go click it and download. That's the sharing of content in China. It's a bit of the Wild West. It's not just movies. The same goes for technology. This is a website called 52rd.com and this is a website in which people trade and swap like mobile phone schematics, data sheets, tips and phones, how to build things, post questions, all that sort of stuff. There's a very sort of open ecosystem where people can go in and very on the surface, they don't have to hide. They're like, hi, I would like to know about everything about the latest Apple CPU. It's like, oh yeah, I have leaked docs from the factory and they're like, okay, cool, I'll trade you my leaked docs from this factory and then they all sort of trade and share. It's a very different ecosystem. I asked at one point, can I download a phone? I took part this phone you saw from the earlier slide and it says I want to find the plans for it. I want to make my own version. Sure enough, you can. So I go and I go on to the website and I'm able to download the data sheets, the schematics, the board layout, and even part of the board layout, these aren't Gerbers. These are live editable like PCB in a LEGO format. So it's not like I was just getting like something I couldn't modify. It's actually like if I wanted to go ahead and take this thing and add like a bunch of flashing lights on the side, it would be about five minutes of work to go ahead and add the LEDs and I could go fab it out and build it. Right? And that's very enabling. That's why you can get like all these little remixes of phones stuff because like you can get that sort of stuff. And of course you can see that like, you know, on the data sheet, MediaTek confidential resources are sent on W, whatever, like you know, it's inevitably some sort of like watermark kind of, but no one pays attention. So then that result of sort of factory culture meeting the internet is that you go from an ecosystem where virality means cat memes, it turns into one of virality means phone memes. So people instead of just at the level in the west of being able to mix pictures and put cute text on it, they can be like, I'm going to take motherboards and parts and components and other brands and trademarks and stick them together. I'm going to build a Ferrari phone or I'm going to build like a Hello Kitty literally an Apple phone that's a Hello Kitty with a Hello Kitty on it. And so if you were to unpack that ecosystem which you have is you have a bunch of small factories who do like industrial design tooling, plastics, components, circuit boards, firmware and they feed into a system integrator who then outputs the device at the end and there's a gray market below sort of like supporting all this by having a buffer inventory to do it and the role of sharing in this network is that there's a lot of people who could be your component vendor. There's a lot of people who could be tooling, a lot of people who could do plastics right? And so the biggest risk to you is not that you're not attributed for it. The biggest risk is that you're just not discovered and you don't go viral, right? And so the more barriers you put between you and being discovered is the less likely chance that you're going to survive in this ecosystem. So there is an incentive for people to share a portion of their IP. Obviously you don't put 100% of what they do out there but there's a large portion of it. Enough for you to go ahead and potentially rip mix burn and do other things with the IP. Another thing that is important in this ecosystem is that embodiment really matters. So there's this phenomenon of the hover boards that you might have been familiar with from a few years back and it was two wheel hover boards. There's a guy who has a patent on United States. His name is Shane Chen and he got it on May 27th, 2014 in 2013, right? And people are like wow, this guy invented it, right? So he should have the right to it according to Western IP theory. However, before this guy even filed the patent all this stuff in this gray box here in this middle column was happening in China. There are people with factories who could build electric bikes, who could build single wheel unicycles, who could build battery packs for like, you know, razor scooters, that sort of stuff, right? And so when people saw Justin Bieber go on stage with his hover board, right? A bunch of factories are like oh, super easy. We have the capability to do it. We already have the tooling and ability to create molds for the sort of stuff. We have the motor capability to do the motors for the wheels, the high torque low RPM motors. We have the control systems for doing self-bouncing unicycles. We have battery systems, all sort of stuff. In a matter of weeks, they're able to sort of turn these things out in huge mass volume. And so a lot of people are like, okay, who is the question in the West is who is the person who's building these self-bouncing scooters? And everyone's like, I don't know. It seems to come from all different factories. They must be really good at obfuscating themselves. Maybe they're using a Torah-like network to go ahead and obfuscate the source. No. It's not actually, it's fully distributed manufacturing. You couldn't shut it down because you would hit one and there's still 99 others surviving at the end of the day. And because there was this impression obviously from the West is that someone must have invested in the factory. They're like, it's amazing someone built a factory in two weeks that can build these things as if they went from like bare earth, poured concrete, raised the buildings, hired people, put the machines in and built the factory in two weeks. The only way it happens is these guys already knew how to do it. They are very, very focused on embodying practical inventions and turning them into practice. And when you have that sort of capability, you can go from idea to invention very, very quickly. So the net result is that it's products over paths at the end of the day. It will take you three months and $20,000 to negotiate even a simple license with lawyers. Or it could just take cash and carry and take the product. Who's going to get the profit faster? The guy who takes cash and carry over product. So it's a reinforcement cycle. The guy who goes ahead and manages to sell product before the guy is worried about getting all his trademark filings in order is building a factory. And that guy produces more product even sooner than the guy who's just signing the license deal goes ahead and makes more money. And then by that point, he's on B2 the product already innovating on the next thing where the first guy is just finally getting his VC money to go ahead and build a factory. So it's a natural reinforcement cycle that happens that really accelerates and gets people focused on products over pattern. And then you end up with a bit of emergent innovation as a result from a very diverse ecosystem. You get stuff like these wacky little phones. You end up with Xiaomi. You end up with stuff like Alipay, which in 2014, across the $770 billion in transactions, Apple Pay only did $11 billion. And this Alipay wallet, if you read the Chinese there, it says we're an open platform for innovation for payments in that ecosystem. So they're really starting to embrace the idea that this openness allows technology and virality to an adoption at the user level. And so what you end up with is that, yes, you have IP like in the West. There's also IP in China. There are different practices of IP. They're both IP protection and they both work in their own way. They're just, they grew up in different environments. The Western IP environment was pre-internet and China IP is post-internet if you think about it. So China itself came of age in the 80s. They were building the computers that eventually built the internet that they eventually used to share their IP on. Whereas the Western IP practices were drafted in Venice in 1700 or something like this. And so it's a very ancient system that we still use and we think is the system they use. It hasn't been calibrated at all for the internet today. So if you're to compare the classic Western IP ecosystem, it does encourage high risk basic research. So if you need to do, for example, a cancer drug or drug discovery, it's 20 years of hard work and then maybe it comes out the end of the day. You're going to want the IP protection. It has a place. There's a purpose for that type of IP protection today even. There are certain processes where it's very, very reasonable. You can give you predictable, defensible earning models which allows you to collect investors. You can say like, look, I'm going to have a monopoly for 20 years and I'm going to milk it and I'm going to make a bunch of money and you're robbing them rich. So give me a bunch of money now. That's the VC pitch. And it's also a very methodical form of invention. You have named inventors. You have a pedigree. You can sort of point to where things came from at the end of the day. On the other hand, innovation in the internet age is an agile, robust network. So it's not just a single person. It's a network of people. And this network of people can handle systemic complexity very easily. So very very complex things, even like the hoverboard can arise overnight because you have a network of people who can collaborate together to create innovation at the end of the day. It does mean there's a constant churn. It's very highly competitive. You're not going to be like, yeah, I'm going to have this factory 20 years from now or I'm going to retire on this wherever it is. You've got to love what you do. You've got to keep working at it to survive in that ecosystem. But a lot of times these people, it's not sort of, in the western model, you're like, we're going to go for an IPO and that's like your cash. Where your stock turns into cash. On the sort of more hardware-focused internet model, it's the day in which you start selling product the first time. Usually your product sales are very high at the beginning and they kind of tail off over end. So you actually, a lot of people who own factories, get the cash out a lot faster than people who necessarily are investing their time and effort in building whatever like your latest social media company that needs a scale of 10 million people to even begin to monetize with ads. And the other thing is that innovation is sort of an emergent network property. So a lot of people are like, what innovation in China? Show me the Tony Stark. I mean there's Jack Ma on those kind of guys. But you don't have Steve Jobs and Reed Hoffman and all these people who are Elon Musk, the named inventors, the heroes and that sort of stuff. A lot of it in China is very anonymous. People just, they're happy to get the virality out there and to make the money and then whatever. At that point in time, we've got our money. We're going to move on and do our next thing. So innovation tends to come more as an emergent network property. And the fact that it's emergent from the network doesn't mean it's not there. The fact that I can't point to the person's innovative or I can't point to a single region or whatever it is as the center of innovation, right? And just say this whole area is innovative doesn't mean it's not innovative. It means it's just a different kind of network based innovation. So the opportunity at the end of the day here is, the way I look at it is how actually can we create a hybrid between the two styles of innovation? Is it possible to, for example, take advantage of this emergent network of innovation, sort of combine it with a bit more of the risk-taking and sort of named inventor style and create products that then end up going to consumers? So a lot of the work that I do as a sort of freelancer in manufacturing is basically like small design vignettes. Like what is the smallest run I can do in do injection molding and not lose money? That's a question I want to answer. What's the smallest run I can do of a laptop and not lose money? What's the most crazy custom thing I could get built out of this factory? So I do a lot of these weird sort of almost, you know, quixotic sort of projects but a lot of them are just sort of trying to find different facets of like what are the parameters we can plug into innovation models and actually utilize and then turn into sort of an impedance-matched form of innovation that you can market and sell into the western ecosystem. And so that's my talk. Thanks. Great. Very insightful, Bani. A lot of photos too. Any questions from the audience about machines and supply chain? Thank you very much for the talk. It's usually very interesting. I've got a question. There's things that you describe would it work only for consumer-facing products or that style of innovation applicable for any aspect of software development or biology or other things? Yeah, right. Can this network style emergent innovation help? I mean, I think the network style emergent innovation is exactly what open source is in software today. You have that already and people are willing to accept that because there's no cost of goods in this solid attribution. I think in, for example, in areas like biology, it can help. There are some examples of it helping but it also can hurt because biology is actually a really difficult field. Reproducibility is like the number one enemy. You do an experiment, you get a good result, the next day you don't get the same result. If you have a supply chain with fake reagents and bits and pieces going through it, you're tearing your hair out trying to do basic experiments. There's a fine line in there but one of the things is you can get reagents much cheaper in China. There's a balance between how much do you go towards a populist model of innovation versus an institutional model of innovation versus the difficulty of the research that you're trying to approach. Most of the stuff you see is people taking components and doing new things with components. They're all already baked goods, the fundamental components. I've been recently a lot of figures involved with the RISC-5 project and they're using like chisel as a means of creating CPUs and I have an idea of like a stack processor or what. At what point in time or do you think they're going to be to the point where I could take a design that I did as FPGA and get an ASIC that's actually going to run competitive to something that I would go through in their chip fab. I think we're basically at that point. We're at that inflection. There's an inflection point we're hitting very soon. It's a combination of factors. Moore's Law slowed down. Even like the top-end chips are still taping out 14 nanometers today. We would expect it to be like 9 by this point or wherever it is if it was keeping up. The slowdown on Moore's Law means that a lot of the fab capacity that is accessible to mere mortals and smaller companies is now coming down in price on the same curve that Moore's Law predicted but now you're basically running the same fab that Intel or IBM might be running a Power9 or an X86 in level. Probably not fortune but maybe 28 nanometers. Still good enough. That's one thing. The other thing is yes, that sort of fab capacity is coming online and dressing those markets. You're starting to see more and more, there are a lot of Chinese companies that do ASICs like Huawei and Rockchips and all those guys who are turning these things out either based on arm or wrist or wrist 5, whatever it is. I think it's more just that ecosystem trying to organize itself. Once a business model is found that produces money people are going to jump all over and it's going to really explode. So I think you will sort of see that level of innovation taking off and there's also another whole section of innovation in the ASIC space which is not around SOCs and the high end stuff but there's actually a huge untapped market at the bottom. So sort of like 250 nanometer, 180 nanometer, like oldish silicon but the wind there is condensing a whole bunch of analog discrete IP into single chips and really making things tiny, tiny, tiny. So completely in your earphones they have a lot more processing pills that you can swallow that do tracking of your intestines and that sort of stuff and telemetry of packages through RFIDs that sort of thing. These all require that level of silicon integration that was very inaccessible before. I think you'll sort of see a second boom of the echo wave. So the leading wave is sort of the high end processors and then there'll be like another wave behind it of people sort of like okay well we have smaller and more niche products that are interesting to do and they're sort of riding this other wave behind. Any other interesting questions to this topic? I can just ask something you've just mentioned in your answer now is that a lot of this is commercially driven and so perhaps one of the points to raise is in the west a lot of the innovation is altruistic, some sort of new creative designs and sort of thoughts to that in this Shenzhen model. Yeah I mean there are people who are kind of sort of otaku-ish. They just build stuff and they innovate and they don't really care they just sort of keep drinks about. But like you know in every sort of organization there's like I had a boss who once said like in every company there's a division of 200 people there's that one guy in the back room who like no one knows he never talks to anybody but he like produces like 90% of the actual innovation in that company right? And so I think the ecosystem does have pockets of those guys and they don't even care like they're so far from the attribution layer they're so fundamental to a bunch of stuff so that does happen but I think that's just a there's just certain human beings who are like that who really get into something they lock in and they and I think that's the nice invariant is that happens regardless of the ecosystem. You see them in Japan, you see them in China, you see them in Europe and the United States. I've seen them everywhere. But overall I do think that particularly when you talk about hardware there is fundamentally a commercial bent to it. In order to build the hardware you have to spend some money for the atoms. No one will give you a pile of atoms for free. They don't even give you water for free. So at the end of the day fundamentally hardware innovation is tied to a commercial aspect which is and that's one of the big differences between like when you talk about open source software, open source hardware, a lot of people in open source software don't wrap the heads around the fact that money and liability are actual things you have to deal with in hardware. Good, I mean just a round of applause again for that money.