Ray Ozzie: Churchill Club

Remarks by Ray Ozzie, Chief Software Architect, Microsoft, on the Potential of Cloud Computing
Churchill Club
San Jose, Calif.
June 4, 2009

MARK LEAHY: Good evening, everyone. Welcome. My name is Mark Leahy. I am a partner at a local law firm named Fenwick and West, and I’m also the Chairman of the Board of the Churchill Club.

Our program this evening is called The Potential of Cloud Computing. We are honored to have with us Ray Ozzie, Chief Software Architect of Microsoft, and Steven Levy, Senior Writer with Wired magazine. Thank you both for being here tonight. (Applause.)

We have two lead sponsors for our program tonight. The first is NetApp, a leading provider of data storage, management and protection. Thank you very much NetApp. And our second sponsor is Microsoft, a worldwide leader in software, services, and solutions that help people and businesses realize their full potential. Microsoft, thank you very much. (Applause.)

So, briefly, about two upcoming programs, on June 18, we will be having our annual Good Wine and Good Company Program. This year it is about the Wine Wide Web, and how the net is effecting wine appreciation appreciation. Five or six premium wineries will be on hand to pour. Next, on Tuesday, June 23rd, it’s the program about The Future of Security with Dave DeWalt, CEO of McAfee and others. As always, you can learn more about us if you go to the Web site of the ChurchillClub.org, learn about our programs and membership. You can also friend us on Facebook. You can join our group on LinkedIn. And you can follow us on Twitter. And you can view our videos on YouTube, and at BoardTV.

Now, I’m pleased to introduce our moderator, Steven. He is one of the most well-known and respected journalists and authors of our time. He began following and writing about technology in 1981. He worked for Newsweek for about 13 years, starting in 1995, and a year ago joined Wired magazine as a full-time writer. His work has also appeared in The New Yorker, in New York Times Magazine, Harper’s, and Premiere. Steven has written six books, including the critically acclaimed “ Hackers and Crypto”, and most recently wrote “A Perfect Thing,” a book about the iPod.

Before focusing on technology, among other things, Steven was a rock critic. He covered sports. And he found Einstein’s brain, I’m told. That’s right, in 1978, he literally tracked down Einstein’s brain that was in the possession of a pathologist in Wichita, Kansas. Tonight, he will hack into the brain of Ray Ozzie in a conversation about cloud computing, and possibly a few other things.

Please welcome our guests tonight. (Applause.)

STEVEN LEVY: Well, thanks very much, Mark, and thank you, Ray, for coming up here. I suspect it’s a little easier to find your brain here, and we’re going to talk as the topic indicates, a lot about just the cloud here. But I hope we’ll get to a couple of other items as well. And, of course, then there are questions from many of the folks.

So, the cloud is being a term that’s being tossed around endlessly now. And one CEO who is really not given to understatement said this whole hype is total gibberish, it’s a meaningless term, but people use it anyway. So, I’ll give you a chance to clarify things. When Microsoft thinks of the cloud, what does it mean, what do you talk about when you talk about the cloud?

RAY OZZIE: That is kind of tough because we in this industry tend to latch on to these things, the peer-to-peer era, and things like that. What it really means to me, what differentiates it, is this self-service, on demand, way of accessing computational resources, storage resources, things like that, at a virtualized abstraction so that it’s relatively homogeneous. We tend to build it on federated infrastructure, so that it’s out there, it’s a public service. Many partners will do it. Smaller providers will do it. But the reality is, it is different from the developer’s perspective because you can treat it as a utility, you can treat it as a computing utility.

My first experience with computing was prior time-sharing servicing in Chicago on a teletype. It was utility computing. IBM, you know, was CTCMS, and VM-370, you know, pioneered virtualization. These are technologies that have happened before. But right now, in essence, this pendulum is swinging.

STEVEN LEVY: But it can’t be the same as saying, oh, this time sharing is the cloud. It’s just a front view. There has to be an above view. We’re really talking about this connected world where information is somewhere not easily identifiable at least to the end user. What difference is it going to make for the users?

RAY OZZIE: And this gets into why I think we confuse this, because it can really be anything that we want it to be in this era. We can start ultimately with the user experience and ask, ‘How will our user experiences be transformed in a world of cloud computing. Within the enterprise, how will enterprise infrastructure such as e-mail and document management, how will enterprise applications be transformed in a world of cloud computing? And then for developers, how do we experience it.’

I think it is fundamentally transformational. I think the fact that even though the paradigm kind of goes back to old, the fact that we have ubiquitous high bandwidth right now, the fact that devices are so cheap, the fact that we can get storage and computation, and build these massive data centers gives us as architects the ability to imagine a solution, imagine the right kind of user experience that we’re trying to do, and piece it up with the right thing delivered to the client, cached on the client, synchronized across clients, and the right piece of it in the cloud, bringing all those experiences together.

STEVEN LEVY: Well, before you came to Microsoft in 2005, you founded a company called Groove.


STEVEN LEVY: That was a very cloudy company. People networked, and had groups, and they met somewhere, where no one will, in the cloud.

RAY OZZIE: That’s right.

STEVEN LEVY: So you get to Microsoft, and was Microsoft sufficiently cloudy for you? Did you feel you had to block the sun?

RAY OZZIE: The hailstorm had passed. (Laughter.) When I got to, as you said, I had the opportunity in ’97 when I left IBM to kind of return to zero. When you’re fortunate, and you’ve had a successful product you get caught up in everything that it takes to make it successful. Ultimately you end up doing a lot of customer work. And then I decided that I needed to get back to technology. I returned to zero and Groove was built in ’97 for about eight years until we were acquired. And it was born to be Internet. And it took a contrarian approach, as opposed to using Web servers as the core infrastructure for how it was built. It was a purely peer-to-peer system, ultimately augmented with cloud servers, and enterprise integration servers, and management servers, and things like that. But, the great thing about it was I did have a chance to kind of get into the ethos of what is the net, and ultimately what is the Web.

By the time I got to Microsoft, respectfully, they were very busy working on things that would ultimately become Vista, and Office 2007, a large part of the company. But, I felt as though it was kind of like back to the future in many ways. There was a lot of PC thinking, the PC was still the center of most of – of how most people thought about things, and it was a little scary, because by that time I had a perspective that there is this transformation happening. I still think the PC is amazingly relevant, but it’s the connected PC, it’s the PC connected to the cloud, connected to other PCs, PC connected to phones, TVs, and so on. So, I worked with Steve and Bill on a plan to change management, basically.

STEVEN LEVY: So, change management, to manage change in Microsoft to accommodate that ethos, you mentioned?

RAY OZZIE: Yes, it’s basically slowly but surely in ways that you do change management in a large organization, communicating with groups, trying to help them understand not just infrastructurally what has changed in the environment, but how customers – how they want to consume the solutions that they do, the way that they themselves are transforming, and how they want to see that technology adapt to what they do.

And in many ways kind of bringing people through the transition of what it was like to develop mass-market software. Microsoft is the best in the world at developing and delivering mass-market, packaged software, and all that that implies. And that model emerged at a time when we wrapped up floppy disks in boxes, submitted them into a distribution channel. They got to users, we got feedback, asynchronously through user groups and so on. For enterprises that is still the perfect model, because they don’t like to consume things at a very, very rapid rate. But for consumers now it’s a real-time, interactive environment. You can see what every customer is doing on a real-time basis. So, meaning that transition, it’s been a lot of fun.

STEVEN LEVY: Well, mass-market, packaged software – 

RAY OZZIE: Different market segments consume – want to consume value in different ways. Enterprise requirements are dramatically different than the requirements on the outside. And every type of software you deliver to organizations is somewhat different than things that people are used to on the net. So, for example, you can get – you can look at something like Facebook, or Twitter, and these are innovations in the communications or social media realm, but they’re dealing with technology and how consumers and social dynamics interplay with technology.

When you throw organizational dynamics into the mix, communication tools in organizations are dramatically different, because there’s a chilling effect if you say something bad in an organization. There are actually repercussions to saying something bad. The – many businesses are highly regulated. They have – their companies are accountable for the behavior of their employees. So the nature of the tools is different. You can’t update systems at many, many major retailers when you’re approaching the holidays. You don’t want to go to a professional services firm around tax season and suddenly find that the user experience has changed.

Those are great pluses for us. It’s great business. We want to be sensitive to that. But, we also serve consumers, we also serve individual information workers on the net. And many of the tools that people use are the same across both. So the opportunity is to figure out how you can build software and serve those enterprise requirements, while at the same time delivering perhaps a different version, a different packaging of that same value in a way that’s more contemporary, and embraces the Internet.

STEVEN LEVY: I don’t want to leave that ethos remark, because it appears that what you’re saying, and I think I learned a little bit about this when I did a profile of you last year, it’s almost as if there is an echo to the cloud by the net, which was going to help Microsoft in a way change its culture and refresh itself.


STEVEN LEVY: Out there, by adopting those values, those cloud values.

RAY OZZIE: Yes, absolutely. And some of those cloud values are – probably what’s core, what’s most, most important is interoperability and openness of data formats. I mean, it is the lifeblood, it’s the DNA of getting systems to connect with one another, whether it’s an enterprise – in an enterprise data center, interoperability for systems integration is extremely important. But, on the net, the net has been shaped because of independent implementation of different packages, and exchange of information between them, of being conservative in what you produce, being liberal in what you consume. These are kinds of things that are fundamental to the organism that is the Web. And coming from an era where instead what’s most important is the API as opposed to the protocol, the tools as opposed to how they’re intermixed, it is a different philosophy.

STEVEN LEVY: So how did you go about doing that? You arrive at Microsoft, you have a view, you have this perspective of being in the cloud, you’re arriving at the company which is busy getting some maybe belated products out the door, and how did you operate doing this?

RAY OZZIE: Well, in any large organization, whether it’s in government, or the military, or Wal-Mart, or Microsoft, any large organization change management is a challenge, because it’s – I’ll just jump to the end. You cannot effect change genuinely effect change by mandate. You can’t just simply say, this is the way it’s going to be, and suddenly everyone snaps and is doing it. In its earlier days, when I was a competitor with Microsoft, we often were amazed by how quickly the ship could turn. But, that was smaller era. It was a smaller company, and a different era.

And in essence what I feel is the most effective is to set up the conditions to communicate what you’re trying to achieve, to set up the conditions, to lay out the principles, the strategic objectives, and so on. And then at the end, at the opposite end, really start to establish conditions where people are telling stories, where you bring people in who have lived in the world that you’re trying to create, and you get them to mix it up with the people who don’t understand that. And there are various phases that people go through. In those early days I wrote this memo back in late ’05 that kind of laid out a big, broad picture that shook things up a little bit, and it was speaking the language that people in Microsoft understood. Bill would write a big memo, people knew what that meant. It sets the pace for the next era.

STEVEN LEVY: Your Internet disruption – 

RAY OZZIE: Yes, exactly, services disruption. And then I created a conference called Soft Serve, which is software people and services people brought together and told war stories, and so on.

STEVEN LEVY: And you actually even changed in your group just completely the way people work, just the lay out?

RAY OZZIE: Yes, to the extent that I also do some incubation, some things that are a little bit riskier than the product groups might take on, I’ve brought some of my beliefs, in terms of collaboration style, and how I believe startups work, into the physical work environment that we did, and so on. And that work style, those teams ended up producing the DNA of what ultimately became Live Mesh, what ultimately became Windows Azure.

STEVEN LEVY: OK. We’ll talk about that in a second. Was that easy to do, did you get any resistance, when you went in and like knocked down some walls, made open spaces and things?

RAY OZZIE: Well, that’s one of the things that the privileges of being – once I was able to find a location, buildings were tough to find at the time, if I wanted to establish some kind of an effect related to the team, because that’s the way Microsoft works, you do have authority and accountability to be able to do that, so long as you are doing what you are committed to do.

STEVEN LEVY: So now let’s talk about some of the things that you did, because I know after a long incubation period you’ve been out for a while to talk about this. So, Windows Azure, what did you originally call that?

RAY OZZIE: It was called Red Dog.

STEVEN LEVY: Red Dog, now I liked Red Dog. I thought that was a great nature. Why Azure? Why change it?

RAY OZZIE: The history of that code name, which I would rather not go into, was something that Dave Cutler came up with in a way that’s uniquely Dave.

STEVEN LEVY: They’ll have to read my article for that?

RAY OZZIE: Yes, they’ll have to read your article for that. But, it was a good code name. We were very fortunate that in the early stages of that project what we did was we essentially said, we examined the nature of how our best consumer properties were building systems that were running, I don’t know, at the time probably 200-300,  200-400 million concurrent active users, Hotmail and Messenger, and so on, probably half-a-billion at this point. And we took – we essentially took that services expertise and took some of our best operating systems programmers and said, if you were designing a system now for the next 30 years what would it look like, what would client operating systems look like, what would server operating systems look like moving forward? They did a tremendous amount of work, and then ultimately brought it back together on the service side with the service group. So there’s coherence between the service strategy and management strategy, and cloud strategy, and the same on the client.

STEVEN LEVY: And Azure is going to let people build their own Hotmail, is that right?

RAY OZZIE: That’s right. In essence, the nature of Windows Azure, at one extreme not talking ahead of what’s shipping today, but the nature of what we’re building will enable people to wrap existing workloads, existing Windows Server workloads in a way that with as little change as possible they can move those workloads up into a cloud environment, and that could be a private cloud or a public cloud environment. And even they need some work, because configuration-wise, in terms of networking, there’s different latencies between operating things in the cloud and on premises, but with as minimal change as possible to bring existing workloads up.

But, more ideally laying out programming design patterns, and building an infrastructure so that you could say, this is what an ideal cloud program looks like, this is the way you factor the roles, the front-end roles, the mid-tier roles, this is what database looks like in the cloud. This is how you build a program with no single point of failure from day one, with elastic ability to scale form day one, and so on. And so that’s, in essence, what Azure is.

STEVEN LEVY: And Microsoft, of course, is going to eventually have all its own cloud applications on –

RAY OZZIE: That’s right. When I got there, if you look at Hotmail, Messenger, and so on, each one grew up, whether because of acquisition, or because of the state of the art at the time, grew up as a stovepipe. Each one had its own management systems, each one had its own storage, cheap storage system. Each had its own ops group. In some cases they had their own data centers. And we, in essence, said, what is the right way of re-conceptualizing this so that they could all go into a common infrastructure.

Another thing is that, and we could probably talk for a long, long time, because we started basically at a much lower level. We started with the notion of, what would a data center look like from a physical perspective moving forward. And at that point we were – we had just transitioned from what we refer to as a generation one to a generation two data center, where generation one is essentially you have screw drivers and people who install OSes, and buy PCs, at the tens of PCs. Gen two is you’re buying more standardized racks, but it’s still fairly manual.

Gen three is essentially what we’re in deployment on right now, with containerized data centers, where you build the – you build the building, you spent $300, $400, $500 million building a building, and power and cooling, and big stalls. And then as you need capacity, semi trailers roll in with thousands of PCs at a time, and they kind of plug the – but even in that environment you still have to pre-invest in the land, you have to pre-invest in the power and cooling infrastructure to build the shell for that thing.

The next generation that we’re in testing now in a few places in the world are, in essence, free-standing, completely modular data centers, where every component of the data center from the UPSs, the cooling power, whatever conditioning you need, are all free-standing with no roots. So, all we do is prepare the property, build a security wall around it, bring the networking and power in, and negotiate the contracts for that. And then, truly, we don’t have to deal with a lot of pre-investment in inventory.

STEVEN LEVY: All right. I know at least one company making data centers, maybe related to that, has become very interested just in the overall subsection of power conservation, a green technology. Is Microsoft going that way?

RAY OZZIE: Yes. We’re right up there. We’re all kind of at the same place. There’s a lot of – you can do it, we all do it, different people can talk about it in different forms. You know that when, with all due respect to all of these competitors, when we’re all operating at that scale, there are environmental issues, significant environmental issues. It’s not just economics. And, yes, we’re pushing the state-of-the-art.

And one of the things that we’ve found that we didn’t expect is that as we – we can only deploy Azure as fast as we can do it in the geographic territories that we’re comfortable deploying them in, but ultimately because of patent issues, because of government regulatory issues, there will be data centers in every country on the earth, there will be.

And we need to partner with different companies, telcos, and so on worldwide to federate their cloud infrastructure with ours, so that they can build things out. And what we found is that, I suppose it shouldn’t have been surprising, the maturity level, even in very large companies, and governments, is not very high. And we can do a service by helping bring these data center technologies that we’re pioneering. But bringing these to partners, because it is going to be an environmental issue.

STEVEN LEVY: You talked about the massive cost there, really. How many companies can really play this game, and build the cloud here? You mentioned Google, there’s Amazon, you know, maybe Yahoo.

RAY OZZIE: I don’t know about Amazon. Amazon is – they are the leader. They’ve done amazing work. But the level to which you need to build this out to serve at least enterprise deployments is substantial. It’s very substantial. And you have to first understand, Microsoft is building infrastructure for ourselves because we have so many companies out there using Exchange, using SharePoint, there are a lot of companies who want to get into this infrastructure because we can save them a lot of money doing it. So, we’re doing it for ourselves. We’re doing it for Hotmail, doing it for MSN, and doing it for their apps. So, it’s a big investment.

STEVEN LEVY: So, it’s sort of being limited to the companies that as a consequence of their core business need to build these big data centers?

RAY OZZIE: I think that will be the primary driver. It’s a very odd situation in that in many ways, the consumer technologies are what drive you to have the self-motivation to do it. And it’s actually what drove our competencies in-house. Had we not had MSN alive, and have invested, and invested in MSN since ’95, since the AOL, Yahoo era, we wouldn’t have had these competencies in-house, and those competencies wouldn’t be there to benefit the enterprise. Now, all of that is benefiting them.

STEVEN LEVY: It’s sort of like a self-ROI in terms of Microsoft?

RAY OZZIE: Exactly.

STEVEN LEVY: So, if I’m a customer looking to build one of these big cloud services, and looking at the timing to roll it out, what is your competitive edge? What will you tell me to build it up on our platform as opposed to one of the others?

RAY OZZIE: I think we really have, I guess I would say we have five primary advantages that give me confidence. No. 1, I started to talk about it in the data center stuff, but it’s technology. Our ability to invest helps, but we also have MSR. MSR has been doing a lot of the technology – Microsoft Research has been doing a lot of the technology in terms of understanding different cooling mechanisms, just simply driving from the atom on up. That investment has been tremendous, but the opposite effect that we have operating system technology and skills has helped a lot in terms of getting the stuff up and running.

The fact that we had storage investments because of search, in essence, we had a big storage. Search requires a tremendous amount of storage across tens of thousands of computers, and that is the storage system upon which Windows Azure’s storage system was built. So, by the time it was born, it had already been well proven out over many generations of that product. So, from the technology, I’m pretty confident.

We have a developer edge, I believe. Now, time will tell whether this is true or not, but between 5 and 7 million developers are actively working to the Microsoft stack in one way, shape, or form. And today those are the people who are more or less the market opportunity for us. The reality is that the market opportunity is broader than that because just like Windows, where on Windows there are a lot of great open source projects and packages, and people use it as the basis, not just Linux-based, they deploy this on Windows Servers. We actually have the opportunity here to have a business opportunity relative to those developers, because if we can prove to them that we have a great infrastructure for their software, they’re pretty pragmatic, and they’ll deploy it. So, that’s developers.

On the partner side, you know, 90 to 95 percent of Microsoft’s revenue comes through partners. That’s just across all of Microsoft. We understand partnering. We understand the notion that we are part of a value chain. And that it’s really good if you set up an ecosystem such that lots of different parties make money. And, as I said, I think, in the cloud realm there are going to be lots and lots of opportunities for partners to make money, even hardware partners, where on the container side we’re essentially raising the state-of-the-art so that you’ll be able to go and buy a container from various vendors because we’re trying to drive them into standardization at that level.

And the final two are really enterprise advantage and consumers. The consumer advantage I talked about, the fact that we, unlike some of our enterprise competitors had that consumer asset. So, if an enterprise competitor wanted to get into the cloud game, and they don’t have that competency, it’s going to be hard for them to get there.

And, on the enterprise side, we have – I’m not sure how many people realize this, but 86 percent of all enterprise – all companies of any size, meaning mid-size and up, run Active Directory. And that means their users are provisioned in Active Directory. It’s one button press to federate those to the cloud. And so, suddenly you’ve got a world of addressable markets for developers to have one button provisioning of those users within enterprises. So, I think that, plus the fact that Exchange and SharePoint are good applications, most enterprises won’t jump right in at the EC2 Azure level. They’re going to come in saying, how can save money? And today the best way to save money for an enterprise, especially in this economy, is to take things like document management, and e-mail infrastructure, or communications, and to have somebody else manage that for them.

STEVEN LEVY: How big a revenue opportunity is this for Microsoft? Is it going to deliver profits the same way that the previous successful businesses have?

RAY OZZIE: It’s a huge revenue opportunity. The margins on services are not what the margins on software are, so it will increase our profit. It will increase our revenue. But, you won’t have that margin on the rest of it. I should say, though, from a taxonomy perspective, even though the margins at the low level, at the Azure EC2 level, are going to be lower, at the top level, where you’re delivering a solution, or something like Exchange, you are pricing those based on business value more than COGS, and so the margins are still very, very good.

STEVEN LEVY: Your competitor is very driven to charge low prices, maybe in some cases free?

RAY OZZIE: Yes, they are. And we have – that competition is great. We look forward to competition in that realm. We’re not competing with them in that area right now. We’re competing more with Amazon, VMware, you know.

STEVEN LEVY: So, this part of it, when is all this becoming available?

RAY OZZIE: The low levels are available in a kind of a beta, we call it community technology preview form. We’ll be going commercial soon on those low levels. But what excites me, frankly, more, is what’s happening at the high level in these services that we call online. They have changed services. The whole investment is tremendous. A year, year-and-a-half ago when we started talking about this with customers, with integrators and partners, you know, they didn’t really understand why they wanted to do this. Now, these integrators are building practices around this. You have partners out there, and our own sales force knocking on people’s doors saying, how can we save you money? Here is a way that we can save you money. We’ll make money. You’ll – it will cost you less, and it’s good all around.

STEVEN LEVY: So, the second part of this, the mesh.


STEVEN LEVY: Explain that.

RAY OZZIE: Well, everything that we’ve been talking about really is more or less the back end side of what cloud computing looks like. What really turns me on, just because I’m kind of genetically, even though I did systems early in my career, I really want to ship a mass-market app, it’s kind of addictive. We like it. And, the thing that excites me is the transformation that’s happening at the user experience level, and how we consume devices. I mean, we’re moving to a world where we have so many different types of devices, and number of devices in our lives.

And stated kind of abstractly, but probably most coherently, if you were designing an OS today for the experiences that need to be delivered today, you would design it differently. You would have the cloud at the center, everything is all connecting to the cloud, and you would use those devices in rich ways that were appropriate to that device, but leverage that connection to the cloud and to the other devices. So, it would be very easy to buy a six-pack of netbooks at the holiday season, and give them to the kids, because you know that all you have to do is drop them on your desk, or on the table, login, and all the apps are cached, all the data is synced. There’s no reason why code that comes to the client, any kind of code, whether it’s assembly language wrapped up in an exe, or whatever, it should all be cached like JavaScript is cached. It should all be sandboxed like the browser sandbox. Data should be synchronized. The Web isn’t there yet on that, but in terms of technology to enhance the Web and bring us in that direction, you know, that’s the opportunity.

I think the world –

STEVEN LEVY: Isn’t that platform, what you described, it sounds almost like a nightmare for Microsoft, because, you know, it goes against –

RAY OZZIE: As long as I’ve been in this industry, everybody who has got some new technology touts it as being the thing that’s going to kill the previous thing. And when it all settles out, it’s the previous thing and the new thing. And we do more. And the reality is, we’ll always need an OS. Every device needs an OS. And the OS that abstracts the hardware and has the best experience is going to be a great OS for that device. But the programming model on top of that OS is what’s changing. And the experiences on top of that OS is what’s changing.

All we have to do is make sure that the way that we approach the programming model is contemporary and relevant for developers at that moment in time, and the apps have to factor in the Web at the center, because that’s what people do. And there are ways to connect the Web for Web developers with what’s going on in the OS in ways that haven’t even been tried today. There’s no reason that experiences in the browser have to be locked into the browser. If you look at things like the iPhone, or many other things, yes, it can talk to the accelerometer, it can talk to this device, that device.

I think the real questions that people should ask are, well, if you’re a Web developer what might you want to do with Office? What might you want to do with the shell? There are ways of lighting up the browser experience, and the OS experience holistically that are much more compelling than what we have today, and I’m not concerned about there being a future.

STEVEN LEVY: A lot of people probably feel that Microsoft has not embraced this netbook solution, but you seem excited about it?

RAY OZZIE: I don’t know what the stat is, 86 percent, or something like that, of netbooks. I don’t know how much we have to do to embrace it. Yes, we have to probably write an OS other than XP that runs on it, and we’ve done that with Windows 7. But, the reality is, I don’t know what a netbook is. A netbook is evolving into an inexpensive appliance-like PC. It’s a laptop. And is there going to be a netbook that does something that has a value proposition that’s different than a laptop? I don’t know. It might actually happen. I don’t know. But, the way it is right now people are buying them as inexpensive laptops, and the more laptops there are out there the happier I am. There’s more opportunity for us to sell applications, there’s more opportunity for us to sell operating systems. Yes, there are pricing differences, and things like that. But, just think of the large addressable market worldwide.

We’ve – in our industry we’ve tended to kind of, with all due respect to a really big place, we think of India differently. We all talk about India is not really a PC market. India is a phone market. They’ll never be a PC market. Well, once you reach certain price points India becomes a PC market. And there is suddenly a lot more opportunity. So I’m very optimistic.

STEVEN LEVY: One more question, to what degree do the productivity apps become totally cloud-based?

RAY OZZIE: I don’t think it’s – they’ll be totally cloud-based in the realm, in the – let me back up. There’s kind of a – in order to get things going across the company you need meetings, you need to say things, say them again, and say them again. So we say three screens and a cloud, three screens and a cloud, three screens and a cloud, throughout the company. And what that means is everything we deliver, from a user experience perspective, will be – will have some aspect of its value delivered across the PC class of device, the phone class of device, and the TV class of device. Every one of them will have something, and all will be connected to the cloud. That will bring them all together.

The Office experience, it’s not software for a PC. It’s productivity. People are paying for productivity. So every person when they buy Office will be doing editing, and looking at big stuff, and big desktop screens, because that’s what it’s good for. The PC, nothing will ever be the PC in terms of hitting page down, or the down arrow on a big spreadsheet, and scrolling around. It’s just so compelling. And so that’s how it should be delivered.

Yes, it has to be cached, it should be delivered from the cloud, but its native code for the PC is great. But, people, most of the world’s people don’t come together on the PC, they come together on the Web. And we do a lot of sharing. You don’t create documents for yourself very often. You create them as part of larger things. So the sharing scenarios, the collaboration scenarios are homed, rooted on the Web. And then there are phone scenarios. Everything that we do, you go to meetings, and productivity, you don’t always take your laptop, but you probably always carry your phone with you. The phone has your location. The phone has a recorder. The phone has a little thing that you can take a picture of what’s on the whiteboard. The phone is an amazing companion to the Office scenario. So 100 percent of Office will be cloud, 100 percent will be mobile, 100 percent will be PC.

STEVEN LEVY: OK, a couple of other non-cloud questions. It’s been about a year since your predecessor as Chief Software Architect – 

RAY OZZIE: That’s right.

STEVEN LEVY: Left the company on a full-time basis. What’s been the effect of that, for you, for the company?

RAY OZZIE: Reviews are conducted a little differently.

STEVEN LEVY: That’s the stupidest thing I ever heard.

RAY OZZIE: Yes, true. Bill, even though he is physically not there, he’s there. Bill will always be there. Bill – 

STEVEN LEVY: Does he write? Does he call?

RAY OZZIE: He writes, he calls, but infrequently, and on things that he’s interested in. He’s engaged in the things he wants to – everyone knows he’s just a BillG@ away. He’s there. But, he knows that he’s not accountable for our success anymore, therefore he knows not to give orders, or directives, or influence things that would blow the accountability of being able to deliver something and the accountability you have to have in a big company.

In essence, every company is shaped by its founders. Bill brought to the company a sense of – a culture of crisis that any day now two guys in a garage are going to be there to take down the company. (Laughter.) That is culture is very – 

STEVEN LEVY: Was he right?

RAY OZZIE: Well, yes, he’s right. No, he’s right. And so that culture and the fact that the company has been through competitive battle after competitive battle, after competitive battle, Bill’s a very – it’s a very healthy culture in that realm. And that’s why when new things come along people don’t get freaked out by it. They think about it.

STEVEN LEVY: Have you figured out what to do about Think Week? Do you have your own Think Week?

RAY OZZIE: Yes, Think Week was – it was conceptually a very Bill unique thing, in that Bill – for those of you who don’t know, Bill would kind of open it up to the company to give – to write papers and give him input. Bill has an amazing ability to consume, consume very, very quickly. So thousands, a-thousand-something papers would come in for each Think Week, which was held I think three times a year. He would go off and sequester himself, and his staff helped prioritize, and he would probably read a couple hundred of them and give feedback in some way, shape, or form. And it was just like a wake up, go to sleep, wake up. And people loved it. People loved getting Bill’s input, and Bill would hand it out and give a bunch of people – send it to a bunch of people who also commented on it.

I don’t think that’s something that we want to reproduce. It was a great thing. But, it really gave a lot back to the culture, and people like feedback, senior feedback, senior technical feedback and visibility. So it’s reshaped, but it’s essentially the same thing, but it’s a broader senior set of technical individuals who are giving feedback in a more – in a slightly different way. So we’ll see how it pans out.

STEVEN LEVY: And finally, before questions, you want to mention it’s in your bio here, and I wrote about it, and you mentioned it once, PLATO. One thing that I found fascinating when I wrote about you was how influenced you were by this system that you came across when you were an undergraduate in college. Can you tell us just a little bit about what that was, what it meant for you, and the effect it still might have on your thinking?

RAY OZZIE: Yeah, I will. How many of you – I’m just really curious – how many of you know — would say you know much about PLATO from that era? Yeah.

It’s kind of a tragedy that the Bay Area culture, given that this is the center of the universe – (laughter) – doesn’t have as much an appreciation for the contributions that that system had at its time. I hope at some point we can do something at the computer museum or something to kind of showcase. It’s live, it’s on the Net today at cyber1.org. You can log into it.

STEVEN LEVY: But what is it?

RAY OZZIE: So, this was the late ’60s when it began. I was exposed to it in the mid ’70s. I was a systems programmer on it. A gentleman by the name of Don Bitzer, I’ll say a creative eccentric, really bright guy, but he’s a solution-focused guy, and he believed, he deeply believed that education was in trouble in this country, and that there was potential in utilizing technology in teaching, in new ways of teaching, new ways of education.

So, he created this thing, PLATO, Program Logic for Automated Teaching Operations, and just said we’re going to build everything that we need to make the most compelling thing at the time.

And the way I was exposed to it was I was at the University of Illinois. I was literally punching cards and submitting them to the 360. And directly across the street there was this building that had this orange glow coming out of the windows. And what they were, were Don Bitzer and his team invented the plasma panel. Do you remember what the plasma panel is? It’s on your TVs, a lot of TVs. But he invented it because he wanted graphics terminals, and graphics terminals need memory to display what’s on the screen. And there were storage tubes at the time, but plasma panels had memory per dot. So, they invented plasma panels so they could have a terminal that had a persistent image on it.

That terminal had a microfiche projector, displayed multimedia on the back of it. It had an audio/video device with headphones. And the users, you know, it was a centralized mainframe, thousand terminals, 500 at U of I, 500 in pockets worldwide, big network, big systems thinking, and students would use it in educational settings.

But from 10:00 p.m. till 6:00 a.m. was non-primetime, and the authors of lessons and we as systems programmers got a chance to play with it. And what people built were instant messaging, character at a time instant messaging, group chat, online gaming. Some of you might remember – well, all of the air simulations and things like that, flight sims, came from that heritage. There were online discussions called Group Notes. There was e-mail; it was called Personal Notes, Lotus Notes, UNIX Notes. All of these things were derived from that.

And it was online community. It was we who were fortunate enough to be exposed to it at the time understood what it was like to have people who had relationships with people across the world or across the country that they never saw. You’ve got to understand what flamers were like, this, that and the other. (Laughter.) And this is where I began to gain an appreciation for collaboration, because I worked on projects with people that I didn’t see face-to-face, programming projects.

STEVEN LEVY: And I think that shaped your career then, and now you’re building that at Microsoft.

RAY OZZIE: Absolutely.

STEVEN LEVY: Thanks, Ray. (Applause.)