Mike Maples, Jr talks to Marc Andreessen, Co-founder of Netscape—the startup that launched the Internet era. What it was like when Netscape was about to blow up? What does he wish Netscape had done differently? -- and what lessons from its success are not fully understood?
Marc Andreessen:
So it was a strange time. There's an old William Gibson quote, the author or Neuromancer, all the great science fiction books, he says, "The future's already here -- it's just not evenly distributed." And so this was a great example of that-
Mike Maples:
That's Marc Andreessen who's now a famous venture capitalist at Andreessen Horowitz. At age 22, he founded Netscape, the company that kicked off the internet age. How did Netscape even happen? And what was it like when it was about to blow up? This is Mike Maples, Jr. of Floodgate, and it's go time with Marc Andreessen.
Mike Maples:
I decided to do some things a little differently with this interview because I wanted to capture things about Marc's startup journey that most have likely not heard. I wanted to get a level deeper about how Netscape and the early web grew out of the Mosaic project. But I also wanted to help people see a side of Marc that few people see in public. While you might see him share books he likes on Twitter, it's hard to connect this to what it's like to spend time with him. He's an incredibly avid reader in diverse areas like technology, science, art, literature, science fiction, history, politics, economics, psychology, and he can synthesize ideas that connect the dots in the blink of an eye. And that curiosity, combined with his technical depth, is definitely a super power.
Mike Maples:
If you've wondered about the origin stories of the web and what it was like to be in the middle of it, Marc goes into a depth that I haven't seen before. And he also reveals the three things he wishes Netscape had done differently that would have made the internet better today. I'm excited about how the conversation went and hope you enjoy it too. Let's talk to him.
Mike Maples:
Marc Andreessen, welcome to the podcast.
Marc Andreessen:
Hey, it's great to be here.
Mike Maples:
So Marc, you've been involved with some pretty big wins. So why don't we just start out by asking you, what's your advice to people who want to build something great?
Marc Andreessen:
So the first piece of advice is, don't do it.
Mike Maples:
It's impossible.
Marc Andreessen:
It's impossible. So, it's so hard. Sean Parker has the best line on starting companies: "Starting a company is like chewing glass. Eventually you start to like the taste of your own blood." And so don't do it. And the reason that's the first piece of advice is because number one, if you can be talked out it, you definitely shouldn't do it. So if you listen to advice number one, you definitely shouldn't do it. If you ignore advice number one, you might have the personality type to be a founder. So that's the first gunshot.
Marc Andreessen:
And then you see this a lot, I'm sure, is there are a lot of people who would like to start a company. The goal is to start a company, and then they kind of try to back into an idea. We call these sort of synthetic startups. It's like, I want to start a company, and I am now going to apply myself-
Mike Maples:
Market white space companies.
Marc Andreessen:
Yeah. Exactly. Wouldn't it be great if x, right, where x is just something I read in a magazine that day or just... And by the way, there have been some successful synthetic startups. Some have worked. The more common thing is an actual, honest to God, organic idea that is actually something that you are deeply immersed in. The odds that it's going to be a flash of inspiration are very low, but if it's a field that you've been working in for five or 10 years, and you know it inside out, and it's obvious to you that it should work a different way, then maybe you've got something.
Marc Andreessen:
But you've got to really, deeply... Number one, you have to deeply believe because you have to really be irrationally committed to it because it is so hard, but the other thing is you have to actually validate your beliefs. It has to actually... You have to be right.
Mike Maples:
And Gates and Jobs were the same way with PCs. You're obsessed with the field for its own sake, and you start noticing things because you're down the rabbit hole.
Marc Andreessen:
Yeah, that's right. It's very hands on. And both Microsoft and Apple founding stories, both companies started very small and very humble because they were literally working in... Apple, they were building their first computers by hand. There was no abstraction to it. There was no master plan, there was no theory. It was, can we build a hundred computers made out of wooden boxes and sell them at the trade fair? And it built out of that. But to get right to your point, they were living in that world.
Mike Maples:
I would love to just know about how you got to building Mosaic. How did you get involved with the internet in the first place?
Marc Andreessen:
It was a strange time. There's an old William Gibson quote, the author of Neuromancer, all the great science fiction books, he says, "The future's already here -- it's just not evenly distributed." And so this was a great example of that. So what I sort of semi-realized...
Marc Andreessen:
When I chose to go to Illinois, part of it was because they had a top-flight engineering program, computer science program, but in particular, they were picked in the mid '80s to be one of four federally-funded centers for supercomputing at the time. And this was a specific government program that basically had two parts. One was to basically buy these $25 million supercomputers from companies at the time like Cray, and thinking machines, and basically put them in basically four locations in the U.S. and then for scientists to be able to use for drug design and studying black holes and doing all kinds of stuff. A $25 million computer occupied a full room, and in fact in those days what they would do is they would actually build an entire building for the computer, and then they would actually, they'd build the shell of the building, they'd build the walls, they'd leave the ceiling open, then they'd lower the computer down through the ceiling using a crane to lower it into place, and then they'd finish the building. So this was a big deal at the time.
Marc Andreessen:
These things were super expensive, and so they only put them in four locations, and they wanted other scientists and researchers all over the country to be able to use them, and so part two of the program was this program called NSFNET, the National Science Foundation Network, which was the funding group to pay for this whole thing. So those programs were authorized in the mid '80s, and then I got to Illinois in '89, and so they basically had been rolled out at that point, and so Illinois in those days was like a fully broadband, wired, internet-native campus. And it was one of exactly four of those.
Marc Andreessen:
But just because you have the internet doesn't mean that it's actually all that useful. If you're a scientist, how exactly are you communicating with your colleagues on different campuses? How exactly are you storing and analyzing your research? How exactly are you going to share your research with the world? That stuff was all yet to be built. And so the purpose of this group was to build, basically to build the early systems that would let people collaborate.
Mike Maples:
So then you're building these early systems. Where does the idea for the Mosaic browser come from?
Marc Andreessen:
So, funny story there. So basically, the project I got hired in to work on was a project called Collage, which basically was sort of, I don't know, maybe Zoom's the current comp, or something, or Skype. So it was sort of general purpose, real-time conferencing, so audio and video, but also all the standard complement: whiteboard sharing, collaborative document editing, Google Docs kind of thing. This is like, what, in 1991? 1992? They kind of had all those ideas back then.
Marc Andreessen:
It's one of those funny things, when people talk about this, the role of timing in this stuff is always so interesting. That stuff's all happening now. Zoom goes public now. It's 30 years later. The lag on these things can be quite something.
Mike Maples:
Okay, so you're trying to help these ideas take some early baby steps. How'd you think about what to work on?
Marc Andreessen:
Well, there were a few things. So then you basically say, okay, real time isn't going to work. It was sort of internet variations on the bulletin board systems of the PC hobbyists in the mid '80s. So sort of those ideas being transplanted with the internet, which is kind of what all those social networking has been doing for the last 30 years. There was that, and then there was this new thing called the web, which had been developed in Switzerland at CERN, but that was super early, and there were two or three websites at the time, and the software was like... It was just very early.
Marc Andreessen:
If you were to read an article about the internet at that point, the article would be like, why would anyone want to use the internet? And then a lot of the articles just ended there. But if they would go on to describe why you might use the internet, it was like, well, you can kind of do email, and there's this FTP thing, you can download files, and if you really want to get weird, you can be on Gopher and [Ways 00:08:00]. You can go to Gopher and do menus. You can go on Ways and you can do searches, kind of clay to be molded.
Marc Andreessen:
So basically the originally idea for Mosaic was okay, how about the universal client? The problem is, all this stuff was like, these are all fragmented programs. They were hard to use. They were hard to install. But most of this stuff was open source. You had to be pretty sophisticated to install it. By the way, it seems obvious that everybody's going to have this stuff. It seems obvious that kids are going to want to keep using this stuff once they graduate from college. It seems obvious that there's going to be software on servers and clients. It's going to be used to share information. Why wouldn't you be able to have a link?
Marc Andreessen:
Number one, you don't have a text-based display, because you want to be able to read stuff. Most computer UIs up until that point were not text based. They were like buttons and switches and levers, like old-style nuclear control panels or something, and this software metaphor was like, put everything in a document. That was new, but we thought that was obvious, because you like to be able to read things. Then we thought, okay, hypertext is probably obvious. So there were a set of these things that we thought were obvious, but it's kind of like, if they're so obvious, why haven't they happened?
Mike Maples:
And I remember, was it Vannevar Bush wrote something. He talked about this thing called the Memex, I think, hypertext in the '40s, maybe even, a long time ago.
Marc Andreessen:
So I started to do a lot of research. Luckily Illinois has a big library, so I started to do research. And again, I want to stress again, this is pre-Google, so this is like, go to the card catalog and try to figure out... Can I swear on this podcast?
Mike Maples:
Yeah, it's okay with me.
Marc Andreessen:
Go to the card catalog, and figure out who the f&*k wrote a book about this? Please God, could somebody have written a book about this. So basically the history was, tracing back in time, Doug Engelbart, who basically developed a lot of these ideas and demonstrated a lot of them working in prototype form in the late '60s, and so there was an intellectual heritage going back to Doug. Contemporaneous even before that was Ted Nelson who had sort of invented the term "hypertext," and so there was this idea of linking that kind of existed, but Ted is the guy who kind of picked up those ideas and kind of put them in the idea that a computer should be able to do this. So there was kind of that heritage.
Marc Andreessen:
And then there was, as you alluded to, there was this guy named Vannevar Bush who in 1945 wrote a paper called, "As We May Think," which is one of those great, old-style, 1940s-style titles for papers, and it basically outlined a hypothetical system, this is 1945, so Vannevar Bush is an important guy. He was FDR's science advisor during the development of the nuclear bomb, and so he was a very important guy. He defined, basically, how the federal government would fund research for the next basically 70 years, so he was a very important guy. Pillar of the establishment.
Marc Andreessen:
So he wrote this document, which the Atlantic at the time published back when they used to publish things like this. Today they'd publish a piece about how bad it is, but at the time, they just published a piece describing it. It basically outlined a system called Memex, which is funny because 1945, so it's actually pre-digital computers. Digital computer was a brand-new idea. These ideas go way back. They're fairly obvious fairly quickly. That would be a good idea. The good idea would be to be able to store information, retrieve it, share it, transmit it over long distances, search it. There we were sitting in like 1992 being like, okay. So basically it's like, are we crazy? Is the rest of the world crazy? Or is it just the right time?
Mike Maples:
And it's funny because it's one of, to me, we both know this guy well, Balaji Srinivasan, one of the things that I really like is his notion of this idea maze. I think Chris Dixon talks about it too. To me it's a way better way to come at the question because a lot of the... All of it will happen someday. The question is just, what's the right time for it to happen? And the idea maze is kind of this technique for understanding all the attempts that have occurred in the past, and kind of understanding why those ideas got blocked, and now do they get unblocked all of a sudden. How do you decide what to build first and which ideas are worth pursuing?
Marc Andreessen:
So what we needed to have was a network effect. We needed to have a flywheel kick in. We did the browser and we also did the web server at the same time, and that web server actually now is... Apache is the modern descendant of that web server over the last 25, 30 years. What we needed was kind of a ping-pong effect. What you want is the flywheel where more people reading with the browser leads to more people wanting to publish with the server. The more people who publish with the server, the more incentive there is to read, and then you get the flywheel effect. So what we needed for that, the advantage we had, is we basically had a hack. One of the things we say at our firm is if you're going to start a new network... Network effects are the best businesses in the world, but to get the network effect going, to get through the bootstrap phase, you need some hack. You need some strategy to get you through that initial phase to the flywheel catching.
Marc Andreessen:
So our hack actually was, it was the internet, the sort of NSFNET, [Vivihack 00:12:31] actually was at the time internet news groups. We're actually the carrier. It just turned out, and this is one of those stroke of luck things, it just turned out there was latent demand. There were enough people on this thing. It kind of goes back to this. There weren't online at the same time, but they were online. And so there were enough people online where you could, there was a big incentive to be able to share, even in the number of phone people who were there. And then there was a big incentive to consume if there was anything at all to read. And then this was the prototypical early adopter crowd, where they just like to try new things. So there was a set of internet, and there was groups that were kind of actively baiting, discussing all these kinds of topics, so we were basically able to use that as the carrier wave.
Marc Andreessen:
One of the first blogs was a page I maintained. It wasn't the first one, necessarily, but it was one of the first ones. So we had what was called the "What's New" page. And the "What's New" page, I would update it every, I'd get into work and update it every morning, or say when I got into work late in the afternoon, as the case might be.
Mike Maples:
Probably started out as just, everything's there.
Marc Andreessen:
No, it's literally every new website. And it was literally like, I would just get emails. Okay, I launched a new website that's got the menu of my favorite Indian restaurant in Cardiff. I've got a new website that has lyrics for the R.E.M. songs. It was just random stuff like that. A lot of it was just people experimenting. It started out being one a day. It was like, okay. And then it started being two a day. And then it started being three a day, and then it went to five and ten, 20, 30 a day, and you could kind of... There were two ways to see the flywheel. One was the incoming email. There were two incoming email boxes where you could see the flywheel kick in. One was entries for the "What's New" page, because the "What's New" page was the main distribution method. It was the main way people were finding out about new web pages because this was pre-Google and everything else, pre-Yahoo. But the other email was customer support email for the browser, which was the thing-
Mike Maples:
The ones you didn't want to get.
Marc Andreessen:
The thing that almost killed me was providing customer support for the entire internet. But that's another part of the story later on. So basically, literally what happened was, this whole thing started taking off. More and more people at Illinois started working on this. The teams started to expand. We became more ambitious. We actually went for the second round of NSF, National Science Foundation, funding. All this is happening, I mean, I was making minimum wage, $6.25 an hour, but that $6.25 an hour was being paid for by the National Science Foundation through a very generous grant. I'm grateful for that.
Mike Maples:
At least you didn't get diluted.
Marc Andreessen:
It's hard to dilute $6.25 an hour, in fairness. They say, keep your burn rate low. That's a good way to do it.
Mike Maples:
Mission accomplished.
Marc Andreessen:
So we actually went and we actually, so literally what was happening was we were just, it was working, and so we were the dog that caught the bus kind of thing where it just, what literally happened was the number of customer support emails per day was 100, and then 200, and then 300, and then up and to the right. So we went to NSF for grant number two to basically make this thing real and kind of fully build it out, and of course they denied the grant.
Mike Maples:
We don't want to pay money to take all these support emails.
Marc Andreessen:
Yeah, exactly, which is literally what... At the time I was like, they have a thing that's working, and they're shutting off the money. Okay, that seems kind of dumb. And then I realized, actually later on, I was like, oh, that was actually smart. Specifically, it's a research institute, it's NSF, it's research. It's entirely to fund scientific research, and so clearly customer support emails are like the 14th version of the Mathf.Clamp is not...
Mike Maples:
So maybe that's an initial sign of product/market fit, is when the university says, you know what, we're done with you using our resources here. You've outgrown this university.
Marc Andreessen:
Yeah, exactly.
Mike Maples:
That happened kind of with Google, too, I think.
Marc Andreessen:
Yeah, right. At some point, Google would have taken over the entire computer science department, eaten all the computers. You get kicked out of the nest.
Mike Maples:
So they were kind of pushing you out of the nest.
Marc Andreessen:
So in retrospect, they did exactly the right thing. They did exactly the right thing at exactly the right time.
Mike Maples:
So did you decide, okay, I want to start a company, or were you just like, well that's a bummer. My grant money just went away.
Marc Andreessen:
The second.
Mike Maples:
So now you're just answering customer support.
Marc Andreessen:
No, I was like, shit, I have to get a job.
Mike Maples:
So what year were you at college?
Marc Andreessen:
I was graduating.
Mike Maples:
So you were about to graduate.
Marc Andreessen:
It coincided with graduation. It coincided right around that time, and I was like, well, shit, I got to get a job, and so they were basically, there's very little actually in Champaign–Urbana.
Mike Maples:
So now you got to get a job. You're interviewing for jobs.
Marc Andreessen:
Nope.
Mike Maples:
No?
Marc Andreessen:
Well, I had a slight advantage in my job search.
Mike Maples:
Okay. You controlled Mosaic.
Marc Andreessen:
I controlled Mosaic. Specifically, I controlled Mosaic, which is like, I've got to decide what people saw. So I made sure they saw my resume.
Mike Maples:
Okay, nice. That's a growth hack.
Marc Andreessen:
Yeah, it's a growth hack for your career. You just go in the browser and you add to the browser a button to see your resume. In fairness, I didn't put it on the homepage, which I could've done, but that was a step to far. But I put it in the "About" page. I got a set of offers, some on the East Coast, some on the West Coast. I almost joined the Java team.
Mike Maples:
No kidding.
Marc Andreessen:
In '94. Yeah, in '94 when I first came out here. They offered me-
Mike Maples:
When it was called Oak.
Marc Andreessen:
It was called Oak at the time. This was Java pre-Java, but at the time it was a spin out from Sun, but it was only a partial spin out, so they offered me something called phantom stock options.
Mike Maples:
It's also funny because all of us were thinking about, okay, there's going to be this network-centric future, but most people were framing it in yesterday's metaphors. They were talking about the digital superhighway, and so as a result, people thought the center of gravity was going to be set-top boxes and video game consoles, and I don't know if others had this reaction, but when I first saw the Mosaic browser, it was instant ignition: this is how it's going to play out. I just immediately knew the way we've been thinking about this is wrong. It's not going to happen on these set-top boxes. It may someday, but the browser's going to happen right here, right now. There was just no doubt in my mind.
Marc Andreessen:
What I would nominate on the point that you made, I think the point you made is an important point that we see all the time in our day job, and we see it all the time today, which is, and I'm going to make two dynamics to that. Top down versus bottoms up. And one way versus two way. And these are still battles that are playing out. This is the whole battle on cryptocurrency that's happening right now. You see this all over the industry, all over the world, this battle... And by the way, many global politics are based on this battle right now. So top down versus bottom up, the view at that time, it was the information superhighway, it was going to be sort of set-top boxes, interactive television. But it was going to be provided by big companies.
Mike Maples:
Yeah, or should the government-
Marc Andreessen:
Or the government.
Mike Maples:
They built the regular highway. Should they build a digital superhighway? That'll help us against "Japan" quote unquote.
Marc Andreessen:
Exactly. That was the big thing.
Mike Maples:
That was the thought process.
Marc Andreessen:
That was the country that was going to take everything at the time. So it was going to be top down. The magazines and newspapers, 100% of the coverage, 100% of the commentary, all the media coverage, it was all around, it was going to be the government or if was going to be Time Warner in those days, or AT&T, or Verizon, or the big cable company with the predecessor to Comcast. So it was going to be these giant media telecom companies. And then Microsoft and Oracle, giant software companies, were trying to kind of wedge their way in to kind of be part of it, this top-down thing. They would decide what it was going to be, and then that goes to the second thing, which is one way versus two way. It was going to be primarily video, but a one-way push. So this is why the whole metaphor of 500 channels. The whole thing was, today you've got 14 TV channels, and in the future we'll have 500. That's the information superhighway.
Mike Maples:
And it'll be interactive TV.
Marc Andreessen:
It'll be interactive TV.
Mike Maples:
You can push a button and order a pizza.
Marc Andreessen:
What's interactive is you'll have a remote control. You won't have a keyboard or a mouse or anything unless you get into trouble.
Mike Maples:
You'll just have a button that has a slice of pizza on it.
Marc Andreessen:
Yeah, exactly.
Mike Maples:
We really had those.
Marc Andreessen:
Or you know, like American Idol. It'll all be like American Idol where you can vote for the person who's singing the best, or something like that. But the idea that an individual user was going to be contributing into this environment, the idea that an individual user would be publishing a video, or making a post, or anything like that was just like, setting up a website, there was no incorporation of that kind of two-way idea at all, and so I think what put the whammy on people was if you came from the established power structure, if you came from the big companies, or the press that was used to covering the big companies, it obviously had to be top down and one way, right? By the way, the press was one way at that time. This was before the audience could talk back, so if you were at Time magazine, or NBC news, or the New York Times, you were used to-
Mike Maples:
You controlled the conversation.
Marc Andreessen:
Yeah, exactly. Maybe you published the occasional letter to the editor, but you can strain that shit. You don't let people get carried away. That's one 48th of the space on one page of the paper. So there was that, and all the people who thought they were in a position to decide, had that view. But then all the regular people, at least especially let's just say the nerds, including ourselves, the regular people who were just like, look, I just want to be able to do things. By the way, do things. I want to be able to consume, but I also want to create, and I want to contribute, and I want to build.
Mike Maples:
The thing that I've just seen time and again, it's like when you think about it, when you were at the University of Illinois, you're hanging out with the supercomputers, and you're building a browser, it's almost like the world was thinking in Cartesian coordinates and you were raised to only think in polar coordinates. Your mind was prepared to receive the insight because like you said, you were living in the future. You may not have even known it at the time that you were living in the future, and like so much of entrepreneurship I've found, it's like noticing. You're living in the future, and you notice something, and you solve your own problem, and you're not necessarily trying to get rich at the time. You're just like, I'm working on cool stuff, and to do more cool stuff I've got to build this thing.
Marc Andreessen:
What I've found with myself, or I've found with the other founders who are like what you're describing is, it's just obvious. It's just like, oh, obviously this should be this way. But then there's cognitive dissonance. If it's so obvious, why hasn't everybody figured this out yet? Now most of the time when people come to that conclusion-
Mike Maples:
They're just wrong.
Marc Andreessen:
They're just wrong, or insane.
Mike Maples:
They have a vision, but it's like a hallucination.
Marc Andreessen:
Exactly. But every once in a while, you've got somebody who really does decode something, and I think to your point, on kind of the preconditions for it. Some of it is, you get to see the early kind of pings. There're pings from the future. You can see these things actually running today. You get in a position where you can see something like that, so that's part of it. But the other thing that happens is, it's the people who get to operate with the new assumptions.
Mike Maples:
Like when you're your age, starting Netscape, you don't have to translate. People who lived in Cartesian coordinates, they have to translate to go to polar. So people who lived in the world of tops-down, one-way digital superhighway, interactive TV, they had to translate that to what you were already doing, but to you it was just obvious. It's like, there wasn't anything for you to unlearn.
Marc Andreessen:
And then on top of that, I had no power. I was not the CEO of AT&T. I couldn't do any of the stuff that all these fancy people could do. I didn't have that level of control. Anything that I was going to do was going to have to be bottoms up. It was going to have to be two way, by definition. If it wasn't two way, I'd be blocked out.
Mike Maples:
So then speaking of powerful people, how did you find Jim Clark? How did you guys connect?
Marc Andreessen:
This is where I got really lucky. Jim Clark, kind of quick recap, Jim Clark at that time, Jim Clark had been the founder and the original CEO of Silicon Graphics, which was at the time the companies... As you said, you were there. The company at the time was probably most analogous today to, I don't know, some combination of Google and, I mean, it was THE company. It was the company that all the smart... It was like if you were a smart person in the computer industry, it was the company you either worked for or wanted to work for. It was like the brain center of the industry. This was when they really drove computer graphics to be what they are today.
Mike Maples:
A total gee-whiz company. Best graphics engineers in the world. Best networking engineers in the world because you have to push the pixels over network. It was just an incredible place to be.
Marc Andreessen:
Fortunately for me, Jim had a problem. And the problem was, Jim had grown very dissatisfied with the state of affairs at Silicon Graphics at the time. He got frustrated by a number of things. He left the company. He decided to start his second company, but he had a very specific problem, which was that he had a nonsolicit at SGS, so he had spent the previous 15 years hiring all the smartest people he knew in the world into SGI, and now he couldn't take them with him. So he literally had a, what's that called, warm meat problem, in that he didn't have any bodies to work with him. So he literally went out to a whole bunch of people in the industry who he knew who weren't at SGI, and he talked to them about maybe do you want to start something or work on something? And then he met me through a friend, a mutual, actually a guy he worked with at SGI who I didn't know, but who knew about me. He was actually one of the guys at SGI who actually is responsible for all the demos. They were a famous demo company.
Mike Maples:
That's how we sold our computers.
Marc Andreessen:
So the guy who ran the, I think designed the demos or built the demos for the briefing center, was basically up on all the leading-edge stuff because he was the demo guy, and so he basically knew about all this stuff, and apparently happened to mention to Jim that A) that I existed, and then B) that I had just recently moved out here. And so I get a random call from Jim Clark one afternoon being like, Hey.
Mike Maples:
Did you know who he was?
Marc Andreessen:
Oh yes. It's literally like Steve Jobs calls you, or you know, Larry Page called you. Hey, this is Larry Page. Would you like to talk about starting a new company? You're like, okay.
Mike Maples:
You think?
Marc Andreessen:
All right. Gee, I don't know, let me check my calendar, right? I'm basically, sure, name the time and place.
Mike Maples:
Did you guys get on pretty well in the early days?
Marc Andreessen:
So a couple things happened. So one is to my enormous shock the other people he was talking to were too risk averse, and I should also, this was during a very sort of down period in Silicon Valley. This was not an exciting time. This was like '93, '94. There had been a really big recession and a lot of companies had failed, and so there were a lot of people who just wanted a job. But he did not get as nearly a positive response as I would have imagined from a lot of people who had actual, had careers, let's say. So through process of elimination, in part, I think it came down to me. And then he and I started brainstorming. And then we started working together on plans.
Mike Maples:
So he wasn't saying, "I want Marc Andreessen because he invented the Mosaic browser." It's like, I can't get any of these killer SGI engineers, he's maybe another new smart guy that I can get and not get sued. That's pretty awesome.
Marc Andreessen:
I think that's part of it, and this is probably just channeling Jim, at the very least, he knows how to build something new, which not everybody does. So at least he's done it once before. But the reason why the Netscape idea was not obvious is because even after all of that, it still wasn't obvious that the internet would be a business. And part of that was, it wasn't a business. Nobody had made it into a business. It was just this thing...
Marc Andreessen:
It goes to the top down bottoms up thing, which is just like, even after all that, and I saw the adoption cycle and the whole thing, and I stayed on all the mailing lists, and I saw everything, it was just like, I don't know. This is all a thesaurus. There's no commercial transactions on the internet. It had been illegal to do commercial transactions on the internet until 1993. And this was only the spring of 1994, so there was no eCommerce. There was no Amazon. There was none of that stuff. There was literally nothing to buy. There was no money. There was no nothing. There was not an obvious business to be built.
Mike Maples:
And I remember for a couple years, Upside magazine at the time, and things like that would say, when's the internet going to have a business model? For about two years, people said, nobody's going to have a business model.
Marc Andreessen:
This is absurd, everybody knows this.
Mike Maples:
Even Yahoo people were saying, well great, but no business model.
Marc Andreessen:
Google had no business model, like all these things. None of these things had business models. Right, exactly. Jim had been enmeshed, Silicon Graphics had gotten enmeshed in two areas that goes back to the conversation we were having about interactive TV. So they were actually the provider of the technology for the Time Warner interactive TV project, which was the most viable... It was the most actually developed version of the top-down information superhighway thing at the time in Orlando, Florida. This was written up in all the magazines and newspapers at the time. It was a really cool system that's analogous to what you have today on a modern whatever Comcast or Direct TV set-top box, or something Netflix kind of experienced, in 1994. It was impressive. It was like the capital cost per house was like, I don't know, $16,000?
Mike Maples:
Yeah, it was crazy. We had some of our high-end gear to power those boxes.
Marc Andreessen:
It was not... So we basically set out, plan number one was to build the software layer for interactive TV, and then we basically realized, oh shit, there's not going to be actually any interactive TV. The economics actually don't work. And then the other decision was, SGI was building the graphics chip for the new Nintendo at the time of the Nintendo 64, the first game console with 3D graphics. So the idea basically was to build what today you'd call Xbox Live, or PlayStation Live to basically the network for Nintendo 64. The problem was Nintendo 64 was not going to ship for another two years.
Marc Andreessen:
So literally we got to the point where we were like, those two plans don't work, and then it was literally like, what are we going to do? Process of elimination. Okay. So if top-down interactive TV isn't going to work, and interactive gaming at that point wasn't going to work, then what's left? What's left is the internet. So it's like, process of elimination. This thing that nobody's taking seriously, that nobody thinks can be a business, that breaks all the rules, that's bottoms up, that's organic... By the way, messy, and hackers, and crime-
Mike Maples:
And you probably don't even think it's a business yet.
Marc Andreessen:
No. Yeah. I have no actual business experience at that point, so I have no basis to evaluate anything in business, but however, it seemed like a stretch. If it's the only thing, then it's going to win. Because the only thing-
Mike Maples:
Is the thing.
Marc Andreessen:
... is the thing. My favorite. It was weird because Jim was like, oh, that makes total sense. What Jim actually realized, to his credit, he was so, SGI was so powerful at that point, that he was in that top-down world, and I think he would say, he was so fully in it and of it that he thought that was how the world was going to work, and he was doing his best to make it work that way when he was at SGI. But then he's got such amazing, he's got a mental flexibility that's-
Mike Maples:
Incredible.
Marc Andreessen:
... extraordinarily rare to have somebody who's this flexible in their thought process where he was just like, when he got into this new context, he was kind of just shedding assumptions. He was able to replant himself into a true start-up context, and able to shed all his assumptions and say, okay, from a standing start, what would you do? And come to a completely different set of conclusions.
Mike Maples:
Yeah. So then you decide to start this thing, and you raised money from... Well, he seed funds it at first. He's probably already been seed funding just your little vision quest.
Marc Andreessen:
No.
Mike Maples:
No. Okay. So you're just on this vision quest by process of elimination.
Marc Andreessen:
Nights and weekends.
Mike Maples:
It's like, okay, we're doing this.
Marc Andreessen:
Yeah.
Mike Maples:
Then you just go raise money from John Doerr Kleiner Perkins, or?
Marc Andreessen:
We were running about six months with Jim's money, and then he's like, I'm rich, but I'm not that rich, and so he was like, we're at a time where we need to go raise money. So Jim zeroed in on Kleiner Perkins because KP had actually backed Sun, which was SGI's big competitor, and he told me he had always really respected how... Both companies ended up being very successful, but Sun had a much faster takeoff rate out of the gate, and then he had always respected how John went about being a board member at Sun.
Mike Maples:
Okay, so you raise some money for Netscape, not sure if it's a business yet, but hey, let's go for it. This must be around the time when Mosaic has started to see even bigger lift, because I seem to remember the summer of '94, it was starting to really become a thing.
Marc Andreessen:
So it was just a snowball rolling down the hill, picking up speed. And it was starting to mainstream. So you were starting to get the first signs of consumer sector coming on the thing, which is normal people, which is a big deal at the time. You're starting to have companies starting to start up and launch websites. So it was around that time, I think it was around that time that AT&T ran the first internet ad on at the time, because Wired had created a website, and they ran the first internet ad. The reason we have all these banner ads is because the first ad was a banner ad for AT&T on the wire.com.
Mike Maples:
When did you switch from Mosaic to Netscape?
Marc Andreessen:
Mosaic had been the name of the project at Illinois. We didn't bring any of the code with us. The code was open source, but copyright University of Illinois, and we needed to rewrite it anyway because we needed a bunch of stuff in the code like security that we didn't have, so we wanted to do a clean rewrite with what we knew now, but we did figure, it's the name of the research project. I knew the name "Sun" was actually named after the research project at Stanford that spawned Sun was actually called the Sun research project.
Mike Maples:
Stanford University Network.
Marc Andreessen:
Exactly. It's just the name of the research project. It's a free thing, so obviously we'll just call this thing Mosaic Communications. We'll figure out the product name later. University of Illinois then did a very sort of clever thing I had not seen before, and I don't think I've seen it since, which is they didn't sue us. Instead, they sent lawyers to all of our potential customers and told them they were going to sue us.
Mike Maples:
Ooh.
Marc Andreessen:
SO they freezed us in our tracks. They basically blocked our ability to do business, because they alleged a broad range of trademark and copyright violations. The copyright violations weren't true because we didn't copy any of the code, but they threatened to sue us for that. And then we had this problem, which was, we had this name, and there's ambiguity as to whether there was a trademark on it or not. But there did seem to be a clear-
Mike Maples:
Hard to explain away.
Marc Andreessen:
Yeah. Inconvenient fact, as they say. And so we actually ended up suing them for I believe it's restraint of trade and purchase interference.
Mike Maples:
Are you feeling kind of pissed at them right about now?
Marc Andreessen:
Yes.
Mike Maples:
Thermonuclear super pissed.
Marc Andreessen:
Right now? Right about now, like today, sitting here?
Mike Maples:
Yeah. Even now?
Marc Andreessen:
Very much so, yes. Yes, I'm still extremely angry. Thank you for asking.
Mike Maples:
How does that make you feel, Marc?
Marc Andreessen:
Yeah. Incensed. We'll pick that up in part two, the cycle of therapy section of the thing. So we actually ended up suing them. We actually ended up suing them, and then we negotiated a settlement. We paid them, and as part of the settlement, we changed the name.
Mike Maples:
Okay.
Marc Andreessen:
And that was the last penny.
Mike Maples:
So you changed it to Netscape. So then you do the Netscape browser. Did it just immediately blow up? I mean, everybody knew who you were, everybody knew that you were the voice of browsers. It just...
Marc Andreessen:
Yeah. It was basically continuation of the Mosaic phenomemon, and we were the clear inheritors of that because we had built it, and so it was just one of these things where it was a cold start as a company, but it was building directly on the momentum from the previous thing. We knew, what else did we knew, we knew Illinois was not going to continue the Mosaic project because we knew they didn't want to do it.
Mike Maples:
They didn't want to.
Marc Andreessen:
Right, exactly. And so we kind of knew it had to be picked up. And then look, the other thing was, there was just a set of things that you just needed. It was time. It was time to be able to do financial transactions. You needed a cushion, which the original browser didn't have.
Mike Maples:
So then was there any sort of palpable moment where you're just like, holy crap, this is blowing up, or had it kind of already blown up even before Netscape really got started? Was there any moment where you're inside of Netscape and you're just like, holy sh*t?
Marc Andreessen:
Well the big moment was the night of the original release of the browser where we hooked up one of the computers to the stereo system and had that be the cannon fire sound effect for every time somebody downloaded it. The cannon started to go off. Before long the cannon was going off continuously, right?
Mike Maples:
Like how long?
Marc Andreessen:
It was a few hours. But again, it was feeding on this momentum. It was like everybody, we knew everybody who was using Mosaic, we knew how to get to them. We just said, hey, there's this new thing. It was much better. I mean, it was built correctly. It was our second implementation. We knew it was better.
Mike Maples:
So now it's probably what, this is probably late '94?
Marc Andreessen:
Yeah.
Mike Maples:
Okay. If I remember, you went public in early '95?
Marc Andreessen:
Yeah. August.
Mike Maples:
With no profits. Did you have any revenue yet?
Marc Andreessen:
So we had revenue. We doubled our revenue every quarter that year. So the revenue that quarter was $5 million, $10 million, $20 million, and $40 million. And we went cashflow positive right around the time we went public. I think we were, if I remember correctly, we were cashflow positive continuously all the way to when we sold the company. One the legends, myths that built up around the company is that it was this early precursor for these unprofitable companies.
Mike Maples:
Really?
Marc Andreessen:
And actually we prided ourselves at the time of delivering cashflow basically through the whole thing.
Mike Maples:
So was the decision to go public pretty obvious, then, at the time? It was just like, your revenues are exploding, people want the product...
Marc Andreessen:
It was obvious to Jim.
Mike Maples:
What's it like? So you're just barely out of college, and all this stuff's happening around you. What's that like?
Marc Andreessen:
It was literally, we were so heads down. It was just like, go work some more. It was one of those things. There was so much to do, there was so much to do. It was building a company, which is incredibly hard, but then on top of that, the whole thing started to work, and then it's just like you have a thousand ideas.
Mike Maples:
And a few more customers support calls.
Marc Andreessen:
And all that, yeah. All this other infrastructure to build. I'm learning basically all the business on the fly. I'm basically learning I don't know whatever is in an MBA plus another 10 years of operating experience as fast as I possibly can. So I'm basically either at work or I'm at home reading business books. Those are the only two things that I did.
Mike Maples:
Now you know about the theories of product/market fit and all that stuff. How do you reconcile what happened at Netscape with the notion of product/market fit? Because it just seems like one of these rare cases where it's just like lightning just struck, and it was huge right away.
Marc Andreessen:
I think for B2B, there are deterministic ways to get to product/market fit. With consumer stuff, it's less clear to me, even still. We actually use the term "lightning strike." It may just be... Well, here's one of the questions. There are these companies, there're quite a few companies that have had lightning strike consumer hits in the last 20 years. That led to the creation of very interesting companies. And there's probably, I don't know, in the U.S. alone, 50 or 100 or 150 of those, right. How many of those ever had another one?
Mike Maples:
Very few.
Marc Andreessen:
In the same company that they didn't have to go acquire from outside? And I think it's Apple, maybe zero.
Mike Maples:
Now people talk about the internet as being easy for hackers to get to, fundamentally insecure as an architecture, or needs to be rethought in some ways. Knowing what you know now, are there things that you think you could have done, or that Netscape could've done, that may have made it play out a little differently?
Marc Andreessen:
Yeah, so there's three big things that we should have done. Would have made a big difference early one. One of them... Well, there's one thing we did that mattered a lot, and I'll talk about that. There's one thing we tried to do and couldn't get there, which remains a hot topic today. And there's a third thing that didn't even occur to us, which I kick myself we didn't think about this, but I'll explain why we didn't think about it. So the thing that we did do is we got encryption in there.
Mike Maples:
That's huge.
Marc Andreessen:
And that was a fight then. By the way, it looks like it's going to be a fight again now. The various Western governments are once again pushing to try to restrict the use of strong encryption. And people don't know the history of that. We fought that battle, and Netscape Navigator was the first commercial implementation of encryption that became widely used. There had been other products before, but we were the first one that millions of people used. At the time we developed Netscape to have strong encryption in the browser meant that the browser was classified under U.S. federal law with criminal penalties as ammunition. It's classified in the same export control category as tomahawk missiles. We were not allowed to export the version with strong encryption. We could sell it in the U.S. We couldn't export it. We had to export a weakened version. Our sales pitch to a user in France or Australia or something is like, hey, congratulations, you get the one that's easy to crack.
Mike Maples:
Okay. The U.S. government can have their way with it. I hope you like it.
Marc Andreessen:
Right. Exactly. NSA has it prewired effectively because they can just crack it easily. So we fought really hard in the '90s, the encryptions wars at the time, and ultimately the government actually backed down in '97 or whenever they actually changed policy, and they actually legalized, essentially legalized strong encryption globally for U.S. companies to be able to build products globally. That battle keeps getting refought, and probably in my view it's come back to life. I think it's just an absurd thing to be... Do you want secure systems or not? Can somebody please decide.
Mike Maples:
Right.
Marc Andreessen:
If you don't want secure systems, fine. I guess we'll stop trying to build them. I guess they'll get built in other countries and the U.S. tech industry will not be relevant anymore. So that's an option. But if you want American companies to win at tech, maybe we should be able to build secure systems. If you want us to be able to protect and defend the United States against cyberterrorism and criminals and all this stuff, presumably that's a good idea.
Marc Andreessen:
Anyway, we got that one in there. The one that we tried to do was integrative financial transactions. So payments, obviously, was something that you would want. We tried to build that in. We tried hard there. The problem there is obviously money, payments, transactions have been historically highly regulated. We made arguable the mistake in that case of asking for permission, which is we went to the banks and the credit card companies.
Mike Maples:
You should have let the pirates go crazy.
Marc Andreessen:
Probably. The problem was, we didn't have the technology yet. We didn't have Bitcoin. We didn't have cryptocurrency. This was pre-cryptocurrency. Had we let the pirates go crazy, we would have just had to implement a system that just let you transfer money with no permission, and probably we would have gotten nuked for that.
Mike Maples:
That would have even been worse than encryption.
Marc Andreessen:
Yeah. People figured out how to bootstrap a system a few years later, but they almost died. They came close to dying. They almost got regulated out of existence. They just barely got through it. And so that was the big one that we should have had and we missed. But we did try. And that's one of the reasons why I'm so strong now, so positive on cryptocurrency and blockchain and Bitcoin and Ethereum, and all these things today. It's just like, internet-scale money and trust needs to happen. It was a huge problem that we didn't have it early. Had we had internet-scale money and trust wired in early, the internet economy today would not be based in advertising. It would not be based on any of this privacy stuff that people are worried about today. I would be based on money and trust. And it would be a fundamentally better, stronger system. Our view is, with cryptocurrency, we have a chance to go back and kind of redo that. Now of course they're trying to, in various forms, keep that from happening as well. "They." But we're going to try.
Marc Andreessen:
And then the one that I wish we had had, but it didn't even occur to us is real names, real identities.
Mike Maples:
Oh, interesting.
Marc Andreessen:
Which is the other part of trust. Okay, who are you dealing with? And so all the issues, you know spam and fraud and all these issues, abuse and harassment and all that stuff, you can't solve any of that stuff if you don't have real names. It's hard enough to solve those problems when you have real names. It's impossible to do it if you don't have real names, I think.
Mike Maples:
Okay. Thanks, Marc.
Marc Andreessen:
Good. Thank you, Mike.
Mike Maples:
It was great talking to you.