Steve Ballmer: Cloud Computing

Cloud Computing
Remarks by Microsoft CEO Steve Ballmer
Paul G. Allen Center for Computer Science & Engineering, University of Washington
Seattle, Wash.
March 4, 2010

ANNOUNCER: Please welcome University of Washington President Mark Emmert.

MARK EMMERT: Good morning.

Well, first of all, let me welcome everyone to this wonderful space here in Computer Science and Engineering. We expected a standing room only crowd. We wound up with a hanging room only crowd. This is great. It’s good to have everybody here. As everyone in this room knows, our relationship with Microsoft is a very, very long-standing one. Ever since Paul and Bill first snuck onto campus and stole computer hours, up until this very current time, the relationship between the UW and Microsoft has been arguably the most important relationship the university has ever had and continues to have, and we’re very proud of that. We send somewhere around a 100 or more graduates a year to go to work for Microsoft. We have internships. We swap faculty relationships in research and outreach and education that are very important to us. And so, for us at the UW to have Steve here to talk is a marvelous opportunity for our students and faculty to hear what’s on his mind.

Steve doesn’t need much of an introduction, obviously. He’s someone who is widely known. He has a resume that is nearly impeccable except for the one singular flaw that he didn’t attend the University of Washington. He has managed to rise above his Harvard degree, though, and become successful as a great business leader, someone who was here, if not at the conception of Microsoft I guess you were kind of a midwife to it, employee 24, or something like that, Steve?

So, all of the great things that have occurred at Microsoft, Steve has been involved in, and been in leadership in that whole business enterprise from the very beginning to create this organization that’s had a transformational impact on the planet, on the way we work, and the way we live.

So, it’s with great pleasure that I introduce a wonderful friend of the University of Washington, a terrific business leader, a wonderful citizen of the Northwest, Steve Ballmer. (Applause.)

STEVE BALLMER: Thanks, Mark. I have to admit, I’ve never exactly had a hanging room only crowd. So, I’ll try to project up. Hello up there. As well as out.

It is a great privilege for me to have a chance to be here, and to speak. There’s some weird irony to me, when I travel around the world, our people always say, well, we’d like you to talk at this university, or that university in Oslo, in Budapest. They threw eggs at me, actually, at the university in Budapest. Some irony, and I’ve never spoken here at the UW. And so the chance to come and be here today, and have a chance to share some thoughts with you, and take some questions is, for me, really quite an honor.

And, as Mark said, there’s no more important research institution, source of great talent, and partner and friend for Microsoft than the University of Washington. And that’s been true for time immemorial. This has always been sort of a place where we could count on finding the best and the brightest, who can help really drive and shape the future of technology, which is where I want to spend my time today.

You know, it’s an amazing thing in a certain sense how the tech industry just keeps pushing forward, pushing forward, pushing forward. I’m no grand history of other industries or other technologies, but they seem to run out of gas at some time. And I feel like kind of the tech industry is the gift that just keeps right on giving. You know, when first computer, 1949, and we’re sitting here 60 years later, and we still have so many things that we all get a chance to drive and go do.

Ed Lazowska from the UW was sort of highlighting earlier that 1969 was really an important year. Woodstock, man on the moon, and those were small, because the Internet was really first used in 1969. The gift that just keeps on giving.

Today, I want to talk about the cloud. The cloud sort of is part of that Internet gift. It’s the next step, it’s the next phase, it’s the next transition, and depending on who you are, and how you think, you could say the cloud started five years ago, ten years ago. You can go back to 1969, if you want, and say that the cloud started 40 years ago, because the microprocessor and the Internet are the gifts that just keep on giving us the chance, and the opportunities to make a difference.

If you just think back over sort of recent memory, the browser was an invention of the early ’90s. Windows 95 was radical because it had a browser built in. I’m not sure I’m going to spend a lot of time on sort of the sub-history about that. You know, when free e-mail, and we first we got involved with Hotmail in the late ’90s. Woo. And then search. Woo. And yet this cloud thing is just another big step that the world is buzzing about, and thinking about, that will create the opportunities for all the folks in this room to do important research, to build important products, to drive forward, and do fantastic things.

Ray Ozzie, who is our chief software architect, wrote a memo almost five years ago now talking about this and its importance, and we were sort of stimulated and driving the company, the cloud, the cloud, the cloud, the cloud. And yet here we are five years later, and there’s still so much unrealized potential. Big problem, what the heck is the cloud?

So, I’m going to give a little bit of perspective on the cloud really from the standpoint of people who get a chance to use it, to drive it, to shape it, to make something of it. And that will be my perspective, but I think you ought to all understand a little bit kind of what the average person on the street, so to speak, here at the University of Washington thinks about the cloud.

So, we did a little bit of video work, and why don’t we roll the video and get some UW perspectives on the cloud.

(Video segment.)

I don’t know if we’re always going to be talking about the cloud, that’s a word that might last three years, five years, ten years, then it’s said the gift will keep on giving. So, the real thing to do today is to capture what are the dimensions of the thing that literally I will tell you we’re betting our company on, and I pretty much everybody in the technology industry is betting their companies on, $3.3 trillion dollars, or whatever it might be, globally industry, all bet on this incredible transformation around the cloud.

In a sense, people say it’s everything. But, thank goodness we’ve got Coach Sarkisian to be quite precise about some of the things that we think are most important about the cloud. So, what I’m going to try to do is share with you five key dimensions in the cloud, five key opportunities, five key things that I think need your best ideas, your best thoughts, your best invention, commercial inventions, academic inventions, product inventions to really drive forward. And I’ll give you a little bit of context about what Microsoft is doing in that regard.

First principle, the cloud creates opportunities and responsibilities. Now, that sounds like some blah, blah, blah, business term, blah, blah, blah. It actually is a whole statement about a range of innovations that I think we are seeing and will continue to see where there are literally new software investments that create new business models, new opportunities to start and form businesses, because of commercial software infrastructure that’s never existed before, but also creates opportunity and responsibility to the user to protect the user, to protect their privacy, protect their confidentiality. And while these are all social issues, they’re also all technical issues and invention issues, and innovation issues.

So, as we think about the cloud as an opportunity, I think a lot about the things that people have done. The ability for literally any small creator to create a piece of content, create a piece of software, to have it available instantaneously anywhere around the globe is fantastic. But, as we start to see more and more advertising infrastructure to support those small creatives, we see the invention of things like the App Store, where Apple has done a very nice job that allows people to monetize and commercialize their intellectual property.

As we start to see more and more work go on in the area of catalogues. For years, for years in the pre-cloud era I spent my time meeting with a brilliant idea and they’ll say we’ve written a great application and we’ve got two customers right here in our own hometown of Dubuque. And now Microsoft, please can you help us find customers in Des Moines and maybe Seattle, and maybe New York. And you think about the commercial infrastructure that now will support a whole new class of creators, creators of a wide variety of intellectual property.

Open Source has been kind of an interesting phenomenon. People said, I’m willing to participate in the creation of software essentially on my own time, and for free. And yet with the advent of these new commercial infrastructures I’m sure some of the inventors who participate are now going to be able to ask, how do I monetize, how can I get economic value from the incredible innovations that I get a chance to create.

The amount of invention that needs to happen is high. The world is still not a perfect place, instead of  in terms of the commercial infrastructure. Yes, you can create a Web page and put on an AdSense ad. There are some things that you can do. But, we certainly haven’t fulfilled the sense, the opportunities to create technology that empowers the creator. That is an aspect of the cloud that I’m excited  a dimension of the cloud that I’m excited about.

Immediately people get nervous, particularly when you talk about advertising. They get nervous, what about my privacy. And that’s why I think we have to talk about the opportunities and the responsibilities. The responsibilities for creators, for business people to respect the consumer, to build technologies that really do allow the user to be in control. We had this huge internal debate inside Microsoft when we came out with Internet Explorer 8 with a feature called in private browsing.

Should you really be able to turn off, and there are a lot of people on the commercial side of our company and in the industry who said, probably that’s not a good idea, because there will be a restraint to commercial activity. And yet I think we have a responsibility, all of us, not just to socially respect the user, but to build the technology that will protect the anonymity, the privacy, the security of what I say, who I say it to, where I go, what’s important to me.

And the level of invention is high. I’ll bet if I did a little poll we’ve got some Facebook users in the world. Facebook has done a remarkable job in many ways. And yet the challenges that we’re all seeing, and that they’re seeing and how you let somebody describe technically user interface-wise, how do you describe who you are, and what you want to protect, and who you want to protect it from. The amount of great IQ that’s got to go into that problem here at the University of Washington, at Microsoft and elsewhere, is really quite high.

At Microsoft we’re trying to take kind of an interesting view on this. We think as a big company we’ve got to lead on privacy. We’ve got to really do the best job we can to make sure that there’s a vibrant and healthy competition amongst all of the companies that provide commercial infrastructure. We kind of understand both sides of that ledger sheet, so to speak, because for the creator to have the opportunity to really build a strong business it will be important that, whether it’s payment infrastructure, or stores, or catalogues, or billing, or advertising, that there be a number of strong and innovative commercial players in the marketplace. So, this is a dimension of the cloud, and it’s a dimension of the cloud that needs all of our best work in my opinion.

Second dimension of the cloud: The cloud learns and helps you learn, decide and take action. If you ask kind of most people what is the Internet, as opposed to the question we chose to ask, what is the cloud, if you ask most people what’s the Internet, they’re going to tell you it’s a place where lots of people and companies are all out there virtually. It has something to do with the creation of a virtual world.

The world is a large, complicated place. So, the first thing that got built to help people navigate were essentially directory services, search services. People built tools to help you navigate and find information, pull it all together, et cetera. And yet, we’ve got to further than that. The cloud, the Internet, the cloud needs to learn about you and it needs to keep learning and figuring things out about the world that has been described virtually.

It’s great to know about 83 million Web sites on the planet, but if you’re actually trying to find something specific, I’ll put my hand up, as part of the U.S. healthcare debate I decided I should actually understand what we spend money on as a society. Try that one out for size. Pick any search engine you like and go give it a whirl. You’ll get a bunch of links, you’ll find a bunch of data, you’ll probably try to cut it, copy it, paste it, but you won’t be able to just sort of describe maybe like a simple, little chart that you would like to see populated. How much money do we spend on healthcare, how much of it gets spent on older people, younger people, poorer people, richer people, people in the last year of their life.

It’s only about eight numbers, there happen to be eight numbers that you can’t learn by following the public debate. But, there were eight numbers that I felt as a citizen I ought to know. But, the ability of the cloud to actually learn form all of the data that’s out there, and the ability of the cloud to learn from me what I’m interested in is not what it will be two, three, four, five years from now.

Machine learning, machine learning is a science, a very big science here at the University of Washington. But, a science that wasn’t of much real practice. I would say in industry if you go back at least 10 years, for sure. And yet today it’s maybe one of the hottest aspects of computer science, because we’re actually trying to take a look at information and glean meaning from it. And we’re trying to take a look at human expression, and understand intent, and that area of endeavor is just going to continue to grow.

And the tools around that that really help somebody get the information they want and make a decision. When I type flowers into a search engine, I’m not really interested in ten blue links about flowers, I really want to buy some, and maybe even at the best price. And, you can walk through the many things that people do, and how do we provide the tools and technologies? And it’s not just a question of the search engine. We’ve got this product that I really love called OneNote, that lets you collect information from a variety of sources as kind of research tool. But how does it fit, and how does it play? Excel, I happen to be a numbers thinking guy, I would create that little healthcare thing as a little spreadsheet. I would want Excel to just go get that stuff from the cloud. And so this notion of learning, learning about me, learning about the world, making conclusions, and then helping me to decide and take action, I think is a very big idea.

It’s not just the province of a few big companies off in their research labs, although companies like ours, and some of our big competitors and partners we have a special role, we have to open this up, though, because there’s tons of people who will have innovative ideas in specific domains.

We were talking, Hank Levy, Ed and I were talking about science. There will be people who know a lot more about the semantics, and will be able to describe, and help comb the Web for scientific data a whole lot better than anybody at Microsoft, or Google, or any of the big search engines. But, how do you let people plug in and take advantage of the cloud to provide new learning, new context, new data. It’s a huge set of opportunities that certainly we’re very excited about.

To give you a little bit of sense about learning, how the cloud learns. The cloud itself needs to learn. It’s got to collect new data. It’s got to sense new data. It’s got to represent the real world, and keep getting smarter and better, so that it can help me learn.

We thought we’d invite up on stage Blaise Aguera y Arcas, who is going to show you a little bit of some of the things that we really mean about the way the cloud can learn and help you learn.


BLAISE AGUERA Y ARCAS: Thanks very much, Steve. (Applause.)

So, what I would like to show you today is the new version of Bing Maps, and it embodies a number of the ideas that Steve was just talking about. This is the Web site, so it’s up. We’re not routing everybody to it by default yet when they just go to Bing Maps, because we are still learning. But, what we’re trying to do in this new version is to really go far beyond just the classic idea of mapping online as it was innovated in the ’90s by the likes of MapQuest and so on, where it’s focused on driving directions, and secondarily on just spatial searching. What we’re really trying to do here instead is to think about mapping as a trellis or a surface on which you bind every conceivable kind of information that relates to space, and to time as well. And that surface area is one that should accrete information from the cloud, from users, and from developers, bring all of that stuff together, and integrate it in a way that you can explore, and decide.

So, this is the site. It’s powered by Silverlight. So, one of the first things that you’ll notice about it is that it has this fluidity that’s beyond what one can do with AJAX. But, that’s not the reason that we did the Silverlight work. The reason really is that the visual richness, as well as the composability of different elements is much greater when you use an environment like this one.

So, I would like you to imagine that  I need to tell some sort of story. So, suppose I’m a prospective student, and I’m trying to figure out if the University of Washington is the place for me. As I zoom in close, the roads resolve into this seamless aerial image. This is a synthetic aerial perspective on the university, and on anywhere that you might want to zoom in, that’s looking obliquely at 45 degrees, which really gives you a much better sense of the environment from the air than the view straight down from the air, would.

And, as I move closer in, notice that there was an interesting transition there, and we’re now looking at a single aerial image. We really want to show you the original data bound to space, not just that composed and synthesized version of the data. And, as I move around, I’m going to transition from one aerial image to the next very much in the manner of Photosynth, which of course was a wonderful joint development project that we did with the University of Washington. Its lineage comes from “Photo Tourism,” which was written by Noah Snavely right here in this department.

Now, as we go even closer in, we get beyond the level of what we can represent. Let me get my bearings again, the fountain. We go beyond the level of resolution that we can represent from the air, and we come down to the ground. Now, this imagery from the ground, of course, you can’t collect just using aerial cameras. These things are collected using a device that looks like this. It rides on top of a car. We think that this is actually rather clumsy, and has problems in scaling up to the kind of depth and richness, and ubiquity that we’d like. We’re experimenting with techniques more like this one, and this one, and this one, in order to continue to collect and bring it indoors, and really scale this out to everywhere.

Now, this is, of course, what that particular camera collected at one moment in time. But if I launch now this map application called Streetside Photos, which we just released a couple of weeks ago, then what I’m seeing over here are Flickr images. They’re sort of creative commons Flickr images that start off geo-tagged. Now, when you geo-tag a Flickr image, of course, it only positions it very approximately in space. But what we do is, we use more of that sort of Photosynth type magic to bind those Flickr images directly to our own imagery. So, these are images of the former home of computer science at the UW. I think everybody is quite happy that you’ve moved.

So, that registration of that Flickr image with the imagery that’s collected by us really shows you how the head content that we collect, and the tail content that users collect bind together in space. This is a simple example, just a single photo. But, as we extend this to entire synths, other forms of collection, video, semantic collection, this starts to get very interesting.

Now, this is not just imagery, which should be apparent as I move through space. There’s an entire three dimensional model that’s been inferred from this photography. And so our information about the semantics of space begins with pictures, and then it extends to models, and geometry. And then, of course, it extends beyond that to all of the different sorts of abstract information that can be connected to place.

If I’ve navigated correctly, I think this is where we are right now. Let’s fly back up. OK. So, now I’m going to switch gears a little bit, going back to my student role play, and let’s take off photos and pull up, instead, local lens. What local lens is doing is binding another sort of information to space, to the map. These are blogs, these are hyper-local blogs. They’re being crawled in real-time. So, this is a very risky demo to give in public, because I never know what kind of thing is going to come up. This is from 10-07. But all of these little pins that have appeared on the map represent the location that was mined, that was referenced in one of those blogs, and it’s mined in real-time. So, this happened at 7:22 this morning, the taco fire truck on the corner of  right. So, the taco truck will soon be repaired. This might well influence my decision about whether or not to come to the University of Washington. (Laughter.)

So, binding all of that information to space lets me explore these things through space, through time, and through content. So, it provides a much more multi-faceted and multi-dimensional way of exploring and experiencing that information.

Now, what’s nearby is another example of a map app. This one has a lot of  this one is useful really often, because what it’s really about is exploring a region in space, and telling me about all of the businesses nearby. So, in this case, if I’m a film buff, I might be interested in what sorts of movie theaters are near the UW, and this is good news, because Landmark is a great independent theater.

So, I’ve got all sorts of structured information that is flowed in from all of these entities. And so, I can see, of course, the ratings, and the data, and so on, and the Web site. And, I can dive in and have a look through that same first-person imagery. So, I’ve bound together the semantic information, and the ratings, and the blog information, and the visual information that we captured ourselves, at this particular moment it looks like it was playing The Maid.

So, this really shows you all sorts of information binding together. It shows you the cloud, the spatial incantation, if you like, of the cloud learning in a variety of ways, from content, from users. The ability for developers to let this environment become richer over time I think is also very key. So, all of these map applications that I’ve been showing you are written by small teams, in some cases only one person in Microsoft research. I’ll close with one that got a good reaction when we previewed it for the first time a couple of weeks ago.

This is WorldWide Telescope. It was originally a Microsoft Research project. So, the integration of Worldwide Telescope is really done by one or two developers into Bing Maps, lets me look up into the sky and as I look up it’s showing me the star field. So, of course, the star field doesn’t stay fixed. So, this is right now at 10:34 a.m., what it would look like if you could turn off the sun. And you can look at different times, different dates, when does Venus rise above the horizon and all that kind of stuff.

And this is, of course, not just a picture. We have here a very, very rich source of astronomical information, thanks to the engineers and the scientists who have worked on worldwide telescope, and all of that is bound to space and time. So, this is a nice high-resolution image of the moon that I think has just been added. And, of course, eventually you should be able to fly up into the universe and explore well beyond the earth, not that I think that really has very much to say about whether or not you’ll go to the University of Washington. OK. Steve, I’ll give it back to you at this point. Thank you so much.

STEVE BALLMER: I hope the demonstration does a couple of things. Number one, I hope it kind of wets your whistle for some of the kinds of things that can be done. And number two, it really helps bring together this notion about learning about the world, how do we learn from others, how do we pool the data that’s available on the Web to learn about the world, and then map it U.S. and make it of interest to somebody in real life.

Second, or third, rather, dimension. The cloud enhances your social and professional interactions. This is one I think if you stop and think about it you say, of course, that’s e-mail, that’s social networking, that’s this, that’s that, that’s the other thing. And the truth is, that’s all correct. And yet the degree of innovation that will continue to go on in the area of people helping them connect with one another and connect in various personas.

There’s what I want to do socially, and what I want to do professionally. I don’t even think you could say is, if you will, and personal, because the truth is we all are people we go to school, we go to work, we have personal lives. We socialize when maybe people don’t anticipate that we will, like when we’re at work, and we also do things professionally that are surprising.

People always like to say no, no, professional means at work. And I say, if you watch my wife who works in the home and is involved with a couple of things to do with foster care, actually, run here out of the University of Washington, you’ll find that she may work out of the home. But, the way she manages her calendar, and her work, and her collaboration with others, these things all blend. And the cloud winds up being the place of innovation that lets us pull our lives together the way we want to, to touch people the way we want to. To find people when and how, and the way we want to.

We want to be able to put these things together, and again, build on the notions of learning. We want to be able to do that directly. We may want to be able to do it in context. I’m in the context of a project, or I’m in the context of a game, or context of watching football and I want to engage with somebody, or I may just want to go out, find friends and interact. The amount of innovation that’s still to come is amazing. We talked today about the social graph, thanks to a lot of the good work that’s going on in the social networking sites. And yet the ability to really connect people, and to help people connect is just beginning really to be tapped. Whether that’s on Facebook, whether it’s on Xbox Live, on Twitter, there’s so much more to come.

The way we connect with each other. I felt like I should come to the University of Washington to give this speech. It’s ironic. I’m here talking about virtual interaction, socialization, and yet I wouldn’t think it was right to not be here in the Allen Center itself to talk to you. And the truth of the matter is, I’ll know we’ve succeeded with this aspect of the cloud the day we all agree that the virtual interaction through the cloud is as good as being here. And that’s a combination, again, of a lot of innovations, hardware innovation, software innovation, some people think it’s all about getting enough bandwidth. We can kind of prove to you there will never be enough bandwidth.

So, there better be enough software to figure out to really create a real-time virtual image of this room, projected at Microsoft and projected  it’s probably easier to project back. This would be a particularly tortuous test, I think, of doing a good virtual meeting. But, it’s part of the way that you really start getting down the path of thinking about social interaction and professional interaction.

The way businesses collaborate has changed and will continue to change. One of our very good cloud customers is a company called Aviva. They’re about the fifth largest insurance company in the world. The name Aviva is a new name. There was one 24-hour period when they switched it on, because this was a merger, a blah, blah, blah, this company, that company. And on the first day they decided we really better get out and tell our employees that we’ve got a new name. It’s going to be a tough day if somebody goes home at the end of the day and their spouse, or significant other has no clue where they work.

So, really, having the CEO engage in virtual collaboration, communication, with that audience through the cloud, because not everybody is at work. You can’t count on everybody and being in the four walls, people with customers, their various places, how do you do that, how do you pull that together. It’s the kind of things where more innovation is necessary. The kind of collaboration that we can do inside the university. If you really try to some of the same things on an ad hoc basis, and still with a level of security and privacy is really quite challenging.

This is probably, if you will, the most mainstream thing for Microsoft. If you think about where we grew up, other than Windows, we grew up with this product called Microsoft Office. And it’s all about expressing yourself. It’s e-mail, it’s Word, it’s PowerPoint. It’s expression, and interaction, and collaboration. And so really taking Microsoft Office to the cloud, letting it run in the cloud, letting it run from the cloud, helping it let people connect and communicate, and express themselves. That’s one of the core kind of technical ambitions behind the next release of our Office product, which you’ll see coming to market this June.

We’re having some success. For the parts of our Office product that are already in the cloud about 90 percent of the customers, at least institutions that we work. with choose us. A lot of good work is still going on for the consumer. We’re pushing forward with Hotmail, and Windows Live. We’ve got other guys doing fantastic jobs on helping consumers in their social persona come together. This is really an area, I think, of important work.

I want to show you kind of a little demonstration of something that you can actually do, anybody can do if they live in England. But, we’re going to show it to you here today. It’s based around Xbox Live. Xbox Live is part of the technology set that enhances social and professional interaction. It’s a cloud-based way of doing that, and so I’m going to invite, if you will, Simon Atwell from our team, who is going to show you a little bit of what Xbox Live might look like if you were trying to watch TV with friends living in the UK.


SIMON ATWELL: Thanks, Steve. (Applause.) Thanks, Steve.

And good morning everyone. As Steve said, I work at the Xbox Division here at Microsoft, and I’m super excited to have the opportunity to share with you the Sky Player experience that, as Steve said, that we built with Sky in the UK.

As you can tell from my accent, I am originally from the UK, and before I moved out here I was a Sky TV subscriber, and I loved the content that they offered. But, I’m a gamer. I love playing games on my Xbox console, but games are great, but I want movies, I want TV, I want it on my console, and I want it now. And so, to me, it just seemed a natural way to converge that Sky content and bring it to the Xbox platform, and deliver it through the cloud.

For those people who aren’t familiar with Sky, Sky is an integrated part of lifestyles within nearly 10 million households in the UK and Ireland delivering comprehensive, high-quality television. So, now I have this ability from the cloud to have an experience on my console having access to this great content. And so, I’m going to walk you through what that would look like. I hope for everyone way out there that that’s working out for you.

So, here I am. I’ve landed at the Sky home page, and had this very smooth and simple navigation around collections of content that helps me discover what it is I want to watch now. I’ve got movies, and sports, entertainment, and other great categories, and that enables me to discover the content. So, let’s say, for example, I’m going to go into Sky Movies. So, again, delivered from the cloud, recommendations around latest content. So, if there’s information about a movie that I want to find more about, I can drill in and get that. I can find related movie items to the item that I’m actually looking at. This is a great way for me to discover content that I want to watch now.

I also have the ability to look by ‘most popular.’ I could even search by title, trying to find content that it is I want. And, because this is delivered from the cloud, it gives the ability for Sky to deliver timely recommendations. As you can see here in the box set collection, where the Oscars are front and center. So, I have people over to the house, we want to figure out what it is. I can view by category, find content that we can all consume, and we consume it right there and then.

So, to complement the movie service, Sky has a great on-demand TV service as well for catch-up TV. And I kind of think you’re looking at that and going, it’s kind of cool. It’s on a console. Cute-looking Xbox, but I’ve seen it before. But, for me, the one thing about Sky that differentiates is the ability to watch live TV on my Xbox console and share that experience with my Xbox Live friends, the same friends that I play online with.

So, here I’m going to go into Live TV, and as you see I have a list of channels that have content available right now. And it is a UK service, so as you’ll see it’s time zone based on the UK. They’re about eight hours ahead. And the way I interact with Live TV, and it’s just playing out for me there, because it’s looking at my last-viewed channels.

I can go into the TV guide, and I’ve got this simple, easy way to navigate around content. I’ve got channels, times, and as I navigate around, I get more information about the content that I want to watch.

So, let me go to something that would be fun to watch. MTV, the cool crowd. So, this is actually connected to the cloud, to the Sky service back in the UK, and this is streaming Live TV. It’s streaming Live TV to the console. I think that’s really amazing. I really, really do.

But we didn’t stop there. For me, watching TV is a social experience. I want to get together with my friends, and I want to share my opinions, and they’re going to hear them, whether they want to hear them or not, but they’re going to hear them. And I want to do that with Sky.

So, what I’m going to do here, I’m going to back out, and we’re just looking at the TV list, and I’m wandering down the list here, and, well, a friend of mine online. So, I’m going to just go jump in, and watch Live TV with them. Obviously we’ve got a 4,700-mile delay on this. But we’re feeling confident about it, it’s going to go. Here we go. And, yes, I want to watch this with my friends. And here I am, I’m – this is how I watch TV when I’m at home with my friends. I’m on a virtual couch, and I’m watching apparently the biggest TV in my apartment I can get my hands on. And I think that’s a great experience. As you can see here, I’ve got three friends sitting on the couch. We could be anywhere sharing this experience.

Now, while we’re synching up with the other party friends here. We’ve got that slight little delay. We’re going to push that off to the whole UK, and we’ll be fine with that. But the thing is, is that I get this experience where I’m just doing something that’s social more than just playing games.

Well, as I said, I want to share my emotions. So, I can go ahead and do that while this video continues to load. So, as you can see here, I can go ahead and be super excited about the content that we’re so close to watching, so close to watching. Let’s do it again. Yes, come on. Again.

But the thing is, it’s a social interaction, so maybe I just want to get a better view of what that interaction is. So, for the purposes of this presentation, Steve, I took a little bit of artistic license. And you have Steve (Ballmer) on the couch, with Bill (Gates), and the president of the university Mark (Emmert), all seen here having some fun. (Laughter.) Now, as you’ll see, the TV is still playing in the background, but now I have a way to more socialize with the people I’m watching TV with.

I’m a huge sports fan. I love sports, get totally immersed in it. And so, we’re watching this great game, and it’s just the worst call ever. And, I’m like, you know what, I’m sharing that with my friends. So, I’m just like, no, that’s just terrible, just terrible. But, it engages with my friends. So, Bill thinks that’s funny.

OK, I just have to – yes, we just – yes. You know what, and so – (laughter) – but it’s just a great social – and Mark is just having fun watching the pair of us duel, and all of that. It’s just a great social experience. I’m watching sports, watching TV, and I’m having fun.

Now, you can take this even further. I can get my Xbox wireless headset, clip that on, and talk to my Xbox friends online while watching the TV, so they get to even hear even more of my narrative. (Laughter.)

And with that, Steve, if I was in the UK, that’s how I would use Sky Player. (Applause.) Thank you.

STEVE BALLMER: Latency, keep that in mind for later. There is definite latency, another software problem. And it could be all about the fat pipes that will run across the ocean, or it could be a lot about solving software problems that really facilitate and continue to drive this notion of social and professional interaction.

I think we still have Sky TV on, if somebody could just turn off this speaker, I would sure appreciate that. I think I’m hearing the sports broadcast from the UK, and I don’t know the names of those football teams.

Dimension No. 4, the cloud wants smarter devices. The cloud wants smarter devices. This isn’t to say that we’re not going to continue to do a lot of work on browsers, and standards, and moving forward, that’s all going to happen. But, the truth of the matter is that when it comes the cloud, the devices that you use to access it do matter, whether it’s the PC, the phone, the TV, as we just had a chance to show you, the devices that you use do matter.

The way in which we can learn about you, the sensors, the cameras, the voice, the gestures, today, this year, we’ll get about 10 billion utterances, speech utterances, submitted to us in the cloud through something called our TellMe Service, which handles call centers, and Bing kind of phone voice response searches, and the like. And so, the ability for the device to participate in connecting to the user, providing a richer interface, to get data back from sensors, and use that to improve the cloud experience on behalf of the users is really quite strong.

Later this year, we’ll ship a thing that we call Project Natal. It’s a camera that comes with the Xbox, and it recognizes you, and your voice, and your gestures. The hardware does matter. It absolutely matters. The great smart device hardware is going to bring together the best of what we think of today as rich clients, and the best of browsers, and the best of a next generation of natural user interface, voice, touch, speech, et cetera, all in one unit.

And you see those innovations proceeding today. PCs don’t look like PCs did five years ago, and the cloud has a lot to do with it. Phones don’t look like phones looked at all five years ago, and they’re not going to look the same in five years. But they’re smart. They’re going to get smarter. That doesn’t mean that they can’t be simple. It doesn’t mean that they can’t be cheap. That’s the job of us innovators, how to give people the smarts that they want to take advantage of the cloud, and at the same time the simplicity and low cost. When you see PCs, just take this one, this thing weighs about a pound and a half. I would invite people to come take a look at it. You can kind of open it up. This is as light as anything you’d think about. Nice big screen, Kindle reader on here from Amazon. You’ve got everything you want. This is an amazing, an amazing device. A little pricey, probably, for most people today, but not out of control. And we’ll start to see these things continue to come down.

The same thing in the case of TVs. I’d love people to put one of these babies next to every television, and let every television have the smarts, not just to play video games, but to give you the kind of experience that Simon had a chance to show, with the kind of user interface, and smarts that people really, really want.

We start with Windows at Microsoft. It’s the most popular smart device on the planet, and our design center for the future of Windows is to make it one of those smarter devices that the cloud really wants. How do you take the billion-plus people on the planet who use this smart device, and how do you make it more and more valuable in the world of the cloud?

How do you take those technologies, and push them into new devices, at least new for us? The phone, we’re kind of in the process right now of re-launching our activities around the phone. We have a new version, our Windows Phone 7 Series, which is a smarter device, but really designed for the cloud. Earlier versions of our Windows phones, I think you’d say, were really designed for voice, and kind of the legacy world. But, as we bring our Windows Phone 7 Series to market, you’ll start to see, I think, some very interesting things, really bringing together these technologies.

I’m not going to walk you through much of this, but what we’re really trying to do is put people, and places in the way that Blaise discussed, in the way that Simon discussed, people, places, content, commerce, all front and center for the user with a very different point of view, perhaps, than some other folks. And we’ve gotten some early attention around that.

A big issue for us, because our opportunity now is to say how do we take the next step to make smarter devices, the kind that the cloud really wants, phones, PCs, TVs, and other next generation devices.

Dimension No, 5, the cloud drives server advances that, in turn, drive the cloud. If you really stop and think about the cloud, originally it was just a bunch of servers. Right, people bought a server. The first Internet site in ’69, there was something that was kind of like a server on each side, and they worked together. And then it was PC servers, and more and more servers, and more and more and more servers, and more and more and more and more servers, until today there’s something like 2 million servers sold around the planet just to power the cloud.

And yet, how we think about everything to do with server hardware and server software now needs to change based on the cloud. The cloud has really driven a perspective that comes from scale. The numbers of servers, the sheer numbers of servers, the amount of data that gets stored, we’re trying to digitize every picture, every image, every video on the planet in the way that Blaise showed you, and that’s a small subset of all of the data that will be available online.

The scale in peak load volatility, you get Web sites that have no action, then, vroom, they scale up, vroom, they come back down. Geographic scale, believe me we had some geographic scale issues with Sky TV, thank goodness they’re not trying to sell us that service in Seattle, WA. But there are scale issues of geography. There are scale issues of hardware and people costs. The number of times you want to deploy, remember, this is the world of the Internet, this is the world of the cloud. Things move quickly. And you need to have server hardware and server software that allows you to rapidly deploy new capabilities, new functionality, new data, and new releases.

So, the cloud is changing the way we think about server hardware and software. If you walk outside the front door over here, we brought a cloud. A cloud in a box. You can see a next generation server, it’s a container that’s sitting there. It includes the equivalent of about 10,000 servers. It’s cooled, it’s a next-generation concept. We used to have to stick fire hoses into these things to cool them down. Next generation technology, you can put a garden hose in to one of these things to cool down. It’s hardware technology, and it’s software technology.

The original software concept is, let’s just virtualize everything. Virtualization, virtualization is great if what you’re trying to do is give a little bit more agility to software people and IT people with yesterday’s ideas. But, how do we take the next step, and help people design, how do you design applications that immediately make sense in the cloud? That was kind of the quest we set on as a company, and others have set on. We set on it when we started Windows Azure. Dave Cutler, who actually spent some time here at the U, he’s been at DEC, at Microsoft, and really out trolling, what are the big ideas in terms of the way we change how you write software?

The goal can’t be to throw out all the world’s software and start again. So, you want to be able to migrate software. You want to be able to build on the skills. But, when you write an application, it should know what physical piece of hardware it runs on, where it runs in the world. It should be deployed essentially instantaneously. If a machine breaks, that shouldn’t be your problem, nor should it be the problem of the person. There shouldn’t be people babysitting all of these machines.

So, you start with that principle, and you design different development tools, different hardware, different management tools, and there is so much innovation that will go on in that area. We have a project here at the University of Washington that we call, I think, just Azure Ocean, where the world’s oceanographic data is being collected and put into an Azure cloud. And what does that mean, how do you – that’s a very large data set that just keeps expanding with the sensor data that comes from around the world. I’m sure a very exciting period, the last couple of weeks, with the earthquake in Chile for the crowd that’s involved with it. But that’s going to continue to grow.

And there’s peak load, there’s data coming in. What do you do with it? Where do you go with it? Super, super important stuff.

We’re trying to drive that in many ways, but the catch point here is not only that the cloud drives server advances, but those, in turn, are now starting to drive the cloud itself. We’re asking the question, how does the cloud become something not just that Microsoft and four other companies run on behalf of the whole planet, but how do we give the cloud back to you, to users, to institutions? How do you instance your own customer cloud? You should be able to, if you want, run your own cloud. It may not have 10,000 servers in it, but you should be able to buy a refrigerator, maybe, instead of a container worth of capacity. You should plop it in someplace here, stick a hose in, stick a little bit of electricity in, and a network connection, and that should be a cloud that you can use.

You might say, why wouldn’t I want to put all my stuff in the public cloud? It could be for integration with other systems. It could be for a wide variety of reasons that might be important to you, but we certainly see that businesses, governments, are now clamoring to buy the cloud and put it in their own facilities, or have it run in their own countries, or have it run in their own states, and own municipalities. And so, this notion of the cloud, certainly the cloud has changed the server. And now through the private cloud, or the customer cloud, the cloud itself is being morphed by these innovations.

Those are the five key dimensions that I wanted to talk about today. Let me give a little bit of context about Microsoft in that picture. And I would say simply, the cloud fuels Microsoft, and Microsoft fuels the cloud. We’re, what, 10 miles from here to our headquarters. We employ across the globe about 40,000 people who are involved in building software, 40,000 people. And if you ask those 40,000 people, or if you measure what those 40,000 people are doing, about 70 percent of the folks who work for us today are either doing something that is designed exclusively for the cloud, or is inspired to serve the five dimensions that I talked about today.

When you buy a new crate and you put it in a data center, is that cloud computing? I can’t even tell you the private cloud versus the next generation of server and enterprise computing. But, about 70 percent of our folks are doing things that are entirely cloud-based, or cloud inspired. And by a year from now that will be 90 percent.

We still have existing customers, and we’ll serve them well. But, the inspiration, the vision, what we’re doing, how we’re thinking about delivering it really builds from this cloud base.

Windows, we shipped Windows 7, which had a lot of work to do that was not cloud-based. The inspiration for what we’re doing now starts with the cloud. And Windows, driving smart devices that the cloud really wants. Windows Phone, Xbox, we had a chance to see. Bing, of course, is a service that was born of the cloud. Windows Azure and SQL Azure, which are really the products that succeed to Windows Server and SQL Server start with the cloud as their design point.

Microsoft Office and Exchange, and SharePoint, which are the back-end facilities in that professional set of tools are all really focused in on the cloud today. This is the bet, if you will, for our company. It’s an interesting environment. Whenever you get a big shift like this in our industry, which comes every few years, 5, 10, they don’t come every year, and they don’t come every 40. The Internet came 40 years ago, but we’ve probably gotten three, or four, or five, or six great opportunities out of the Internet and the microprocessor since then.

But, those are interesting times. They’re times certainly when people want to say about companies like ours that are larger, that have had some success. Can they move, can they turn, can they dial in and really focus and embrace kind of the new opportunities and the new disruptions. And for us that’s where we’re built today. That’s where we’re programmed. In our industry I tell people you shouldn’t get into this business if you don’t want things to change. That’s true at the individual leave, it’s true at the company level. The technologies that are really relevant today are not the technologies that were state of the art when I joined Microsoft 30 years ago.

I mean, if you just look at the allocation, I’m sure, of tenured slots, and the faculty here, the disciplines that people are involved in looked different. Yes, there was a department called systems a bunch of years ago, and there’s a department called systems today. But, the field of endeavor keeps moving forward. And for all of us companies, and individuals alike that is a great opportunity.

For many of you, you’ll say, hey, not that important to me. I’m just in school, I’m coming out, I’ve got a lot of great ideas, and all I’m telling you is you’re entering one of the most fantastic industries on the planet, with the chance to do great academic work, to build great products, to be involved in great commercial opportunities. And I will give you a plug, as the biggest local employer, we’d love to have you come do that at Microsoft.

By the way, my e-mail address is [email protected], send me a resume if you want to test that proposition. But, seriously, this is the time, it’s the opportunity and the cloud forms the basis between the microprocessor and the Internet, we did give the gifts that never stopped giving. And they’re giving us the cloud today, and as I like to say at Microsoft, for the cloud we’re all in. And I think for all of you and all of us who are participating in this industry. It’s just a great time to be all-in and really drive the next generation of technology advances.

I’m well over the time that I wanted to speak. I appreciate your patience, but I’ll look forward to your questions, and comments. And again, for my first talk here in the Allen Center, to all of you up and down, I say thank you it’s been a real pleasure for me to have this chance. (Applause.)

I think there’s a couple of mics. And if people have questions, go to the mics. Or if you’re not at mic level, shout out or something. I won’t even tell you to raise your hand, but just shout out. I have pretty good ears, and I’ll repeat any questions people might have.

Yep, Ed?

QUESTION: OK, I’ll take the bait. I’d like to hear a little more about privacy and security. In some sense there’s a market failure. You can’t really measure how good it is, therefore you can’t really sell it. And it seems, as you said, that it matters more to the major companies because there’s reputational issues. But how do you convince people that the cloud is secure? How do you – it’s not just making it secure, it’s a matter of giving people confidence that they can believe in the privacy and security.

STEVE BALLMER: Yeah, I think it’s going to be an ongoing challenge and an ongoing area not only of technical innovation but of communication innovation. The cloud is probably – if you take the population at large, some people think the cloud is pretty darn safe, private, and secure, some people really know that it is less safe, private, and secure, and some people don’t think about it all that very much.

And the question isn’t, is there one privacy and security standard that fits all. The question is, can we give people the tools that let them feel in control, let them feel responsible and then go back that up.

And I think people know this when they see it. The question right now is most users really don’t know what’s going on. I mean, I won’t – how do I say this? Let’s not pick on – I can pick on ourselves, but I think there are some celebrated cases where people don’t – didn’t even know what’s available.

I went to my kids the other day, and they said, ‘Oh, dad, you know, we don’t interact with you socially on the Internet.’ And I said, ‘Well, I saw the following five pictures, I know you were at this party, and I know you’re friends with so and so.’ My kids – and my kids are, you know, they’re reasonable kids with good brains, they know what’s going on. And so really making the tools and the technologies that make it easy to manage kind of this interaction is important.

In-private browsing, the example I used earlier, it’s a good absolute standard, it’s either on or it’s off, and yet there are many cases that are far more nuanced.

So, we need more investment not only in the underlying technologies, but in the user interface that surfaces it.

Yes, sir?

QUESTION: Steve, this was great. Finally we understood what cloud is. So, if you go around and ask us students, I’m sure you’ll get better answers.

But my specific question is, first of all, also we would like to see you more often here. This is a good start.

What – my specific question is that it sounds like this cloud is much better platform for large research groups like we have here. We have international collaborators, as well as national collaborators. I personally would really like to have in contact with them, not like a Twitter or not like a Facebook, because they are really, in my opinion, impersonal. I’d like to be with them in their offices almost all the time. They don’t like it, but I like it. (Laughter.)

STEVE BALLMER: We’d better give everybody the user in control features they need, but go ahead.

QUESTION: But I think Ed also asked, this private issue is very important, because they want to keep their privacy. We also want to keep our research groups’ privacy.

So, is there a way that we actually as large research groups work with you to develop some of these technologies, so that you could actually implement that in other settings?

STEVE BALLMER: Go ahead, sorry.

QUESTION: So, the specific question is, do you have in this 40,000 people working for you a certain percent on this project probably, can we actually increase that to 100,000 so that we can also join in a very collaborative fashion?

STEVE BALLMER: The opportunities to do collaboration between research institutions in general and the University of Washington specifically and Microsoft around these issues, which are of mutual importance, I think is quite strong.

Now, it usually takes the right advocate with the right energy with the right relationship on the other side to really drive and be the backbone of that kind of collaboration, but certainly for us with an academic institution, Rick Rashid, who runs our research group, is always probing to find those collaborations. We’ve got more going on at the UW than anywhere else, and I certainly follow up, if you like, I’ll do an e-mail introduction if you don’t know Rick, but let’s make sure we’re driving on those dimensions hard.

Yes, sir?

QUESTION: You talked a lot about adding smarter devices to the cloud, and creating smarter devices that the cloud wants. I’m sure that’s going to be a relatively large collaborative effort between Microsoft and a bunch of hardware vendors and things like that. But there’s also sort of me in that picture, the end user, the software developer that’s going to end up using those devices.

I was just curious if you could talk a little bit more on what’s Microsoft’s approach to making those devices available to me the end user. Are you guys working toward a standardized set of software systems to access that sort of thing? What’s sort of your approach there?

STEVE BALLMER: Yeah, I won’t say we have one tool. I’d say we use multiple tools, and I’ll just give you the two examples.

I think there’s been a lot of merit to the way we try to collaborate with many hardware vendors to get things that are relatively standardized and then are consequently offered in much higher volume with greater diversity at lower prices.

And we’ve done that with Windows. We’re going to keep trying to push not only Windows PCs as we know them today, but new form factors. The PC form factor is not a static thing.

In the case of the phone we’re trying to do that. The first time I think we erred, we didn’t standardize enough, and the cacophony that came in front of you, the user, was too high.

On the other hand, I think there’s an opportunity to do things that provide more options than some of our competitors who do a very nice job, like Apple and RIM, but you do wind up getting exactly what they choose to build for you, and it doesn’t allow the same level of hardware innovation that might fuel a number of interesting consumer scenarios.

So, in the phone case we’ll take the Windows kind of counsel, but we’ve got to recast it, I would say, from where we’ve been.

In the case of the TV we’ve got both strategies. We actually have a TV implementation in some senses built into Windows. It works really well for small screen TVs that you might call a PC, but for that big screen device here’s a piece of hardware that we build, there’s no diversity. You get exactly the Xboxes that we build for you. We may have more form factors in the future that are designed for various price points and options, but we think it’s going to important.

And when you see this Natal camera that comes with the Xbox this Christmas, I think what you’ll wind up saying is I’m glad they’re doing some hardware, too, because it permits a different kind of innovation, and we want to have both of those muscles working on behalf of you the user and the developer.


QUESTION: Hey, Steve, hi. Great job. We really enjoyed your talk.

And like you said, Microsoft goes back 30 years. I think in the last 10 years or so there’s been a lot of misses with Microsoft being late to certain technologies like the Internet. And then looking at what you’re doing now with Azure and with companies like Amazon has done, and maybe SalesForce has done, and especially Google, and does it seem like maybe your strategy is more reactive to what they’re doing out there, and how do you differentiate yourself and the company and your services with what Google is doing with their app store, and with Amazon Web Services and with Force kind of being a big innovator, and how do you bring more value to the entire space as opposed to just being an innovator? Because you’ve had some key people just in the last year leave your company to go to Amazon and other competitors. So, how do you guys keep the innovation going as opposed to just being reacting to what maybe is happening ahead of you?

STEVE BALLMER: Yeah, I think all great companies have a mix of proactive and reactive muscle, and you’ll use it in different times and in different ways. Nobody ever wants to have to react to anybody else. Everybody would love to have invented everything, and been first with everything, and that’s probably not practical.

I’m certainly keen on increasing our hit rate in terms of early and often, and everybody should be doing that in all companies.

You know, if I take a look and say, hey, look, where am I proud of where we are relative to other guys, I’d point to Azure. I think Azure is very different than anything else on the market. I don’t think anybody else is trying to redefine the programming model.

I think Amazon has done a nice job of helping you take the server-based programming model, the programming model of yesterday that is not scale agnostic, and then bringing it into the cloud. They’ve done a great job; I give them credit for that.

On the other hand, what we’re trying to do with Azure is let you write a different kind of application, and I think we’re more forward-looking in our design point than on a lot of things that we’re doing, and at least right now I don’t see the other guy out there who’s doing the equivalent.

Certainly in the case of search we weren’t the first guys to move. Hey, frankly, even Google wasn’t the first guys to move, but they were the first guys to achieve commercial success, and you’ve got to give them a lot of credit.

But I don’t think anybody should want us to go away and stop innovating. There’s been more innovation in the search market in the last year than in the preceding three. Why? The old fashioned American answer: Because there’s been competition. And we may only have our little 11 percent market share, but, vroom, we’re keeping our heads down, and we’ve been hiring a lot of great people, including a lot of great people right here at the university, vroom, and we’re trying to continue to push forward with great ideas.

Obviously, always opportunity to improve, but when it comes to most of these cloud dimensions, I feel like we’re at the front or tending to the front. And there’s a few where we’re – you know, definitely in the phone case, I kind of explained that, and certainly in the search case we’ve got work to do, but in the TV and some of these other areas I really think we’ve got some pretty exciting leadership stuff.

And we continue to get great talent. If I look and I say, yeah, occasionally somebody will go from us to one of our competitors, sure. Does occasionally somebody come the other way? Sure. But at the end of the day, the thing that’s really going to be the key for us is what percentage of the best and brightest do we get who are going to graduate off of this campus and equivalent campuses this year.


QUESTION: Yeah, thanks so much for your presentation. And as you said, this is a fascinating time to be a part of technology. Things are just going rapidly, increasing at an incredible rate.

Ray Kurzweil has written a book called “The Singularity is Near,” a very thought-provoking book. And I’m just curious about your viewpoint on that, and basically the book talks about the exponential growth of technology. It seems like this cloud computing that you’re really focusing on at Microsoft is going to contribute to that exponential growth of technology. Can you talk a little bit about your viewpoint on a singularity and really your anticipation of really this cloud computing contributing to global growth in technology?

STEVE BALLMER: Well, the book is still on the nightstand, I’ll be honest, but let me talk generally about this notion of contribution and exponential growth. And let me just take one field that Ed Lazowska and I were talking about, which is e-science.

Many of the world’s problems come down to science. Bill Gates likes to talk about climate change. Climate change is a huge issue. And yet if we’re also going to help many of the disenfranchised people on the planet come out of poverty, the world will consume more energy five years, 15 years, 50 years from now than it does today, despite the best conservation efforts that we will have in countries like the United States.

So, how do we make these two things come together? The truth is we need science to move faster: energy science, climate science. We need to speed up the rate of scientific innovation. And the best hope to do that is actually things that come out of the information technology world.

You talk about cloud computing and research efforts and global. The ability to collect data, to model and simulate the real world, to run experiments quickly, even our search team will tell you their number one constraint to improvement is how quickly they can run experiments on how well they understand user intent and semantic knowledge, because in a sense they’re trying those problems.

The oceanographers will have a set of issues, the client scientists will have a set of issues, the people who are exploring and looking for new sort of natural gas sources and other cleaner energy alternatives, a lot of this is the speed of scientific exploration. And I think technology itself has an exponential path in front of it, and the ability of that to propel science in ways that are important to society I think has never been better.

I’ll take two more. I think I’m close to 75 minutes now. Yeah?

QUESTION: Great, thank you, Steve. My name is David. I’m a student here.

We talk about the cloud like it’s one monolithic thing, but it seems to me that there’s lots of maybe little clouds. It’s cumulus clouds. But Microsoft and other companies want to have the biggest, best cloud, the cloud that has the most resources in it, has the most security to offer.

But you talked at the beginning about not being able to find some information about the healthcare debate, and near the end you mentioned that some companies and countries and municipalities all want to have their own little private cloud. How do we avoid their being walls between clouds, as it were, and help people find the information they need?

STEVE BALLMER: Well, let me say I don’t think it’s wise for us to all say the cloud must be monolithic. I think we ought to permit, and I think the approaches that I’ve described, whether it’s in social interaction, in learning, in the way the cloud gets constructed, the private cloud, the public cloud, in some senses everything at least in our approach, and I think our approach on this one is a little bit different because of where we grew up, is to make the cloud something that is very open to different instantiations, different kinds of innovation. If somebody’s got a great private cloud idea that they want to experiment with in oceanography, we should explore that. If they’re willing to be public, that should happen. We need to enable third-party innovation around the cloud.

Now, is it possible some silos will grow up? Absolutely that’s possible. But I’ll bet they’ll only grow up for very good reasons.

I’m sure in every country the military would like to have a more private cloud. I use that as an absurd example on the extreme, but let’s face it, there is a set of things that people will want to keep private.

There’s a set of things where the author should decide what happens. If somebody pays to collect oceanographic data, I hope they make it available to the world, but they should have the right to decide how to treat that as somebody who invested in the creation of something valuable.

So, as we talk about the cloud, whether it’s in the area of the platform itself and the way the server platforms come together, whether it’s in this area of machine learning and data analysis and simulation, whether it’s in the area of interaction, we have to enable and permit the fragmentation that you fear, but encourage people to see the benefits of the kind of thing that Blaise showed you in the demonstration when you can mash together information in exciting ways.

QUESTION: Excellent, thank you.


QUESTION: Hi. Thank you very much for your talk.

So, I’m curious that we shouldn’t care where information is because it should be completely abstracted away, but it seems the laws and regulations do care where information is. I’m just curious how we should manage and take care of that.

STEVE BALLMER: That’s why we talk about a partner cloud, a customer cloud and a public cloud. I mean, I think for a lot of reasons it will be many years before many government organizations will grow comfortable with the notion of their data or citizen data living outside of the jurisdiction.

As technology people we can talk about whether that makes sense or doesn’t make sense, and why the protections can be the same, but it turns out the regulatory environment, as you highlight, is imperfect. I mean, the truth of the matter is – our guys were trying to explain this to me a week or two ago – the same data held in the same place but under different operating circumstances has different regulatory blah, blah, blah, blah, blah.

And we can’t assume all of the world’s important countries are going to even standardize the regulatory framework. That’s why when you walk outside and see one of those containers, it would be OK with me if we have to dump one into every country or sell some to some people who want to implement them.

I love Slovenia, it’s a great country, but there’s only a million and a half Slovenes. This company is not likely to build part of our public cloud in Slovenian anytime soon. So, somebody should be able to implement a Windows Azure cloud in that country. They should be able to buy a device that looks like that or a set of devices and go do that and have that be affiliated for the rapid advance of technology with other things going on in the world.

So, I hear you and I agree that there’s a set of issues, but they don’t have to be constraints.

Here’s just one simple way to think about it. Will all of the world’s centralized compute, storage and networking infrastructure all be built out by four or five companies, Microsoft, Amazon, Google, the cloud guys? Will we buy every server computer and every piece of storage in the world? No, that isn’t going to happen. I don’t think that – if you just think about the level of capital investment that involves.

We need to permit the private cloud, and the kind of thing we’re showing, the kinds of things we’re doing with Windows Azure is about making sure there’s a public version and there’s a customer version, and there can be a government version, all based on the same core technology, and there’s some innovation to go make that happen.

With that, let me end and say again thanks very much to all of you. I want to thank everybody here at the University of Washington for the opportunity, and particularly for everybody whose days we must have disrupted to set up and have this meeting. Thanks again, and if you’re interested in Microsoft, [email protected]. Thank you. (Applause.)