Marc Andreessen, the Man Who Makes the Future
Google Chrome OS, for example, is a fully browser-based operating system, while most of our favorite applications, from email to social networks, now live entirely on the network.
Anderson: A quote of yours that I’ve always loved is that Netscape would render Windows “a poorly debugged set of device drivers.”
Andreessen: In fairness, you have to give credit for that quote to Bob Metcalfe, the 3Com founder.
Anderson: Oh, it wasn’t you? It’s always attributed to you.
Andreessen: I used to say it, but it was a retweet on my part. [Laughs.] But yes, the idea we had then, which seems obvious today, was to lift the computing off of each user’s device and perform it in the network instead. It’s something I think is inherent in the technology —what some thinkers refer to as the “technological imperative.” It’s as if the technology wants it to happen.
Anderson: As in Stewart Brand’s famous formulation that “information wants to be free.”
Andreessen: Right. Technology is like water; it wants to find its level. So if you hook up your computer to a billion other computers, it just makes sense that a tremendous share of the resources you want to use — not only text or media but processing power too — will be located remotely. People tend to think of the web as a way to get information or perhaps as a place to carry out ecommerce. But really, the web is about accessing applications. Think of each website as an application, and every single click, every single interaction with that site, is an opportunity to be on the very latest version of that application. Once you start thinking in terms of networks, it just doesn’t make much sense to prefer local apps, with downloadable, installable code that needs to be constantly updated.
Anderson: Assuming you have enough bandwidth.
Andreessen: That’s the very big if in this equation. If you have infinite network bandwidth, if you have an infinitely fast network, then this is what the technology wants. But we’re not yet in a world of infinite speed, so that’s why we have mobile apps and PC and Mac software on laptops and phones. That’s why there are still Xbox games on discs. That’s why everything isn’t in the cloud. But eventually the technology wants it all to be up there.
Anderson: Back in 1995, Netscape began pursuing this vision by enabling the browser to do more.
Andreessen: We knew that you would need some processing to stay on the computer, so we invented JavaScript. And then we also catalyzed Java, which enabled far more sophisticated applications in the network, by building support for Java into the browser. The basic idea, which remains in force today, is that you do some computation on the device, but you want the server application to be in control of that. And the whole process is completely invisible to the user.
Anderson: Unlike with Mosaic, where your original ideas were proven correct within a year, it seems like this idea has taken 15 years to come to fruition.
Andreessen: Right. And only with the arrival of tablets and smartphones, really. If you draw a pie chart of all the personal computing devices in use, smartphones and tablets are now over 50 percent and growing rapidly. It took a lot longer than we expected, but these really are the network computers. Now, in an ironic twist of fate, the devices do have all these local apps …
Anderson: Well, exactly.
Andreessen: … but I can go on an iPad or an Android smartphone or a Linux tablet and I can access all the same websites and all the same applications and all the same services that I get on my desktop.
Anderson: But we do still have lots of desktops and laptops out there. Let me ask you in 2012: Do you still think that the web and browsers will render computer operating systems a “poorly debugged set of device drivers”?
Andreessen: I will pull a full Henry Kissinger and answer a different question. The application model of the future is the web application model. The apps will live on the web. Mobile apps on platforms like iOS and Android are a temporary step along the way toward the full mobile web. Now, that temporary step may last for a very long time. Because the networks are still limited. But if you grant me the very big assumption that at some point we will have ubiquitous, high-speed wireless connectivity, then in time everything will end up back in the web model. Because the technology wants it to work that way.
Idea Three: 1999 - Web Businesses Will Live in the Cloud
Idea Three: 1999
Web Businesses Will Live in the Cloud
In September 1999, Andreessen cofounded Loudcloud, a firm that would enable whole businesses to move into the cloud; it would host and manage their web services and software so that companies wouldn’t need to run any servers at all. That business didn’t last — despite an IPO in 2001, Loudcloud changed its name and business model in 2002 and was sold to Hewlett-Packard in 2007. But its vision has been vindicated in the phenomenal rise of Amazon Web Services, which serves as the backbone for hundreds of thousands of businesses online.
Anderson: With the name Loudcloud, did you make the first use of the word cloud in this context—as a place where applications run on the network?
Andreessen: It was a common term in the telecom business. AT&Tused it to talk about their Centrex service, which — going way back here — took all the hassles of switching phone calls out of the individual enterprise and turned it into a service. So our idea with Loudcloud was to offer a similar proposition, but for software. When we first announced it, I described it as Silicon Valley Power & Light.
Anderson: Tech companies would use it as a utility.
Andreessen: Exactly: the software power grid. We actually used the electrical metaphor more than the telecom metaphor. When electricity first came to factories, every factory had its own generator. But eventually that didn’t make any sense, because everyone could draw electricity off the grid. At the height of the first dotcom boom, we saw the exact same thing happening in Silicon Valley. You’d raise $20 million of venture capital, and then you’d have to turn around and write $5 million checks to Oracle , Sun, EMC , and Cisco just to build out your server farm. It was literally like everybody building their own electrical generator over and over again.
Anderson: You were the first company to provide software as a service.
Andreessen: I would say we were the first cloud provider in the modern sense of the term. Our pitch was, you should be able to buy all this software by the drink, instead of having to shell out for the bottle up front. By capitalizing on economies of scale, Loudcloud could provide higher levels of service than you could get in-house, and a startup could get its product to market almost instantaneously. It could spend its time and energy building the actual product instead of trying to figure out how to host it and keep it live. That was the pitch.
Anderson: It didn’t really work.
Andreessen: Well, it worked beautifully right up to the point when all the startups went bankrupt, and then all our big clients decided they didn’t have to worry about competing with the startups anymore. After that, it went completely sideways. Literally every other company we were competing with went bankrupt; we were the only one that got through it. So we just went back to basics and we said, OK, we couldn’t make it work as a service provider, but we think we can make it work as a software company, selling the back-end software to manage big networks of servers. We changed our name to Opsware. That ultimately worked, as a business.
Anderson: You were acquired by HP for $1.6 billion.
Andreessen: That whole transition happened during an unfun time in the tech economy. Everybody went through a crisis of confidence between 2002 and 2006. Up and down Sand Hill Road, VCs would refuse to fund consumer Internet companies, because it had been decided that those simply weren’t going to work.
Anderson: Looking back, it’s somewhat ironic that you started with the right name, Loudcloud, but abandoned it. Now the world has come back to cloud. What did it take?
Andreessen: In retrospect, we were five or six years too early. Besides the rebound in the startup economy, there have also been two huge developments in server technology. The first is commoditization: We were running on expensive Sun servers, but now you can buy Linux servers at a fraction of the cost.
The second is virtualization, which makes managing the servers and apportioning services to clients far easier than was possible back in 1999. And that’s why Amazon’s cloud service has been so magical. It’s the same core concept — but with supercheap hardware, which makes the economics far more attractive for everybody, and with virtualization, which makes the entire environment far more adaptable.
