🎂 Jay Adelson & Matthew Prince Fireside Chat

Jay, awesome to have you on Cloudflare TV to celebrate Cloudflare's 10th birthday. You've been someone who I think we met right at almost the beginning of this story and you've always been someone who I've really admired and looked up to and appreciated your advice over the years. You've had this incredible career, you started Equinix, you co-founded Digg and Revision3, then SimpleGeo and Opsmatic, then you were a VC for a while and now you're building sort of the smart brains inside of even classic pinball machines. We'll talk a bunch about that, which is exciting, but every step in this seems like it's been an adventure. Is there a thread that links them or what is the sort of overarching story of your career? Well, I think like a lot of folks who kind of got into the Internet early, I was one of those kids who was really into like bulletin boards and like dial-ups and communication and I've been obsessed with this idea of communication sort of bridging and sort of expanding consciousness. It's like a drug that once you see the power of it, it's hard to kind of back away and I remember back when Tiananmen Square happened and we were finding out about it and that part of the way that that all got organized was actually university email prior to like Internet taking off and you see how it can be great, you know, Arab Spring and all of these great things that could happen and it's just a question of like what you want to do with it and so you know all of these companies in one way or another, they have something to do with that universe. I don't know if I know it when I start it or I do it. It just sort of happens and then you just sort of go with your heart. You know, I was trying to think of what the thread was and it seems like everything you touch inherently has this sort of network effect quality to it. I mean that was what you really pioneered at Equinix. I mean, Digg was a big piece of that, even Scorbett now. Do you just walk around the world just sort of looking for network effects or is that how you see things? Wow, you know, the network effect, I mean there's sort of two ways to look at it, right? There's kind of obviously anything where the more people use it, the more valuable it becomes is an attractive idea. I mean, it could be a very addictive idea but there's also kind of a Buddhist thing going on here, you know, like the Anada, the sort of the, you know, the idea that we're all really not separate from each other and that we become more powerful as we sort of interconnect. It's easy to do that on a network logically and you can see the value as your routing table grows and all these kinds of things. But the same is true with ideas and really Digg and so forth and ideas can become more powerful. There's pros and cons to this network, these network effects, you know, but overall I think I just love the interconnectedness of things and it happens to lend itself well to these network effect business models. You know, I think that we were obviously in the news every day seeing some of the cons to this. With Digg, it really was one of the first social networks. It was, I mean, it had a lot of the gamification and, you know, it was like buttons and upvoting and you'd dig things. What were some of the, like, did you, what were some of the stories, I mean, the content moderation issues and everything that you had to, must have had to deal with, like what was going on behind the scenes there? Well, you know, my idea of a perfect social network was one where the users moderated the content themselves, right? And the idea of that, which now is really Reddit, you know, but back in the day, Digg, I think, really sort of pioneered it. We actually patented the like button, by the way. LinkedIn bought that patent when Digg kind of disappeared. But I think that, like, you know, that was where you sort of learn who really owns, who really owns this democratized medium. You know, we had that famous incident with, when Blu-ray and HDVD were battling it out. And we had this, somebody posted the code that you could crack the HDDVDs with on Digg. And then it got voted to the homepage. And we got a DMCA takedown notice. So we pulled it down, not knowing what would happen. And of course, people started submitting the code once a second to our infrastructure. And we started to have to spin stuff up just to be able to handle it. And then we were taking it down automatically. And we're like, why are we doing this? And we contacted the EFF and said, you know, what should we do? And we ended up, because people were like, actually like singing a song with the code numbers with like a guitar, just so that our algos couldn't like pull it off automatically. And finally, we just posted a blog post with the code number and said, you're right, this isn't ours. This belongs to you. And I love that idea conceptually. I mean, and I left Digg, it was really still in the hands of the users. Like I left probably about six months or eight months before they sort of turned that off. And the users went to Reddit. But I mean, that is the ideal situation when you can make something that is truly owned by its users, right? It's just, I don't know if anyone's come up with a model yet that really works for that, unfortunately. Yeah, I was gonna say, does that, at scale, does that just always break? Because I mean, that was the vision of Digg. That was the vision of Reddit. I mean, it was a lot of the vision of Twitter and even Facebook, that they had that. And then you start to see, you know, the negative effects really creep in. And then the slope slips at some level. Is there, if you were running Twitter today, if you were running Facebook today, do you have any thoughts on how you make those more healthy platforms? Well, you know, I've worked with people who've tried very, very hard. And I think the kernel somewhere in here is this idea of fact checking and sort of like, being able to distribute the ability, sort of apply the same notions we have towards reputation of users in these systems, user reputation systems, apply it to content, and be able to create filters that make it impossible, impossible to widely distribute misinformation. That's it. That is, I think that's achievable. I just don't think anyone's really, it kind of, unfortunately, is, is the opposite of what, you know, the venture capitalists really want, because they want the, you know, control of the as much revenue creation from these sites. And there's a natural point where you basically are going to be forced down that road that Facebook has taken. And there's almost no way out. It's almost like it, you know, back in the Wikipedia model, it's almost, you have to sort of roll this out, almost in a neutral way. And it's super hard to do. It's, it is. And it's, you know, one of the things I've always enjoyed about you is that you sort of understand up and down the entire stack and understand, you know, how that's both possible, but incredibly difficult, understand what the kind of incentives are, but then understand, like, what a routing table is and pull all that together, you know, and so going back to sort of, you know, the beginning of your career, you know, Equinix is one of those just absolutely critical parts of Internet infrastructure. And, and even, you know, the sort of the way that peering has, has allowed networks to come together. What, walk me through kind of how that, how that, how that started, where, what were you doing in your career where you thought, oh, we can take these data centers, which seem like these totally commodity things, and make them actually this, you know, really critical part of how the Internet works? Sure. Well, I'll give you a short version. I was, I was running network operations for a company called Netcom, which was AS2551, you know, way back in the day. And, and Netcom really was one of the first commercial international Internet service providers. It wasn't the only one, but it was one of the first ones that wasn't sort of born out of like a regional university network. And, and I got recruited into the R&D division of digital equipment by Al Avery, who had been working with Paul Vixie and Steven Stewart and all these guys, creating some of these ideas in the lab. And really, Paul and Steven had, had really built this basement interconnect point between Alternet and Barnet, which would later be, you know, UUNet, BBN, Genuity, Level 3, whatever we call it today. I don't know, even know what it is today. Century. Now they're called Lumen. It's like, it's so, it's like, I can't even track. And so, and so basically my job was to sort of combine the idea of a neutral exchange point with some kind of business model around, you know, sort of like, how do you subsidize this sort of idea? And one of the principles at the time, which by the way, we didn't know for sure would work was if we flat rated the interconnect. So there was no like real, you know, price gouging, which we can get into a whole other debate another time about what that means. But, but if we can make it flat rate, could we make enough money from the real estate so that we never, so that we can allow that core to grow infinitely. And it was too successful because the Palo Alto exchange exploded, became way too, too overcrowded. And right about that time, Compaq bought digital equipment and we asked them what they wanted to do with it. And they said, we're going to turn it into dial up points of presence for our, for personal computers. And I'm like, that's not what we do. And they're like, well, you know, we're going to become an ISP, which they never did. But that's one of the reasons why Alan and I left, got some venture money and founded Equinix because we wanted to continue this idea. And really the plan had been to then absorb the Palo Alto exchange immediately. But Paul took over the exchange and really sort of grew that into switching data's business. And then ironically, I don't know how many years later Equinix acquired switching data for like $600 million. And I was like, okay, well that was a long time ago. But it's- It is crazy that, I mean, this building still just off University Avenue in Palo Alto, which was right near, it was two blocks from our original headquarters at Cloudflare. It's still, I mean, it's one of the critically important places where so much of the Internet comes together. And it does come back to this idea of how important network effects are. Once you get networks to interconnect, it really does centralize and generate an enormous amount of value. Who is the brains behind making Equinix look like a Battlestar Galactica-like set? Because you still, you walk into this day, when we were on Cloudflare's IPO roadshow, I remember we were in Chicago and we'd been talking about how interconnection works and everything else. And I said to the bankers, do you guys want to see a data center? And so we went to CIRMAC, 350 CIRMAC, and walked them through. And it's still, it's like, it's theater as you walk in. Well, I went to film school. I mean, let's call it what it is. So my theory was kind of sort of like, if you're having trouble with discipline of, say, a group of people you're managing, and you make the office look really clean and sort of a little bit more disciplined, it kind of impacts the behavior of your employees. And I always believed really strongly in that idea that you can do that with visual. And we actually hired a designer, an artist to help us with the Palo Alto exchange. And then when we started Equinix, we hired a design firm to help us. And the goal was to improve security, make it look cool, help it close deals, and help people come in. But ultimately, I'm a sci-fi. I love it. I mean, come on, who doesn't want to use a biometric scanner to get into a building? That's just cool. I mean, the doors literally go as they open. It's like, is there a speaker here? There's nothing that should be making that noise. A little speaker tucked into the ceiling. I mean, it does seem like exactly it is still to this day. And I will say that I love the sort of dark feel, but when you're in a rack, and you drop one of the rack screws, and the rack is black, and the screw is black, and you're searching around for it, I'm like, damn that jail. Could you have just put a little bit more light in here? Yeah, it's a bummer. And by the way, I'm a customer now. And so I go in, and I lose my rack screws in there, and I'm just kicking myself constantly. But at least it's cool. And the data centers do stay up. Of course, these days, why use a data center? We got you. So that's your problem. There's still a lot of atoms below all those bits. So at the end of the day, if the power goes out, it's still a problem. But yes, it's increasingly worse. So you've done all these interesting things. I remember talking to you a few years ago, when you launched being a VC, and you did that for a while, and you still invest in a lot of things. It never seemed like your heart was totally in it. So many of my friends are like, oh, wow, if I'm really successful someday, I'll get to be a venture capitalist. What was it that was unsatisfying about that for you? Well, I like the teaching part of advising. I like going to an entrepreneur who has never done it before, or maybe has done it once, and helping them gain the system so that they can maximize their value and execute and navigate the minefield that is startups. And that's what I thought I could do as a VC. And what I learned is that that's not what a VC really is. I mean, with all due respect to very successful VCs who are friends of mine, they spend 80% of their time raising money from people who know nothing about this world that we live in. And then when they are with entrepreneurs, and then they're with the entrepreneurs, and the structure of venture capital is actually designed to really, there isn't really shared risk between the entrepreneur and the VC. And that just pissed me off. And every time I tried to change some of those structures, the very LPs that invest in venture were like, well, what's worked for us for the last 20 years is the way it is. So if you try and change it, we're not going to invest in that. And so it's just, I don't know, I think it was Mike Volpe at Index Ventures once called me a communist, because I wanted to share more equity with my employees. And I was just like, what is, you know, and I get it. But the bottom line is, it wasn't for me. And you know, the fun we did, we invested in some really fun companies, and of course, Internet infrastructure and that kind of thing. But now I'll just focus on Pinball, because it's way more fun. Pinball is way more fun. So tell me about that. Like, what is Scorbit? And what are you trying to do? Oh, I mean, my son and I restore pinball machines together. And, and really, that got me really obsessed with all the electronics of these vintage machines, and even the new machines, like I loved it so much. And a couple buddies of mine, who I knew from Revision Three and Dig, Ron Richards and Brian O'Neill, they both were serious pinball players, like tournament players and would take me to the, you know, I just like to play with them. Like, I'm not really good at it. But I really enjoy it. And I love the art and the music and all this stuff. So they started taking me to expos and they were doing the tournaments and I would watch and I just was blown away that these things have computers in them since the late 70s. And yet they were writing down scores on post it notes. And so I just sort of like as like a little hobby was like, well, maybe I can figure out a way to connect these things up to an API somehow. And I don't know what happened. It's like somebody hit me in the back of the head five years ago. And like I blacked out. And then like five years later, I've got like pinball parts all over my garage. And we've created hardware and a SaaS platform. And, and now we can take any game from like the mid 70s till today, like any of those games, put this little device in it, and it automatically pulls the score and game mode information and puts it into an API. And then we have apps and visualizations and sort of like Xbox Live for like real world games. And we're just, we just did this for fun originally. And then it got to this point where, oh, it's, it's a real, it's a business. It's a business for a niche market. And you know, there's all these verticals we can expand to if we want to. But right now we're just, it's just amazingly fun. And my son is working with me and my family. We're all like boxing and shipping units to people. I've never done a smaller medium business before. I got to say, it's really, really satisfying. So, so from the end user's perspective, you know, so I'm, I'm playing a pinball machine. And what, what is, if it's Scorbit powered, like, what is that? What is that? What do I get? What, like, what is that? So, so if you walk up to a pinball machine, whether it's in your house or it's like in an arcade and you have the app, you basically are telling our service that you are in fact player two. Well, now that score in real time is being broadcasted. So not only can people spectate on scoreboards, but leaderboards are infinite, like global. And, and you can, you know, challenge someone to play that game. Imagine getting a notification that your friend beat your score at Zeitgeist, you know, and okay, I want to go play that game just to kind of, and achievements. So you take all these games made in the seventies and eighties, and now you layer all that stuff that Xbox and PlayStation did with, with achievements and, and make it fun and, and try and make it accessible to people at skill levels. Because if you walk up to a machine somewhere, like in the basement of Cloudflare and the top four, you know, scores are all Matthew Prince. It's like, you're, you're kind of discouraged because it's not who runs, who runs our, our network team is a pinball wizard. So, well, I mean, there's something to this and what you do with that data. I mean, I think the idea is like any API, you build it, you put some sample ideas out there, and then you let other people run with it and have fun. So that, so at some level, that seems like such a simple thing, like let's interconnect all the world's pinball machines. But on another, I've actually opened our, like the pinball machine in our office, and it is this remarkably physical, like, you know, and it, you know, it was built in the, in the, in the, in the nineties. And, and, and again, if you've got ones from the sixties and seventies, how did, because these things were never intended to like pull the data out. How does, how does that work? We, we build hardware. We had to, we actually got, so one of our developers is the lead robotic scientist for NASA's ISS mission and was knighted for his work on Hubble Space Telescope. Another one of our, of our guys, Olivier Galliez is in France, who, who's basically a professional reverse engineer. So we put these devices in that can read display data and CPU data and actually speak the language of these older machines, all the way up to new machines, and then interpret that data, kind of normalize it and then push it to the API. So it's an IOT little device. Now, new machines, we partner with new manufacturers. They don't bother with our hardware. They just connect to the Internet, but the older machines, I mean, this was, it was, I'm not going to lie. I mean, no one should ever do this unless they're crazy, but we did it. We got like every single brand and every single manufacturer and we tested it and it was super hard. I was going to say, how did you even track down these machines to, were you like breaking into bars and saying, hey, can I borrow the key? And yes, exactly. That is exactly what we did. We went to like the owners of arcades, like there's a place called Free Gold Watch in the Haight that is like a fantastic, you know, little arcade and people who own them in private collections. And we said, hey, can we have access to it? And of course my poor wife, Brenda, my house is now an arcade, you know, because as I was like, oh, I need a system 11 to test on. So I guess we're going to have to put it in the garage, you know? So, so this has been our, our process and, you know, it's, it's imperfect, but that's, what's great about a tiny team and, and, and, you know, building the software and saying, oh, you know, it'd be really great. We need a feature to, to allow X. And then some passionate person in the community creates that for us. It's really, it's a wonderful thing. And, and it is a, there's a business here, but the business is sort of orchestrated back to your network effect to allow the community and people to make it more powerful themselves and then be kind of a neutral platform, almost like a neutral sort of unaffiliated Xbox live. Yeah. And we'll see what happens. Like it could be really fun. That's so what's your, what's, what's it been like doing this with your son? Well, he's, I mean, I, I know that in Silicon Valley and in the startup world, you know, people say nepotism and having family members can be challenging because of all sorts of, you know, people's belief system around that. And that might be true in corporate America. I don't know. But when you actually work with somebody like that and you can mentor your own kid, who by the way is now 20, but we won't go there. And it becomes a really incredible, like some people throw a ball around with their kid, you know, with me, I'm building a business with him and he's seeing all this. And even if he's not, it's, it may not be a family business, but that process has brought us together in ways that, that nothing could have done before. And, and it's been, you know, it really sort of changes your perspective. And in his case, as he's launched into the planet, he finally knows what dad does and he can apply it to something in his life. And it's, I, I strongly recommend it's sort of like when Tony Hsieh had people who were going to work, sales had to work support, you know, as part of his designing happiness or whatever it was. And I feel like in the same way, there is something really wonderful about being able to truly know what each other are doing at a deep level and it improved our relationship. It's really great. And, and I'm going to try and convince him to keep, to stay with us. We'll see how that goes. Well, that's so what's your, what's your, what's your favorite pinball machine or top three? Well, I mean, like way above the rest is theater of magic, which was designed by John Papa Duke. It's a 1996 machine. Everybody loves medieval madness and Adam's family. Those are great, but brand new machine that just came out a few years ago is called total nuclear annihilation. And it's, and it is absolutely, it's done in the style of eighties machines and is super fast. And the soundtrack, it has like a subwoofer in it. And the soundtrack is kind of like electronic, a little, little trancy, but it's like, it's really intense and it pumps you up. And it's, it's an, it's an incredible work of art. And I, if you get a chance to play one of those and you know, it's like, there's also the funny ones that have come out recently, like Rick and Morty pinball. And, and I think there was a new one that just came out. That's really fun to play called Jurassic park. That's based on the original movie. Not any of this strongly recommend that one too. There's some really fun ones. All right. So we've got about a minute and a half left. And what we're supposed to, what I'm supposed to ask you about is not pinball, but I'm supposed to ask you about what's going to happen over the next 10 years of the Internet. And so given your wide experience, literally having built some of the key foundational technologies that, and systems that have brought together, you know, in, in a minute or less, what are sort of your predictions for the next, the next 10 years? Everything, everything because of virtualization is going to shift the economic base away from the core and to the edge, right? I mean, the, the idea of the worker, the idea of that now is going to be an economic shift and that's going to impact things like interconnection as well. And I'm on the board of Megaport and, and we're watching this happen. And it's not just because of like COVID and people going virtual because of that, I'm talking about just where people make money in this supply chain is going to push to the edge. And as a result of that, I, I'm not sure the data center really matters anymore. And, and I, I, let me rephrase that. It absolutely matters, but who the customer is has changed. And I do think that the enterprise just isn't going to be thinking about, you know, tethering anymore. It's going to be, you know, the, the SD-WAN universe. It's just a different planet. Of course, the core is going to continue to grow. The Internet is still going to be operated by the same, you know, 500 people that go to Nanog. Thank you, Jay. Thanks for being a friend. Love you all.