Anna Rose (00:00:05):
Welcome to Zero Knowledge. I'm your host, Anna Rose. In this podcast, we will be exploring the latest in zero knowledge research and the decentralized web, as well as new paradigms that promise to change the way we interact and transact online.

(00:00:27):
This week, Tarun and I chat with Pramod Viswanath, Professor of Engineering at Princeton, and the co-founder of Witness Chain. We talk about his early interests that led him to work on wireless, the early wireless industry, and how research was brought to market at that time. We then talk about the introduction of adversarial thinking in his research, then move over to blockchain, how the incentives of blockchain can be used within the wireless industry, and his new project Witness Chain, and how they aim to reshape wireless. Now, before we start in, I wanna share that one of our ZK Summit partners, Aleo, are currently in their third incentivized testnet phase. Aleo allows for programmable privacy and with their testnet developers can now build their own private and scalable applications. If you wanna try it out, check out their repo at github.com/aleohq. I also wanna let you know that there's a new ZK Hack multi-week event coming up, starting on November 22nd. This is our third edition, and again, we'll be bringing you a ZK puzzle hacking competition, as well as a number of workshops with the best teams in the space. Be sure to sign up for a kickoff event. I'll add the link to that in the show notes. Now Tanya will share a little bit about this week's sponsor.

Tanya (00:01:35):
Today's episode is sponsored by Polygon, introducing Polygon, zkEVM. We all know that Ethereum needs to scale, and Polygon believes that zero knowledge tech is the best way forward. Polygon's vision for zkEVM is simple. Developers can seamlessly deploy any Ethereum smart contract to a layer two and benefit from the scaling power of zk proofs. It's also permissionless, meaning anyone can use it, and open source. Polygon zkEVM was built by Polygon, but it's for anyone and everyone who wants a cheaper and faster way to use Ethereum without sacrificing security or decentralization. Public testnet is now available, so you now have an opportunity to test their work and make improvements. Join them for this journey. If you'd like to learn more about Polygon zkEVM and stay up to date on the latest, go to bit.ly/startonzkevm and check out the Getting Started guide. Again, that's bit.ly/startonzkevm. So thanks again, Polygon. And now here's our episode.

Anna Rose (00:02:33):
Today, Tarun and I are here with Pramod Viswanath, Professor of Engineering at Princeton, and co-founder of Witness Chain. Welcome to the show.

Pramod Viswanath (00:02:41):
Thank you, Anna. Happy to be here.

Anna Rose (00:02:44):
So we've had two of your kind of associate colleagues on the show already, David Tse and Sreeram Kannan. I'm gonna add links to these episodes in the show notes. I think one question that starts here is like, because each of you mentioned the other, and I'm very curious, like who nerd sniped who?

Pramod Viswanath (00:03:03):
You know, this blockchain is a chain of blocks and there's a chain here too. I was David Tse's very first PhD student. He was my co PhD advisor and Sreeram Kannan was my PhD advisee.

Anna Rose (00:03:15):
Nice

Pramod Viswanath (00:03:17):
And I'm sort of the missing link between the two and I also had some role in encouraging both of them to join this whole research saga and adventure on blockchains.

Anna Rose (00:03:26):
You did? Okay and I saw, I mean, a lot of the work we actually mentioned in those two episodes, Prism, Blockchain is a Race and Nakamoto Always Wins. Like these papers, these works, you're also co-authors on a lot of these. So is this like, I just called you colleagues or associates, but what are you actually, because you're all at different universities too, right? Like it's not always.

Pramod Viswanath (00:03:46):
Yes, but we talk probably three times a day every day for the last many years. My daughter's very first English word was 'Day' for David because Dad was talking to David like every day.

Anna Rose (00:03:59):
Nice. That's cool. Would you say it's in those conversations that a lot of these ideas are being shared and sort of the the origins for these works are happening?

Pramod Viswanath (00:04:08):
Indeed, you know, research we become, even though we are professors and they're students, it sounds like a hierarchy, but at a PhD and a graduate level, we are really friends and we are very, very close to each other and just sheer number of hours spent, thousands of hours over a four-five year period and you get to know each other at a very close level, especially intellectually how you think and that stays on over a lifetime. And it's really a privilege to be part of that.

Anna Rose (00:04:40):
Cool. Let's start kind of at the beginning of your research and I mean, I know that wireless P2P networks, this is what we're gonna talk about for this first part of the interview. The work that you were doing on that kind of in the lead up to this though, like what were you interested in that led you to work on that field in the first place? What was it? Was it the math? Was it the engineering? Was it the challenge? Was it because it was like relevant at the time or is it something that you were very interested in from the start?

Pramod Viswanath (00:05:13):
I mean, as a kid, as a a typical sort of a student who's interested in math and science and engineering, you're interested in what the things that Hollywood promotes, you know, AI, science tech, and I was interested. I came to Berkeley in Computer Science as a graduate student to do AI and Machine Learning. It was a bit early for its time. This is 1995, it was at the cusp just before the real AI boom took off and maybe seven to eight years too early. But I've always been interested in trying to bring tech to as many people and improve the human condition. It was a bit early. The AI sounded a bit too when I actually started reading and get to grad school, I felt this was not right there whereas wireless was just happening.

(00:05:59):
The internet was happening not so much wireless. The internet was happening and it was very clear. The first search engine Inktomi was my lab mate and I could see the internet just growing around me. So I wanted to be part of this revolution and Berkeley allowed me to pick any advisor I like, the first year is open and I took that opportunity to explore what I would like to do, and I spent my PhD years just trying to understand how to efficiently bring communication to the masses, really wireless communication?

Anna Rose (00:06:31):
So is this idea of like trying to actually get it into the hands of many, but at the time, was there, was it mobile or was it like, what was wireless?

Pramod Viswanath (00:06:40):
Wireless was a old thing, right? A hundred plus years and you know, Marconi and Transatlantic Wire from the US to the UK and so well over a hundred years. So it's a long thing, very old history, but 95 / 96, it was clear that there was a time where the technology was growing and phones were available. It's just that they were very clunky and not the greatest of all coverage everywhere and definitely not Manum and child, but it was clear that this was as an important role in bringing people together.

Anna Rose (00:07:16):
Did you think at the time, like was it already clear that there's going to be like cellphones in every hand very soon? Or was it sort of still a question mark?

Pramod Viswanath (00:07:25):
I was just a student, maybe you, I should just say I was a nerd and I didn't think of these broader questions but it was very clear even to my, you know, when I was explaining my research to my grandmother, it was very clear what this would entail. So didn't need to convince many people what this would, if phones got to everybody, it would what it would be. But the vision was there that we want to bring this to everyone and there's one more vision, Anna, which was that phones were just used to make phone calls like a telephone, you know, phone but we really wanted to merge the internet, which is a network which was taking off around us with another network, which is the wireless network and these two networks were as disparate as can be. They were disparate technologically. They were disparate, like literally separate sets of things aand there was also, diseparate in terms of applications.

Anna Rose (00:08:18):
Was it different companies running each of those as well?

Pramod Viswanath (00:08:20):
Entirely, Right.

Anna Rose (00:08:22):
Fully different industries

Pramod Viswanath (00:08:23):
Decentralized. Yeah, totally different. The internet was sort of decentralized. It was up and coming, you know, America, AOL and the company Yahoo. This was way before Google, right? So this is the early days of the internet and basically .com and well it had a lot more decentralized aspect of kind of a feel of what today we see in crypto and wireless was very old large companies. I mean, the, the biggest gorilla of them all was AT&T and it was so big that they got broken up into two parts AT&T and, and Lucent. And they had these big operations where they would literally draw wires all over the country and you know, tell AT&T telephone. So it was they were very centralized versus decentralized, very hierarchical and physical networks compared to packets and switching on the internet. So they were very different.

Tarun (00:09:21):
So, so I think one thing that maybe the listeners of this podcast might not know a ton about is sort of the, also the fight that sort of took place over buying spectrum in that era and like, sort of how the commerce side of wireless interacted with the technological side. So I guess, you know, how would you kind of describe the kind of evolution of the wireless industry? Because, you know, it sort of started as car phones and very like low energy radio things to like, hey, now we have people competing in spectrum auctions and people buying other companies just for the spectrum. And how did that like impact the kind of research? Like, was there sort of a feedback loop between what research was done and sort of what new technologies were people were willing to kind of sponsor? Or how, how did that go about? Was it the research that certain technologies to be sponsored or?

Pramod Viswanath (00:10:17):
Yeah, thanks, Tarun. I mean, it's a complicated question and I was there at that time. So the issue, I would say the main way to characterize the state of affairs at that time was that it's a very permissioned system. You know, to use a blockchain language, very permissioned. You have to be, it's a country club. There are only a few people, and you have to be in the room literally to be part of any conversation, whether that conversation meant on the business side, buying spectrum or on a technology side. you know, actually impacting what kind of protocols and what algorithms and what tech goes over. It was a very low risk appetite, very permissioned country club, I think a feeling to the whole industry, to be fair, it's a complicated industry. It remains so because it's capital intensive, you wanna cover the whole earth, really, like literally cover it, right?

(00:11:04):
You have to be everywhere. You have to put access points. It has also got hardware aspects to it. It has antennas and you alluded to spectrum, which is regulated by, you know, governments. So it had multiple aspects to it nd at the end it became very risk averse and very permissioned and the internet on the other hand was so open and the wires had been done for decades earlier right. During the DARPA project, and the setups were there and the intelligence had just come to the computers and the browser was just born at that time, early to mid 90s. And now that you have a browser, you can build applications on it and e-commerce was an obvious application.

Anna Rose (00:11:45):
When you were doing this, like the wireless research, were you actually designing new forms that these networks could take, but like, probably wouldn't take at the time because of this incumbency, because of the like capital intensive nature of the industry? Was it sort of like experiments that you were putting out there, but not necessarily getting to see them happen in real life to see them become alive.

Pramod Viswanath (00:12:10):
That's entirley true. We wanted to make the protocol simpler, make the hardware simpler, just de-risk the business in a way from now that I look back, I wasn't so much a business person, but not have these vertically integrated stacks and spread them apart and now in the hindsight, I can look back, that's because that was what the internet was. Every individual person could take their own relatively low risk, an easy way to get onto the wagon and experiment and explore and see where the market takes you whereas wireless was so discovered. So research really was about how to build decentralized or peer-to-peer was also a big thing. BitTorrent had come around that time, that's again on the wire side. And we were trying to understand what the implications would be. And how would you, in any practical way, bring such way of thinking?

(00:13:02):
I wouldn't see the same technology, but a way of thinking to wireless, which is so vertically integrated and, you know, everything is controlled together. So in other words, all the pieces work. For example, I'll give you an example, Anna. So and what I mean by they're all completely tightly centralized, controlled, right? You have a phone, you're moving around, you go from one place to another, and you will go from the view of one base station, which is where you're talking to, to another one, simply because you moved and this could be, let's say a couple of miles. The, those base stations had to be talking to each other in a closed way, so you could hand over the conversation from one end to the other end as you move. And this was a basic requirement of phone mobility, mobile phones. Right?

(00:13:44):
That's what it means to be mobile. I mean, if you're static, then you're a phone. It was on the wall. So the whole point was to allow mobility. And if you make a very centralized decision where all the base stations are coupled together, they're synchronized to milliseconds or even hundreds of microseconds, the crystals are so clear, they're all synced, and then the signal can move from one to another without breaking your call and that was how the designs were and so they were very careful controlled electronics together with the protocols. The internet was completely opposite. It was packets, packets would get dropped. They would be resent. It was really a free for all, wireless was the opposite side.

Anna Rose (00:14:27):
Interesting. Do you feel though, over time, did wireless move more towards the internet model, or was it more that tech got good enough that you could use that sort of perfect model better?

Pramod Viswanath (00:14:39):
The market forces were clear that you should move towards decentralized and it happened in two different ways, and it's a little bit subtle and I'll come back to this later in this in our discussions, but the tech part was not there and that is something that we were working on. And in 1999, I did an internship at Lucent at the labs, and I met a great engineer scientist. His name is Rajiv Laroia. We get along, we got long spectacularly. It was a moment where I had spent time thinking of how to build such decentralized wireless networks, how to make them efficient, what would be ways in which people can join asynchronously and still be connected and not got dropped off together with Rajiv Laroia's vision of how to build a whole system around it that can scale in a global way and connect to the internet efficiently. And so that was a spectacular summer of 1999. It was also the peak of the .com and Rajiv was at the cusp of saying that we should take this technology out of the labs and commercialize it.

Anna Rose (00:15:48):
Oh, but did you do it?

Pramod Viswanath (00:15:49):
Yeah. So Rajiv Laroia started this in late '99, and I started off joined them in early 2000 just as I finished my PhD. And this company was called, well it was called Radio Router. Oh, you know, it's a radio, but the router, so really want to be both, radio as a phone and router as a core component of the internet. So, and in fact, our tagline was 'Every Bit Wireless'.

Anna Rose (00:16:17):
Interesting. What could you do at that time? What was actually possible?

Pramod Viswanath (00:16:20):
The phones were still clunky, so there was just a few phone companies, and the phones were very clunky. This and in our pitch slides where things like, imagine you can do email on your phone and so people had to imagine that but it was clear that email was there. MapQuest, I mean, Google Maps weren't there, or Apple Maps, but MapQuest was there. So people would download maps from the online and you can print it and so it was changing many things that people were used to from a physical world to the cyber world. So they could imagine it, but the hardware was not there, but we were doing the algorithms and the tech underlying it.

Anna Rose (00:16:59):
Did you, so this was your first foray outside of academia, I guess, was this?

Pramod Viswanath (00:17:03):
That's right. This is my first job right out of PhD.

Anna Rose (00:17:05):
Okay.

Pramod Viswanath (00:17:06):
Sort of first foray. That's right.

Anna Rose (00:17:07):
But this wasn't your company, this wasn't, you weren't found co-founding this, you were joining someone.

Pramod Viswanath (00:17:12):
Well, maybe I was employee too or very early. I was also working with them beforehand. Yeah, I learned a lot through this experience of being in the industry and being an entrepreneur and,

Anna Rose (00:17:24):
Cool.

Tarun (00:17:25):
Sorry, what year? We're now just after the .com bubble?

Anna Rose (00:17:28):
We're in '99

Pramod Viswanath (00:17:29):
It was still when we raised the money, and we still did it, it was still the very peak of .com March, 2000, and then September came and the whole thing crashed. And 2001 was really bad because there that's when the communication crash happened. And people remember the .com crash of 2000, but 2001 to 2002 is where all the communication companies went belly up. These are, you know, like the Titanics of the era, Global Crossing had put fibres all over the oceans and connected the continents really and they went belly up. I mean, they had no business, most of the, there was no data flowing. There was no YouTube. Yeah and all these fibres were dark and they went belly up and sold on pennies on the dollar. Really fire sale at that time. So it was a bad time to be in communications and networking those years.

Anna Rose (00:18:17):
That's interesting. It was like the rails, they were setting the rails in place for something that didn't exist yet. This sounds

Pramod Viswanath (00:18:23):
That's right.

Anna Rose (00:18:23):
Oddly familiar to our industry. But anyway

Tarun (00:18:29):
Well, I think, I think the funny thing is like, didn't you write your kind of wireless communications textbook around that time? Like, it's funny if I look at the timing, it's like, right, you know, like right at the bottom of the market was the perfect time to write a textbook

Pramod Viswanath (00:18:43):
Because that's when I felt I think we had done the research right during my PhD and David Tse's research, he had just gotten tenure. And here I was, I had come from this startup experience being in the real world, how to design protocols, which algorithms, which research makes practical sense, which ones go together and also they have to be, this is another thing that has a commonality to blockchain, is that wireless at the end has to be very robust and very simple in two ways. First, it has to scale globally. Like man, woman, child, everybody will hold a phone and you can't be worried about interference. So you have to handle it at a core level in a robust level. The other thing about wireless is that these are gigahertz radio signals. So every leaf of every tree will bounce off the signals in different ways. So if you are not designing for that in a robust way and upfront, you'll never get there. And I got you trained as an engineer, as a system engineer to think about robustness and scalability. These two points you know, very, very it got into the blood basically, you think in, becomes instant that any protocol, any algorithm that you think about any tech, does it pass robustness and scalability? And this came in handy when we did blockchains.

Anna Rose (00:20:00):
Before we go into that, I actually, that book the Fundamentals of Wireless Communication, I learned in the interview with David that this was like the book that described it at that moment in the best way, sort of like the formalized version of all that research in a way that could be taught. I actually wanted to ask you about, like in turn, sort of, that you had sort of alluded to it, like you did this at the bottom of like, where the market had already crashed out in a way. Do you think that the reason you could do that textbook then was that there was enough research and maybe the research had crystallized? How did you know it was ready to be kind of like, fixed in print?

Pramod Viswanath (00:20:38):
I think it's partly both of the factors that you mentioned first is that we had done research for roughly, I would say seven, eight years by then. Seven years through my PhD and and David's own personal research. And at that time we had also built this as a practical system and on the field, this was, someone said Radio Router is not a good name. So it was called Flarion Technologies and it sounds like toothpaste or something.

Anna Rose (00:21:04):
Radio Router also sounds a little like it's from the sixties or something, so maybe it's better. But yeah,

Pramod Viswanath (00:21:12):
and I think one of the things we did in this book, and something I clearly, clearly remember is that at the end, every few pages, at the end of every chapter for sure, we had something called a system view. Which is to put together all the ideas and how they would all fit together through these two lenses, ruthless lenses of robustness and scalability. What would it mean individually? There could be an idea that's a breakthrough. I mean, this is truly deep tech, wireless is you know, rocket science, like literally rocket science. Rockets like because we can communicate and you wanna do wireless and so this is deep tech, but at the same time, one can get into a very narrow aspect and like do it exactly the right way, like optimal but it doesn't make sense if you put them as a whole system they don't fit together. So every now and then in this book, we had the system view aspects, taking a whole stack view of how these protocols, this deep tech would fit in in the broader context and we had these two ruthless lenses, like I said, they're very sharp sharp instruments, robustness and scalability.

Anna Rose (00:22:17):
Are they like either or? Like, do you get more robust for less scalability or more scalability and less robust? Or can you get like high on both?

Pramod Viswanath (00:22:25):
Yeah, they're complimentary. Okay. So you can get efficiency and optimality at the risk of robustness and similarly, you can get optimality and efficiency at the risk of not having scalability so you can just have more performance, but not be as robust or scalable. So usually there's a third component, which is performance. You know, you want to do one gigabits per second or high throughput. That's a normal thing people think of. But then you may not be, you may not have coverage everywhere, or it may be less robust. Something happens and you suddenly start getting into a full car and start driving and your car drops. So you want to be efficient and have high performance at the same time. you want to be robust and scalable. This is the central engineering challenge. Efficiency is kind of, everybody wants high efficiency, it's clear. but wireless was especially so because you are paying dollars for spectrum. So it was roughly a few hundred dollars per hertz of spectrum nationwide license and you would need something like at that time, 1.3 megahertz, but today, like five megahertz channels to transmit on the phone

Anna Rose (00:23:34):
This is really bottom line then it's like, you need to be able to get as much into that to make it worthwhile to run these whole structures. I guess,

Tarun (00:23:41):
Can I ask maybe a stupid question, which I actually have not never heard the story too. When did governments realize that they could have a monopoly on spectrum? Because like, you know, think about this way. Like if you're, you know, Marconi coming up with the radio, it's not like the Italian government was there like, no, you have to pay us per hertz. Right? So like when did governments versus industry kind of realize like, hey, this is a sort of public good that we're gonna commandeer cause blockchains have the same actual problems in some ways

Anna Rose (00:24:11):
Ooh.

Pramod Viswanath (00:24:12):
Oh, excellent. I mean, I'm no expert on this history, but World War II was a era of radar and wireless. So it was clear that wars are won by having wireless technology, very, very clear, it would've made a pivotal difference and cracking the code, the enigma codes and so there was a whole history of communication and wireless in general, having central roles in how human civilizations it affected, you know, lot of, all of our civilization. I think that had some big role in governments realizing, gee, I mean, not so much making money as much as, you know, controlling our nationhood and statehood.

Anna Rose (00:24:52):
So even at the start of the wireless industry that, like, in the way that we understand it and the time you were there, like the government had already been keeping tabs on this for a long, long time.

Pramod Viswanath (00:25:02):
Oh, definitely. I mean, I think it's only gotten deeper in measurement. It's like transportation networks, like roads, and it's just in the air

Anna Rose (00:25:12):
Yeah. True. It's where the information goes. When did this book come out? And actually, who is it tailored for?

Pramod Viswanath (00:25:18):
The book came out in 2004, so it's roughly you know, I started thinking about this in 1996, about eight years of both research, thought process and on the field experience in an entrepreneurial way and bringing this tech in a scalable way. So it's, and the book indeed was a mix of both. It had algorithms, it had fun. The title said it, it was fundamentals, but the same time had a very system view is what we called it. So the audience was both universities, we taught at our own home universities, it was taught in dozens of universities around the world including all the mainstream ones in the US and to the companies. So we gave lectures, we've given tutorials and it's used at most of the wireless companies

Anna Rose (00:26:04):
Even today?

Pramod Viswanath (00:26:06):
Even today.

Tarun (00:26:07):
Actually quick question. How much of the book would you say, just like ballpark guess, was theory known prior to 1999 and you working in industry and how much of the like exercises or problems in the book came from like, hey, I was working and I ran into this thing and it didn't work. Out of curiosity, because like you wrote this book after having both like an academic experience as well as a practical experience like how would you divide up the book in that way?

Pramod Viswanath (00:26:33):
Yeah, it's the first few chapters I think were somewhat known but even there, there were subtleties, especially when you started looking at robustness, when we were summarizing work all done in a short four, five year window and some of the chapters and exercises were really brand new because they were done because we understood the landscape. So when you want to, of the engineering design space, once you understand the design space, it's somewhat high dimensional, but you know the design space now, you start saying, oh, there's a vacuum here, you know, with this dimension with antennas and robustness, there's no protocol and we don't know how it fits in and how they sit together. So that's a new component to study. So I enjoy this process of this pedagogical aspect is not just one way. So that means all the things we did, we can explain, but really it has gone back the other direction several times where it gives me ideas of new ideas to pursue.

Anna Rose (00:27:32):
Kind of moving from that point forward in your story, did you stay in industry? I kind of know the answer to this, or did you return to academia?

Pramod Viswanath (00:27:42):
Yeah, I returned back to academia. I became a professor at the University of Illinois Urbana-Champaign and I've always been close to the industry, so I've always had both hats at different time scales over my career and rarely doing both at the same time, but one or the other in an intense way, but without ever taking my foot off of both and mostly because I find them very synergistic with each other. When I'm teaching, I get new ideas that have different time scales of being useful from a practical standpoint and when I'm in the practical era, I get a very visceral feel of exactly what the burning questions are and some of these questions are business questions and some of them actually have deep tech questions and you don't know these things unless you try both and I enjoyed this aspect of being of being both.

Anna Rose (00:28:36):
When you went back though, did you continue to work on wireless or did you sort of switch directions?

Pramod Viswanath (00:28:40):
Yeah, indeed. One of my first students at the time when I went back and early students was Sreeram Kannan

Anna Rose (00:28:47):
Okay, that was when he was

Pramod Viswanath (00:28:48):
You met Sreeram

Anna Rose (00:28:48):
Yeah, yeah, nice.

Pramod Viswanath (00:28:49):
Ans his PhD was on basically generalizing this idea of how to make wireless as decentralized as possible instead of having these bay stations so coupled with each other where their clocks are coupled to, you know, tens to hundreds of microseconds. Can we have base stations that are separate? They somehow find a way of synchronizing, basically bringing BitTorrent-like protocols on the internet to wireless. And so that was Sreeram's PhD thesis. And I had several, a few other students who did PhD thesis broadly on this topic, which is how to do it efficiently, how to route data, how to manage interference, because if everybody talks at the same time and they're not being synchronized, that's no good. So how do you handle interference? These were things that we worked on with a series of research topics.

Anna Rose (00:29:42):
There was this period though, where like P2P at least sort of had a drop off. Did you ever shift focus? Did you ever look at other kind of angles?

Pramod Viswanath (00:29:51):
In other words, Web2, right? The other angle at the time the industry moved decisively from decentralization, the vision of Web1, the internet, to closer and closer to Web2 so from 2006, 2007 when Sreeram started his PhD with me all the way to 2012 to 2014, by the time the Web2 had been firmly entrenched, which is just a few players control all aspects. They control the source, which is YouTube, where you upload data, they control the output where you interact with the data. So both sides of the major platforms of today they had strongly enmeshed themselves in the economy and I mean, they were the economy you could say. And Apple, the big event that happened was that Apple came just before Flarion was acquired by Qualcomm. Oh, so I should say that the technology we did at Flarion today finally came into fruition today, it's called 4G LTE.

Anna Rose (00:30:44):
Oh, cool. Yeah.

Pramod Viswanath (00:30:45):
But it took some time because technology takes time and wireless is slower to adapt, partly because it's just as hardware aspects and we are at an era where hardware has become commodity enough that it can be programmable but it's still still not there yet and that's something I'm working on very actively.

Tarun (00:31:03):
Well actually one question about that era was, you know, I remember in the 2000's there were all these projects by, you know, now the big tech giants, but at the time I think they were maybe a little more in puberty about like, they all tried to do their own wireless projects and like none of them really survived. Do you remember that era? Right? Like Google bought Motorola and stuff like that.What kind of lessons can we learn from the idea of like whatever new technology is trying to buy the old technology? Because I kind of feel like in some ways crypto has had things like that where people have tried to like buy old technology or buy sort of, I hate the phrase Web2 or Web3, but like Web2 in air quotes technologies and try to bolt them on. Do you think there are any like, lessons to draw from this idea of like, hey, like the innovators who try to do the old thing usually fail somehow once they're big enough?

Pramod Viswanath (00:31:58):
Yeah, I think it's the innovator's dilemma. I mean, I think I mean, I can't speak to exactly what the big 10 companies' interest in wireless are, but to the extent I interacted, it was mostly about making sure Google product and Facebook could be on more people and so it's just to get to the end mile. But wireless was more than that, it was about opening up platforms to others and yeah so I think sit with the open aspect, the composable, the programmable aspect of wireless and the vision that you would get by having such open radios, just having a link and doing what you can do over it. It was not fully aligned with the platforms sort of core business goals at the same time, I think wireless became part of our society. I have an anthropologist friend who talks about how humans started wearing clothes, that's a major thing we carry on our body and the second most carried thing is some form of money, which is some form of money. It could be a wallet, but it could be some form of money. And the third most carried thing on the human body has been a phone

Anna Rose (00:33:06):
Of all time

Pramod Viswanath (00:33:07):
And it so changed us as an animal species and how we interact with each other through this device and it became, it was a force of nature, in other words, there is some biology that drove wireless so much.

Anna Rose (00:33:22):
Well, let's bring it kind of up to present and the topic we tend to cover on this show. So that's kind of going back to the second thing, the form of currency that people hold on their person. But at what point did your work start to intersect with the blockchain world and like yeah, did you sort of ignore blockchain at the beginning or Bitcoin at the beginning? Or do you feel like you you spotted it right off the bat?

Pramod Viswanath (00:33:47):
Oh, I didn't spot it, I mean, I'd heard of Bitcoin and but you know, just like anybody else probably read some articles about it and did it, and I'm not, I was not following it well. I was in wireless, which was kind of separate from that and I was not following it, but sometime around 2015 and I was looking for what other things to work on. I still was interested in wireless. I had a student who was very, very passionate about rebuilding wireless networks that are robust but in a different way, robust, not just like I said, every leaf of every tree bouncing off signals, but robust to adversaries and that was my first time I was used to adversity, but my adversity was usually nature when I was doing wireless. So you could say it's not adversity

Anna Rose (00:34:31):
Oh, it's adversarial. Oh, I get it. Okay. So you're

Pramod Viswanath (00:34:34):
It's benign. It's relatively benign. But but this student was very interested in people adversarially trying to do wireless for you could say malicious purposes and the student's name was Giulia Fanti, and she drove this research in a very strong way. You know, sometimes people have this view that maybe outside in the show, right, people think professors and students are like a hierarchy but I think I alluded to the fact that it's more equal but sometimes it's the other way around too and in this case, it was very much, I was driven by her passion and focus on trying to build wireless tech that was decentralized and we're trying to do decentralized in two ways. First, at that time, this was the time of Hong Kong protests and Arab protests in Arab Spring.

(00:35:22):
So people were using phones to team up to meet in a certain place to protest and the authoritarian governments were using cell towers to triangulate who is using this particular app to coordinate and zero in. and like literally triangulate physically where you are and come and catch you while you're on the phone, this kind of thing and the technology was, because it was centralized, you could do such a thing. The second thing was when you send a message, they would try to track down which IP address sent it and go back and catch it through that and so we were interested, and I think Guilia was very interested in trying to avoid both types and basically provide anonymity layers on top of regular communication and in some sense, this was my first project where I always thought about high performance as my main thing for wireless and decentralization. But now we have a new angle, which is adversary resistance.

Anna Rose (00:36:19):
Interesting.

Pramod Viswanath (00:36:19):
And adversary resistance were in network level and physical level resistance and these were, we did two projects. You can already see that there's some aspects of blockchains coming in.

Anna Rose (00:36:29):
Definitely.

Pramod Viswanath (00:36:30):
In fact, that's exactly where we started, how we started,

Anna Rose (00:36:33):
Were you looking for adversarial testing for decentralized models and that led you to learn more about this financial system that was sometimes getting attacked? I mean we've heard a lot about the adversarial nature of the blockchain space, but yeah, I'm curious how that linked up.

Pramod Viswanath (00:36:49):
No, I hadn't heard of the blockchain at all at the time and when we were doing this as purely wireless, and we were driven by very practical issues, we wanted to get a technology and an app out for individuals, literally actually ship it to the protestors and to provide A. resistance against tracking of the IP address, B. resistance against tracking of physical location. So we did two projects, we did the messaging protocol, and this was also the time when lot of messaging apps were coming along and WhatsApp was a big thing at the time and so forth. It had already become a fairly big thing and right, you still have Signal and Telegram. And so this, were all born at that era, by the way. So we had an app called Wildfire, we put it on the Google Play Store

(00:37:31):
but and the tech underlying it won some research awards, we called it Spy vs Spy and we had good fun from research and we were also very practical. I had this practical bent of mind and Guilia Fanti was also a very hands on student. I mean like she built her own bike and stuff like that.

Anna Rose (00:37:49):
Cool.

Pramod Viswanath (00:37:50):
She's very hands on

Anna Rose (00:37:51):
Nice.

Pramod Viswanath (00:37:51):
And so we built this app as an Android app and we released it online, so that didn't take off, but nevertheless, I think we got a feel for this adversary resistance. And we did a separate project called LTE Shark, where you would put an antenna on top off your phone and with this extra antenna on your phone, now they cannot triangulate you up to let's say 500 feet so they can say you're roughly in this area.

(00:38:18):
But previously there were the cellular towers can triangulate you up to a few meters, but a sub-meter. And so basically catch you, but this time you could have up to a hundred to 300 meters of obfuscation.

Anna Rose (00:38:32):
Nice.

Pramod Viswanath (00:38:32):
So this was a physical device, like an antenna system that you could put connect on top of your regular phone. So we did the research underlying it and the technology, and we brought it to sort of practical products, but they were products at a university level and so they were proof of concept you could say. They were funded by the National Science Foundation and the National Security Agency. So we are thankful they gave us a blank check to do this research.

Anna Rose (00:38:57):
That's cool.

Pramod Viswanath (00:38:58):
With nothing at stake.

Anna Rose (00:38:59):
I think, I'm still trying to find the link to blockchain stuff like, so this is still within wireless and this is almost like physical, you're talking actually like techniques in the physical world, but yeah where does it, where does it lead to this sort of decentralized network blockchain land?

Pramod Viswanath (00:39:16):
Exactly. So we were giving this talks around and we talked to a variety of people in both academia and in the market space around that time. we met a colleague we had just hired him as an assistant professor at Illinois. His name is Andrew Miller and Andrew Miller had been in blockchain since his high school from the days of Nakamoto and he was, blockchains were kind of all he did and he said, you know, this is very related to a major issue in blockchains, Bitcoin transactions. Bitcoin has anonymity at the wallet level as an address, but IP addresses can be tracked to wallets based on your transaction. It's an issue and there were some heuristics and hearsay about it on Reddit forums that people can do it. And he put this idea in our head, I clearly remember that sort of February cold afternoon in Illinois when he said that, and we got into this, we said, okay, we were trying to do de-anonymization and anonymity protection algorithms for wireless, but at a conceptual level, it's still a networking stack and we can bring it to Bitcoin.

(00:40:20):
So we did our first project where we tried to de-anonymize. Basically when you transmit a transaction on chain, on the Bitcoin network, it goes through the IP network. And if an adversity is just listening to the timestamps of which transaction package were received in which location where, then they can go and triangulate as to who the source, which IP address was probably likely the source of this transaction. And we showed, we built sort of machine learning protocols that can do it very, very cheaply. That means just a few hundred dollars of bot nets that you can build on the network and you de-anonymize really every transaction on Bitcoin. It's not de-anonymization to a physical person, but it's de-anonymization to a IP address, which is close

Anna Rose (00:41:12):
I think I see your link here. So it's kind of like you started from sort of the land of research, dealing with these big companies. You went into industry, you kind of come back to academia finding new problems and what you learn is like this adversarial mindset where you're like, you have to, like, in your case, you were trying to build to counter an adversary, but the minute you introduce an adversary, you're thinking like an adversary and then you were able to go into this new field and basically use the adversarial perspective. This is interesting. Yeah.

Pramod Viswanath (00:41:41):
and it was very close. We wanted to just prevent IP addresses from being tied to Bitcoin transactions. You lose your entire anonymity on chain if your IP addresses are tied to every transaction.

Anna Rose (00:41:54):
Totally.

Pramod Viswanath (00:41:54):
And so that's how we got in. And we, once we showed that the existing networking protocols they hadn't changed since the days of Nakamoto and so we designed the first sort of networking protocol with focus on how to bring an anonymity in your IP address when you transmit in terms of which transaction was sent. And we had this protocol called Dandelion. It was a lightweight protocol cryptographically very lightweight, and it was also compatible with existing, so not everybody needs to use it. So you use Dandelion on your networking stack. It was networking stack change, nothing to change the L1 layer or consensus or anything. It is just how you do the packets forwarding and Dandelion was a way of putting randomness in where you send and whom to send and when to send that provided this anonymity layer.

Anna Rose (00:42:46):
Dandelion, what era are we talking now? So this is

Pramod Viswanath (00:42:49):
Dandelion was 2017, 2016 through -17. We put that paper up on a public research forum called Archive and in the morning we had dozens of messages from people and a community we had never heard of. This was the Bitcoin community,

Anna Rose (00:43:03):
They had found your paper and they wanted to talk

Pramod Viswanath (00:43:06):
Just overnight and I said, oh, wait a minute, this is a very different pace of research and activity and engagement than I'm ever used to in the wireless world and it was really extraordinarily exhilarating to meet an open society, which is so interested and they wanna try it like right now.

Anna Rose (00:43:25):
Yeah. Did you leave academia then at this point? Did you like actually go for a back out into industry?

Pramod Viswanath (00:43:32):
Yeah, so we first tried to get Dandelion into Bitcoin Core. So we spent a lot of time, we met all their devs and it got into several cryptocurrencies, you know, Monero, Mimblewimble, several currencies used. It was a networking stack, so it doesn't change their consensus. So it's kind of harmless. Why not? And we showed that it doesn't harm your performance anyway so it's an extra level of anonymity that you have. And at some time around late 2017, Guilia had had just gotten a job at Carnegie Mellon as a professor, but before joining and just at the time of joining, she took some leave and I took leave and we got into the crypto world. So, that's our first crypto startup

Anna Rose (00:44:10):
2017. That's kind of the same time I started this show actually.

Pramod Viswanath (00:44:15):
That's right.

Anna Rose (00:44:16):
So, a lot of people jumped in then. Yeah.

Pramod Viswanath (00:44:19):
Yeah and I think we wanted to build a good scientific base. What's the best way to do proof of work? Proof of stake was a thing at the time, Finality gadget, charting, they were all topics that Ethereum Foundation and Vitalik were just talking and we wanted to understand them. They're all networking and communication aspects and I had just gotten flavor for adversarial resistance and I was always interested in robustness and scaling. So all these put together, I realized this is a thing for people like me and my collaborators in the wireless world. So I went and talked to a bunch of my friends from the previous era including David Tse and Sreeram Kannan they were doing computational biology and it wasn't a hard sell, wasn't because I think they can see that I'm fully in and I said, the water is warm and it's very easy to work in the area because we are communicating, we are storing and it's still networking at its core. Decentralized networking.

Anna Rose (00:45:17):
But that first project that you had them leap in with you in 2017, that's not what you work on today? What happened to it?

Pramod Viswanath (00:45:25):
No. So, 2018 crypto winter hit, you know, it was a rush, like within 12 months everything started and closed, that was a rapid winter and the same rush with which people can come in, I realize can also go

Anna Rose (00:45:41):
for sure

Pramod Viswanath (00:45:42):
In wireless. Somehow things are slow to come, but they also are slow to leave. I mean there this like some physical aspect to it, whereas here at 2018. I mean you were there, Anna and Tarun was there and

Anna Rose (00:45:54):
Sea change was fast.

Pramod Viswanath (00:45:56):
It was fast and a lot of people said, it's actually over. This was just a bubble for one time but I, and you know, as you can see, my collaborators that we worked together at the time, we were in it for the long haul.

Anna Rose (00:46:07):
Cool.

Pramod Viswanath (00:46:07):
We had enjoyed working on this area and we sort of went back to academics. I went back to Illinois and Guilia went to Carnegie Mellon and we were working on variety of research topics. We just wanted to understand the principles and the basic protocol. What's the best proof of work protocol? That was present. What's the best proof of stake protocol, which has dynamic availability and handle dynamic stake capabilities? What's the best way to mix proof of work and stake? The fungibility, you know, like some parts work, some parts stake. This is the minor art protocol and what's the best way to do sharding. So all the Layer 1 methods and then we've spent time on what's the best way to do data availability. So you may want to do new cryptographic methods that bring hash accumulators, Merkle trees together with information theoretic methods of coding and you put the two fields together. So we have this coded Merkle trees that are the sort of the right libraries to construct for data availability. So I think we developed what now I call the principles of blockchains at the Layer 1

Anna Rose (00:47:08):
I'm wondering all of a sudden if there's a textbook in the works, because that sounds like that era of like certain parts of it becoming formalized. Like when you, when you say you can like identify the best, then it suggests that like enough research has happened that like you feel confident that there's like a standard that could potentially be written down. Do you think we're there for some of those things?

Pramod Viswanath (00:47:28):
I feel comfortable to say that we have the principles. In fact, I teach this class right now. I'm teaching it at Princeton, Principles of Blockchains. and you know, anyone who's interested, it's web3.princeton.edu.

Anna Rose (00:47:39):
Cool.

Pramod Viswanath (00:47:39):
But it's a 2022 book. You know, when I wrote the book with David Tse in 2004, it was a book, like literally a book. I think 500 to 600 pages and it's that style had, it was published by Cambridge University Press. Not changed since the days of Newton, you know, they also published Newton's Principia Mathematica. But this is a 2022 book and so, which means it has about six, seven pages of lecture notes. There is a GitHub page and venue to program what the protocols are for every lecture. There is a YouTube channel and in a not to my very young millennial or even sub-millennial student, I don't know what the right Gen Z student

Anna Rose (00:48:19):
Zillennial, what are they? I don't know.

Pramod Viswanath (00:48:21):
Yeah. What the students are, we even have a three minute summary video for each lecture.

Anna Rose (00:48:27):
Okay, for the, the TikTok crowd, I guess

Pramod Viswanath (00:48:29):
TikTok generation. So there's a three minute and some students are asking me to make a YouTube short, you know, this one minute YouTube short. So it's a modern book with a Discord server, some parts unavoidable l-tech and old leaf documents. The GitHub and YouTube videos. So it's a modern book that I teach. It's a large class by Princeton standards, but it's open to everyone.

Anna Rose (00:48:52):
Cool.

Pramod Viswanath (00:48:52):
And I enjoy just discussing the basic principles of blockchains. Mm. I think, because again, this was done during the crypto winter from October 2018 through 2020.

Anna Rose (00:49:05):
Cool. Yeah. This reminds me a little bit of something that I'm doing over at the, at zkhack with the zk, I'm gonna do a little shill for something I'm doing too. zkWhiteboard sessions. If people aren't aware, we actually have like video series, Discord server and we are also experimenting with shorts, which is really funny when you're dealing with deep math. I don't know if it works because it's like, anyway, but we're trying it out. We're gonna see if, if yeah, if people like it.

Pramod Viswanath (00:49:31):
And I just want to say more power to you, you know, let the power of the universe come to such efforts. It's all very good. It's open permissionless. It's the opposite of the, and early in that conversation I mentioned about a country club environment. This is the opposite. Definitely. That it's an extraordinary feeling to be part of this democratic movement where knowledge is democratized.

Anna Rose (00:49:50):
So besides the book and the creation of some of these standards to better formalize and understand why some things work and then communicate it, what else are you working on? Because I know that, you know, we know you come in and out of academia, so yeah, what are you working on now?

Pramod Viswanath (00:50:04):
Yeah, and the crypto industry also came back in late 2020 and early 2021 and but wireless had never come back. It was still and in fact, Anna and Tarun, the wireless industry had gone the other way. So in 2000 when we were doing Flarion Technologies, there were a dozen carriers and a dozen equipment makers and several handset makers. But now over the two decades since the industry had extraordinarily consolidated, there were just a few carriers. There were four to begin with, and by the time 2020 came, it became three in the US there's AT&T, there's Verizon, and then there is T-Mobile, which took over Sprint. And so just three carriers and really just AT&T and Verizon, they're very highly consolidated. Whereas the tailwinds of the industry were blowing nicely, which is that the hardware became very inexpensive.

(00:50:56):
So base stations are like Wi-Fi++, you know, Wi-Fi router, you buy it, it's just a couple of $10, $20, $30 and these base stations that you can put together now 5G bay stations are all not that cheap, but not a whole lot more expensive, couple of hundred dollars and much of the networking stack went to Cloud. So the other thing that happened with Web2, which is helping us here in the democratization, is that compute became cheap and easily accessible. You no longer need a super computer or heavy equipment. You could just get it on Amazon or Azure, AWS or Azure very easily. So the networking stack, what used to be very vertically integrated, got splintered and went to the Cloud. And so the networking got controlled there. So the industry tailwinds have been blowing, and I'm still interested in wireless very much, even though my blockchain had happened, you know, still part of the same person. So in 2021, late 2021, I started thinking about how we should bring crypto native protocols to wireless and that's it's like an old flame meets new flame normally not a good thing actually but,

Anna Rose (00:52:08):
But in this case, maybe they were meant to be

Pramod Viswanath (00:52:12):
Very well and it's, I'm the same person and that's what I've been spending all my energy and intellectual energy and entrepreneurial energy on.

Anna Rose (00:52:21):
Interesting. Has anyone tried to tackle this, like mixing wireless and blockchain? Like and how would you? Is it sort of, is it in the incentivization of, I don't know, running infrastructure? Is it in something else that would like, I don't know, help to grow some sort of network?

Pramod Viswanath (00:52:41):
Yeah. So when I started, at some point, I think someone naturally mentioned to me this project called Helium, which is a blockchain. It's a way to incentivize people to buy Helium hotspots. So they sell hotspots, or at least they have vendors, approved vendors, selling hotspots. And you could buy them, put it in your balcony, put it in your house, my neighbor in Illinois, I had two, one in the front, one in the backyard and so I met mom and pop people who have these hotspots, and they used this to earn Helium tokens and by saying that I put it up and the way the proof would work so what Helium showed was the power of tokenized incentives, the idea that you could cover a large network and Helium at some point and maybe still now has a million hotspots.

(00:53:32):
So that's a large number. People put in a few hundred to a thousand dollars. So that's already a billion dollars, close to a billion dollars of <inaudible> coming in and using this and build the network organically. And I think it showed the power of the blockchain spirit, which is incentives, allowing individuals to participate together, you know, as a personal good, incentivize for personal benefit, but having a global effect and a global network effect. But Helium did this in a way that there's no, the incentives are, trust me incentives in the sense that they're Helium branded boxes, they're closed shut, they're not programmable so the trust is basically with Helium, but incentives are open.

Anna Rose (00:54:15):
I see, I see. I do see a bit of the kind of mapping of like these hotspots that are incentivized to minors that are incentivized, like minors all over the world are incentivized by the rewards that they could earn through running this stuff. Here you're trying to use that model in the real world to run actual infrastructure, but for something else. But what you just said was like, this particular project sounds like it's still pretty centralized. It's not necessarily built in the spirit of like truly decentralizing wireless or like actually creating this hotspot mesh that's truly decentralized. Is that true?

Pramod Viswanath (00:54:53):
The technology definitely is centralized because the trust is in the hardware. So you cannot participate in the Helium network with our own equipment in the lab. You have to buy it from Helium or approved vendors. The trust is very much in decentralized but on the other hand, the spirit perhaps was decentralized. I mean, I can't speak to that, but I think the tech was not in line yet and partly because, or maybe mostly because the tech was not there. So what is this tech I'm alluding to? Right? So you put up a bay station or you put up a hotspot and you give it a fibre connection. So that means you really switch it on, so it's now connected to the internet. How does anyone verify you are connected to the internet and what speed do you have? In today, we have things called speedtest.com.

Anna Rose (00:55:36):
Yeah, yeah

Pramod Viswanath (00:55:37):
You can do a ping test, you can go to speed est.com and you'll just check your up and down link. But speedtest.com is a centralized, trusted machine and you trust whatever numbers they give you. I mean, if they say you have a hundred MVPs uploading, then you accept it. I mean, what are you gonna do? You have no way of checking yourself and internet this, do the area of network, what is called telemetry. So that means you measure networking performance, like for example, I say my phone, I'm getting 10 NPS connection on my to while I'm downloading YouTube, like who's to check that? If Verizon says I got it, then I mean that's what I got.

Anna Rose (00:56:17):
I guess like the telemetry is always just given to you by the host, whoever's like providing you with the service

Pramod Viswanath (00:56:22):
That's it, it's centralized

Anna Rose (00:56:23):
You're doing great. It's great, or it's down down.

Pramod Viswanath (00:56:26):
Yeah.

Anna Rose (00:56:27):
The little bars on the phone, it's not like that's some yeah

Pramod Viswanath (00:56:31):
Yeah, there's not even he said, she said, actually there's only you know, they said and it's been that way, which is fine because wireless, and like I said, communication was always centralized and top down and so in some sense the basic tech was not there, and even the underlying inventions were not there and so this was an opportunity for me to think from first principles. So we rebuilt classic, very classic decades old internet measurement protocols, Ping, iPerf, Trace Route, which measure how much a certain link is carry capable of carrying in terms of throughput. But this time with Byzantine resistance,

Anna Rose (00:57:10):
Interesting.

Pramod Viswanath (00:57:11):
That other side people don't need to follow protocol, but you can do speed test without having a speedtest.com server.

Anna Rose (00:57:19):
That's cool.

Pramod Viswanath (00:57:19):
So we call this proof of backhaul.

(00:57:21):
We basically did this series of inventions, what we call proof of bandwidth. I'm using bandwidth metaphorically by meaning, you know, I have bandwidth. So that various network elements, and this could be backhaul, it could be service. Backhaul means I have a hotspot, I'm connected with the 1 Gbps fibre. Like how do you prove to other people that you have a 1 Gbps connection? And this should be trust free. That means a third party should be able to check that this proof was legit in the true spirit of proof of work, in the true spirit of proof of stake. Right, the stake is on the ledger. I can check if your proof of stake worked, and I can check if the nons worked in proof of work, how do you do in a similar trust, crypto native trust free manner, proof of bandwidth? So this is networking protocols meets basic crypto primitives.

Anna Rose (00:58:15):
Interesting.

Tarun (00:58:16):
One question I have is building kind of these like basic networking primitives in a Byzantine resistant way makes a lot of sense, but do you imagine that the adoption of this is from crypto networks first? Or do you imagine that the adoption comes from, say, you know, someone who's a pirate cellular carrier and like, you know, they're kind of like, hey, like we're using spectrum illegally and like this is like how we're going to like pirate radio style. I'm just kind of curious like where, where you think this fits on the spectrum of like who's the initial sort of user consumer of this?

Pramod Viswanath (00:58:55):
I'm looking at it on both sides. I mean, Helium sure that there is a demand for building ground up, and right now Helium boxes are closed. But if you can have an SSH port or we are working to sort of open them up and the Witness Chain is the company that commercializing these inventions and witnesschain.com is where we are. And these inventions, they're lightweight stack that lives on top of these radios. So it should be existing hotspots, right. Should be able to bootstrap and benefit from having trust free incentives. The other thing Tarun, is that once you have trust free measurements, you don't have to use Helium branded or any specific company branded equipment. You can buy equipment off the shelf for, for much less prices and this also drives the economics to make them because now there's a market for them.

(00:59:51):
And there are several other projects along with Witness Chain that are trying to bring relatively inexpensive hardware based stations from a ground up perspective. You also ask what happens to the other side, you know, like existing carriers, what would Verizon and AT&T and we've been talking to carriers on three continents, and the interest there is not because they already have some base station, they would like to grow organically, but more like they would like to resell bandwidth. So bandwidth is like airline seats, you know, if you don't fill them, they're gone. The flight is gone, the seat is not taken. Same here. Carriers are in the business of sending packets over the air. If the packets don't contain any data, they're gone and even a very good network like Verizon or AT&T would only have about 35% occupancy. It's kind of crazy. Most of their network is just going vacant because of the economics and it's again, because of billing. You know, I have a service from AT&T, it's $90 all you can eat service, it's a buffet. And there's no incentive for the carrier or me to do anything better.

Tarun (01:00:59):
So you're saying we're getting obese off excess spectrum?

Pramod Viswanath (01:01:04):
Well, I mean, but it's not it's best effort. You know, you get what you get. But many companies would like a lot more they're more flexible, but they would like a lot more service. I'll give you an example. Here's a contract, 500,000 devices, all Bay Area zip code between midnight and 4am, I want 1 GB of data, wireless, all Bay area zip code. This could be Tesla updating all their cars in the middle of the night between midnight and 4am every day. But there's even for such a mighty sort of even between them, they have not had a contract like this. And this I have from my own decades of experience. It's just a hard business because carriers have, I mean, they have lawyers and lawyers have to sit down and do the contract. In other words, data and networking is a commodity, but it acts anything like that.

(01:01:54):
There is no smart secondary markets. There are no futures on data markets. I mean, there's not a flexible marketplace and Witness Chain protocols by separating measurement and billing from service. You can separate the two. Historically they've been together basically, and which is why lawyers are there and you need legal contracts to enforce it because they're doing the same, both the service and the billing. So you need sort of legal contracts to enforce it but if you can separate them out and cryptography and crypto protocols provide trust on the measurements and the service that you gave and tie them together, then you have economic, what shall I say lubricant to move the market to become more efficient. So there are these economic forces, they are slower than to put it politely, way slower than the blockchain side, which is open and composable and permissionless

(01:02:54):
because these are only a fewer entities, but the market forces are in the favor of such technology being there. I am bullish, someone who's been around for some two decades in this industry. I have no doubt in my mind that, you know what you said you sort of hate this branding of Web3, right? I think it's true, but this would be one of the large last dinosaurs because I haven't changed. Since Web1 did not really change. Web2 didn't change them and this would be one of the industries that would become more nimble, more efficient and the forces are in favor of making this efficiency.

Tarun (01:03:28):
One actual I guess other point to this is, you know, in the same way that trying to be a new type of money fights the nation state, trying to unbundle spectrum implicitly does sort of have the same subversive like cipher punk thing because you're sort of, you are the pirate radio except for many different band communication in some ways for communication. Do you imagine that that will be a sort of fight like where like the big carriers are like look at like let's suppose this whole space grows a lot and do you think the big carriers will try to go kind of lobby governments to be like, no, no, no, this illegal government's like monopoly over spectrum is being violated by this type of sort? Because like it could actually have that kind of iconic, classic end state because it is challenging a state monopoly in the same way they were challenging a state monopoly on money, which is actually kind of interesting in the sense that I don't think anyone's framed it that way. Obviously, I'm not trying to say that we should try to make it as adversarial as that, but it could be viewed that way very, very easily

Pramod Viswanath (01:04:39):
It could be that would be I think we've already won at that stage when we get there and so I would be that would be I think a good day already because when you are having pushback at that level is a good thing, that means you well past the noise level. But the third aspect Tarun, which is sort of you said is the, is the role of the government from a national security perspective. The National Security Agencies from the Army, the Military, the Navy, they have been always been bullish on decentralized networking, partly because they work in hostile areas where there is no way to enforce the rule of the law so in a way, sort of government is already against crypto, but in this case, actually they could be in favor because of just making it more robust to operate in environments which are inherently hostile to operate.

Anna Rose (01:05:27):
Interesting.

Tarun (01:05:27):
I guess I just think about all of the kind of like the famous sort of like, I mean, depending on how you view history like the robber barons of spectrum, right? Like the Ann Schultz of the world and stuff like that and so there's sort of this question of like, they've sort of effectively enforced, you know, like Dish and like a few people who've like won the spectrum auctions by perhaps unscrupulous means they have a lot of incentive for this not to work, right? In some ways. I could imagine that ends up being a war, like I agree with you, like if it even is a war, it will have already been like you would already have been successful, right? But like, if we just like play this out to like that piece of the world, what do you kind of imagine there are any, like the adversary could do in response? Like let's say this network grows and like, let's say there's a million devices using it. How do you think it could get, do you think there is some sort of like regulatory crush or do you think it's like, it really is like a Layer 1 blockchain where it's like once the chain is around, there's enough people and there's enough density that it's just very hard to censor at some scale?

Pramod Viswanath (01:06:38):
Wireless is decentralized by geography. You know, there's so many people in so many places and just literally distances. So I would say that this decentralized is hard to crack down as you would crack down on the internet. You know, like the Great Wall of China. So you can crack that down but here, this is truly decentralized. There's, you are to go block by block every few hundred feet and sort of shut the things down. It's, there is something very innately decentralized and open about wireless that is sensor resistant, but the tech and underlying networkings around it should be commensurate with it and I think for a variety of reasons, the tech had never sort of been invented to make adversary resistant measurements and protocols and Helium sees it actually, right. What kind of adversary action do you envision? Well, Helium, you see it because it's Helium branded tech, and they have people report, I put it hotspot here, and I have several of them. Well, I mean, you can see some things in the middle of the Pacific Ocean. Well, how did, I mean how did that happen? And someone if

Tarun (01:07:44):
Your argument is that there could be a supply chain attack against them. That backdoor is like the actual device, which

Pramod Viswanath (01:07:49):
That's exactly what is happening.

Tarun (01:07:50):
Yeah and the fact that you can't have yeah, yeah, yeah.

Pramod Viswanath (01:07:53):
And the protocols are, trust me, then you don't have the trust and it's vulnerable and you can see it actually already happening on that network. So what we are trying to just do is to first principles bring measurement techniques that are native crypto, so you can verify from far,

Tarun (01:08:09):
This is the most violent usage of metrics in some sense, because if it succeeds it actually actually does like, sort of break a state monopoly, which is very, you know, I think that that to me, that's like the thing that's most interesting about it

Anna Rose (01:08:22):
Cool. I wanna ask you a little bit about the proof of bandwidth and the proof of backhaul. I'm just curious, like you sort of said it, but that's not your consensus mechanism is it? Like, what is, where's the blockchain part? Is it sort of like there's incentive to run these things yourself just by having a token? Or is there some crypto like cryptography, blockchain part in the actual system?

Pramod Viswanath (01:08:47):
So excellent question, Anna. So this is not a Layer 1 by itself. So what these are the, these are crypto protocol, just like your proof of work. You have a nonce and you produce it nonce to do some activity. That activity in Bitcoin and proof of work is propose a block and go inside a chain and that's part of a consensus protocol here. Proof of backhaul is a crypto, again, just just like proof of work would be of showing a nonce. Here you have some other proof. It's a cryptographic proof or a witness of the network. That's why we call it Witness, because the network is literally witnessing who got what and so the witness proofs are taken and these proofs can be put on chain as an OP code. So I think of these proof of backhaul or witness proofs as OP codes, which are network level OP codes that can be used to tie incentives on. So there's not a Layer 1 by itself. There could be, it could fit these up codes can sit on any other existing L1 or all of them. In fact, you can draw them as an oracle, pull them in. So they're naturally implemented as A. proofs, which are, there are state channel that can update and keep track of these proofs over time and from a blockchain perspective, pull them in through an oracle.

Anna Rose (01:10:02):
Would somebody build a Dapp then? I'm just trying to picture like even the incentivization token part. Like would it just be an ERC-20 that lives on another blockchain that has some connection? Or do you imagine there being like a Dapp that manages this somehow?

Pramod Viswanath (01:10:18):
Yeah, so what is allows is two things actually, right? So I think first is, I think I already, we talked about it, which is that any device can come in, you don't have to have branding by a company and trust it. So that part is good. So any device, any tech can come in, you can have Wi-Fi, 4G, 5G, and our demos are on all these technologies. The second part is the, what you're alluding to, which is the composability and programmability. What do you do with these proofs? Who will use them? Which kind of an app or which contract will use them? Couple of use cases. First use case could be a carrier itself, which is selling and reselling bandwidth. This is a marketplace for data and just like you have futures and marketplace for data, you want a contract, you're a business, you're a dash cam company that wants backhaul in certain zip codes for certain amounts of time, you can buy them through contracts on chain.

Tarun (01:11:08):
This is the subversive part by the way. This is the part that s actually kind of anti the state's monopoly on spectrum. I will just, I just wanted to point out that is that is actually, you know, we've spent so much time in algorithmic game theory designing these auctions for a monopolistic state and like this is basically getting around that. So like, I think that's like kind of the most, you know, salient point about this because such products don't exist because state has transfer restrictions on any transfer spectrum. Right?

Anna Rose (01:11:37):
Interesting.

Tarun (01:11:38):
Whereas like here you're arguing that, hey, I can just sell my own, I can resell my spectrum whenever I want.

Pramod Viswanath (01:11:42):
That's it. We call it ECN, Elastic Cloud Networking as opposed to EC-2, which is Elastic Cloud Computing. The equivalent of freeing up networking and having a marketplace and compute was in one place like Amazon sells AWS and Microsoft sells Azure. But this is decentralized because networking is mom and pop putting it up. That's one type of an app and we have an ECN demo liveing over the air, and it's live in South India, Bangalore, we have spectrum experimental license to try this out. So ECN is live over the air in India. Another app could be, and I'm imagining this, this doesn't exist yet, is a DeFi app that wants to have hard assets as collateral. Many DeFi apps could be living on their own tokens and they would like to collateralize it or bring hard assets. Well, you can bring hard assets perhaps through a stable coin maybe, right? But you could also bring like literally hard assets by physical assets of networking. If you actually have a hotspot that's serving so many people and getting such and such revenue and such and such, backhaul, chances are you won't exist overnight. I mean, you have physical devices and physical service, so this could be hard collateral to put on chain and then offer as basically make the market flexible. Make it so there is no bring debt financing to the networks. So this could be a way in which people could finance hotspots.

Anna Rose (01:13:06):
And it's kind of, because you have these proofs, I guess it's digitally native already.

Pramod Viswanath (01:13:09):
Yeah, the proofs tie them together. Exactly. They're trust free proofs and these are hardware. Right, how do you show, and we didn't get into the deep tech, but I think this is the, that's the fun part, what I meant by old flame and new flame.

Anna Rose (01:13:21):
Do you wanna actually be part of the hardware creation though? Or are you sort of like, assuming these hardware pieces already exist and you're just sort of the glue?

Pramod Viswanath (01:13:29):
Yeah, they exist. They exist already. So in fact, one of the, so I talked about system view Anna, which is to think how different parts of a system put together and the same system view comes in here. So the hardware are very inexpensive ones that sits on the cloud. Witness Chain is a very simple app through a standard interface, GRPC interfaces on the cloud, and a very simple app on your device site, which is, could be a phone, it could be an app on the phone, on a Wi-Fi router, it could be a simple firmware upgrade. They are simple, lightweight software only stitching together. So devices exist, the pieces exist, we just wanna stitch them together in a trust free manner.

Anna Rose (01:14:09):
Cool.

Pramod Viswanath (01:14:10):
That's what a carrier does. A carrier today stitches together by bringing the trust. So you trust Verizon, you trust AT&T. So they're the trusted entity stitching together everything and billing and metering everything. And we are basically saying this crypto protocols should replace a carrier today.

Tarun (01:14:28):
If you, you know, let's say Helium was the Bitcoin of sort of wireless distribution. Where do you view kind of Witness Chain? You know, what, what sort of, how would you compare it to some, you know, when someone's like, hey, I've heard of this, I heard of this Helium thing, I read some good press and I read some very bad press and then, you know, how would you convince them that, that there's like something kind of greater here in the sense of you know, avoiding some of the pitfalls that we've seen before?

Pramod Viswanath (01:14:59):
I would appeal to the better you know, the good side and the angel side. I'm always on the, on the <inaudible> side and I would say that you want, these boxes are closed. You can't do anything with them to open them up. So we are starting this in Witness Chain. We have a sister site within Witness Chain called heliummev.com so Tarun they are doing MEV but there's no MEV on wireless because it's a closed box with no way to assess their chain. No programmability. It's a fixed box. But for every fixed box there is a way around it and we are bringing some basic MEV activity, which is a force of good, I think, of innovation to this area. So this is a more incremental, Witness Chain is a more, is the final stage but heliummev.com through Witness Chain is a starting point. If you already have Helium hotspots and we'll have more to come. So help to put box in the middle and just to open up these boxes that one has already and we are hoping that they can be opened up. And if not, we'll try making them open ourselves, you know, these are physical devices and we also have the tech to open them, so they're never really foolproof and so that's our effort with heliummev, just to provide some programmability, the boxes exist and just open them up.

Anna Rose (01:16:15):
You're basically saying like, it's cool that they existed no matter what, because the boxes got out into people's like hands and that's something you can use.

Pramod Viswanath (01:16:23):
That's right. And why would people throw them? They already paid really good money for that and and I'm always a favor of using the emotion and the energy associated with that what you are alluding to into a positive force.

Anna Rose (01:16:36):
Cool.