Remarks by Rick Rashid, Senior Vice President, Research
Microsoft TechFest 2008
March 4, 2008
RICK RASHID: Well, it's great seeing you all here today. I must admit, this is really one of the most fun weeks for me out of the year. You know, not only is it a great opportunity for you to get a chance to see some of the great research that we do, quite honestly, it's a great opportunity for me to see some of the great research we do. It's great getting everybody in one place. This is the only time of the year, really, that we get all of our research labs in here, people showing off their work, talking with each other, it's really a great opportunity.
Now, I'll have to tell you a secret -- actually, it's not a secret since I'm telling you -- but I didn't ever really want to do this. In the early days of Microsoft Research, people kept saying, "We really need to do some kind of a technology fair, we really need to show things off." And I kept saying, gosh, you know, that would be such a bad idea. You know, we would be -- you know, we put a lot of work into it, it would be kind of a waste of time. You know, I don't know that anybody would show up. So I kept pooh-poohing it. You know, finally I got convinced that, okay, as an experiment. We'll do it as an experiment.
So we set up TechFest the first year and I was completely blown away by what happened. You know, we had such a great response from the employees of the company, you know, they were excited. I got so many pieces of e-mail not just saying that they enjoyed seeing the research that we were doing but, frankly, saying that this made our employees feel good about Microsoft, it made them feel proud of what we were doing and the long-term investment that we were making.
You know, and then I talked with the researchers and they said, you know, we had the times of our lives. We had people that didn't even take a break to eat because they were so busy talking all the time to everybody that came by. So it was just a fabulous event. So, okay, well, maybe we'll do it again sometime. Needless to say, we keep doing it over and over again, and we've been expanding it all along.
Now, I started Microsoft Research not quite 17 years ago -- 16 1/2 or so years ago now. And when I started Microsoft Research, I would not have imagined that we would be where we are today. We have research labs around the world. Obviously, we started here in Redmond; we were here for five years before we expanded out. But we moved on to create groups in San Francisco, Cambridge, Beijing, Silicon Valley, Bangalore, and our most recent lab that we've just announced. We'll be opening up a new research lab in Cambridge, Massachusetts, (not England -- trying to confuse everybody with the naming). We'll call it MSR New England, and that will open up in July. And so we're really excited about that.
We've grown from basically me to now over 800 Ph.D. researchers around the world. And to put that in perspective, you know, the growth of Microsoft Research, it's like saying that we created a Berkeley Computer Science Department faculty every year for 16 years. That's what we've done effectively. And, really, it's been an extraordinary effort to pull all that together and the results have been extraordinary.
It actually understates the size of Microsoft Research to say we have 800 Ph.D. researchers because we have so many visitors, so many interns, so many post-docs that work with us as well. At peak, when we have all of our interns in the summer, with our post-doc program and so forth around the world, we actually have more than 2,000 people working in our research labs. So it's really a tremendous effort of a tremendous size.
We've had exactly the same mission statement for the entire time we've been in business. And I always emphasize this. I know some of you have probably heard me say it before, but it's extraordinary to have one mission statement for this long and really have this singular vision of what we're doing.
We're about moving the state of the art forward in the areas that we do research. That has nothing to do specifically with Microsoft. What it has to do with is excellence in our field. It means we're out there, we measure ourselves against the work of our peers, using peer-reviewed publications, we're really doing the same kinds of things that you'd see being done in the very best research universities around the world.
Obviously, when we have great ideas, you know, we then want to move those technologies into our products rapidly. And this TechFest event is actually one of the ways that we've developed to do technology transfer, by getting so many Microsoft employees in here, we really use this as an opportunity to move our ideas out, to get people to think about what kinds of new technologies they can incorporate into their products. And every year, we have great leads coming out of it.
In fact, when you walk around the booths, you'll see little card readers at each booth. Those card readers are there for Microsoft employees who come through who want to learn more about a project. They just swipe their card there, we maintain a database, and we'll contact people and talk with them throughout the year trying to build relationships and really find ways of moving technology into our products.
Ultimately, we're about making sure that Microsoft and its products are around ten years from now, and I'll talk a bit more about that in a second. We're organized a lot like a university would be. In fact, you know, I was a professor at Carnegie Mellon University in Pittsburgh for 12 years before coming to Microsoft. We're organized a lot like Carnegie Mellon University was organized, the computer science department, back in the mid 1980s, because that was the model I learned, that was the model I grew up with, and that's the model I feel works really well.
You know, we believe in critical mass research groups. We don't believe in just having a few smart people here and there. We really believe that we need to have an organization large enough to have the kinds of interactions between researchers that really produce serendipitous results, that really make things happen that wouldn't happen unless you brought ideas from multiple areas together.
We believe in working openly, and we publish our research. We're one of the top publishers of computer science research in the world. I'll make a point about that in a second, but the key thing there is that we believe that unless you subject your work to external peer review, you're not going to be able to be at the state of the art for any length of time.
Now, that's not something that all other industrial research organizations believe in, but we believe in it very strongly. We work very closely with universities around the world. I'll talk more about this in a minute, I'm not going to belabor it, but the size of our program with universities is something that, again, doesn't really have a parallel in the industrial world. You know, we have over 1,000 Ph.D. interns working with us each year. Just here in Redmond alone last summer we had 300 Ph.D. interns. And to put that in perspective, the United States only produced 1,500 Ph.D.s in computer science in 2006. So it gives you a sense of the size of our program as it relates to the total pool of students that are out there.
I mentioned that we measure ourselves the way a university might measure it. We subject our work to peer review. I can't remember any other research organization that has had the kind of success we've had at top conferences in getting our papers published. You know, it's now almost rare that a top conference in computer science doesn't have 10 percent of its papers from Microsoft Research.
SIGGRAPH -- we published more papers at SIGGRAPH over the last 11 or 12 years than any other three organizations combined. So it's really extraordinary. There have been conferences, major international conferences, where 30 percent of the papers had a Microsoft author. I mean, again, that's extraordinary. That has not really happened historically. I mean, we work closely with the academic community, we're part of that community. We're involved in the committees, we're involved in the organizations, we pay our dues in working with people.
We've had tremendous results in moving technology to the products. You know, if you look at Microsoft today, it's so changed because of Microsoft Research that it's really hard to say that what are we doing is caused by that versus what are we doing that isn't. You know, if you look, for example, all the work we do in digital media. That really began with a group I started in 1992-1993, looking at interactive TV and streaming data over the Internet. That then turned into the digital media division, and that led to the creation of the Windows Media audio and Windows Media video, the technologies that now underlie playback on high-definition DVDs. You know, it really has made a big difference, and that has had an impact all the way through to things like our Media Center products today.
Our graphics, and 3-D graphics technologies have lead to the creation of the DirectX technology, and had a direct impact on the creation of the Xbox product. Even today, so many parts of our products—if you play Halo 3, if you use Xbox Live, if you use our matching technology for gaming—you know, all those things derive from work we've done at Microsoft Research. We have great collaborative relationships. The tablet PC, all of the technologies that underlie the Surface technology (Microsoft Surface that we've announced) these are all things coming out of Research. And there are so many pieces. Basically, if there's a Microsoft product, there's probably something in that product that came from Microsoft Research, or the product was created with technologies that came from Research.
Now, a lot of people ask: “Why does a company like Microsoft do research?” I don't know how many of you have seen this -- this is a slide called the “tire track” slide. You know, it was in a report of the Computer Research Association and tried to really capture the idea that technology has a long cycle from research, academic research, industrial research, product development, a long cycle to moving from idea to billion-dollar businesses. But if you look at the investments the United States has made in computing technologies, they've had enormous impacts on our economy, enormous impacts on the country and the viability of the country in the international arena.
Well, you can think about Microsoft Research in sort of a similar way. We clearly create a lot of important technology. You know, we're a source of intellectual property. We create new technology and new ideas that are used within the company. We create the intellectual artifacts that become products or become features of products. So that's certainly an important thing, that's a value to the company, and both Bill Gates and Steve Ballmer have been quoted as saying that the return on investment for Microsoft Research is the best return we get for our money compared to any other investment that we make.
We're a great problem-solving organization. A lot of times people say you need research because you could have a problem and you need to have it solved and you need smart people around to do it. You know, well, we do that. You know, we love solving problems. I love solving problems. I really get -- it's a thrill for me to do that, it's something I really enjoy.
Obviously, we're a great early warning system. You know, we'll say things like, "You know, this Internet search thing, it could be big someday." You know? People don't always listen to us. But, hey, we're there to talk about it and we get people thinking. We lay the groundwork. You know, if you look back at some of the early days of Microsoft Research, we were doing work on what eventually evolved into the smart phone and Windows Mobile going back to 1993. So we're doing things like that.
But I would claim that although these are great outcomes, I mean, these are things that are important, but this is not why you do research. And that “tire track” slide earlier. That talks about the results of research and the impact on the economy, but that's not why you do research. You do research because you don't know what the future is going to hold, you don't know what's coming around the corner. You don’t know who your next big competitor is, what the next big technology is, what the next problem that you're going to have to solve is going to be.
Research creates a reservoir of technology, of ideas, of people that can be brought to bear when times are bad, when things go wrong. Research gives you an ability to survive when things change. It gives you agility. You know, if you look back on the writings of Vannevar Bush, who was influential in creating the National Science Foundation, he talks about this. He says we should invest in basic research not really because it's going to have all these outcomes. We'll get these great outcomes. But really, if we have a famine or if we have a disease or if we have a war, we need to have this in reserve, we need to be ready. And look at what happened in World War II. You know, the fact that we had all those smart people helped us win the war. So that's really why you do basic research.
Now I mentioned earlier that we work closely with the academic community. I mean, our belief is that it's not just important for us to do great research, but it's important for the field of computer science to be viable, for the ideas in the field of computer science, to keep driving new value because we're a company that's all about computer science. And if the field of computer science prospers, we believe Microsoft prospers as well.
So we work closely with the academic community. We fund students, we do internship programs, we fund faculty fellowships, we support innovative research and that's critical to us. If you look at our footprint around the world, we have our research labs -- you know, I mentioned those earlier -- but we also have extensive research collaborations with universities all around the world. And, you know, the yellow dots there, those represent joint research labs, research collaborations, joint centers that we created with other organizations.
I'll give you a specific example. I'll take one part of the world, which is Europe, and give you a very specific example of the depth of our interactions there. We created a few years ago something we called the European Science Initiative. You know, today we're engaged with more than 100 universities in the EU. You know, we've got fellowship programs, early career programs for faculty, over 100 Ph.D. scholarships. We've worked and reached out to the scientific community to help set an agenda for science in Europe.
There's a 2020 science report — some of you may have seen it — that came out of a conference we put together of top scientists throughout Europe. And we've created awards and programs to help drive things. Over the last 18 months or so, we've announced three joint research centers in Europe where our researchers are working with researchers in Italy, in areas of computational and systems biology, in France with INRIA where Microsoft and Imbria are jointly tackling a large set of issues, and the most recent announcement was the new joint research center with the Supercomputer Center in Barcelona. So we're expanding those kinds of relationships.
This video will kind of expand on it. Steven Emmott will be talking about the sort of breadth of things we're doing with this European Science Initiative. Please roll the video.
RICK RASHID: Hopefully that gives you a little taste of the breadth of activities we have going on in Europe. One thing that always, frankly, surprises me when I walk around TechFest is the breadth of things that we do. I mean, I think it's the one single fact that I always carry away from this week is, wow, we're working so many different areas, whether on the one hand we're doing things that feel to me like traditional sort of plumbing-level computer science, programming languages, operating systems, hardware architectures, things of that sort, but we're also stretching out to areas that you might not expect. I mean, we're working in things like vaccine design and computational biology, quantum computing.
Sometimes you just wonder, like, well, what does that quite have to do with Microsoft? Well, the answer is it doesn't necessarily have to, although historically we've found that when our researchers do great research, there's almost always positive impact on things that we do in the company. So it's been good for us to be able to take this approach.
Let me just give you some examples. These are things — some of these are things that you'll see out on the floor as you walk around. Over the last decade, Microsoft Research has really been pioneering tools to be able to prove properties of programs of hundreds of thousands or millions of lines. We've really taken initiative in this area, really pushing the state of the art in a significant way.
These tools have both yielded great things for us internally, it lets us develop much larger systems that we might be able to, but we've also begun to distribute some of these tools and make them available. Windows Vista, for example, as part of its device developer kit, has a proof tool -- a software proof tool -- that is used by our device driver partners, so the hardware vendors that are producing new devices, they can now use this tool to prove something like 38 different properties of their programs so that their device drivers can be more reliable than before. This really gives them a tool -- puts the tool in their hands that allows them to do better for their customer.
We've pioneered not just these tools, but there's also an example that's called safety tools, meaning it's able to improve safety properties of programs. We've also been able to make a breakthrough just in the last 18 months or so in what's called “termination properties” or liveness properties of programs. So for the first time on a large scale, we can actually prove whether a program can terminate. We can actually demonstrate a set of liveness properties that complement the safety properties we've proven before.
These new tools that really let us analyze software at a large scale in ways we were never able to do before have led us to look at new ways of thinking about developing software too. And that has lead to the creation of something called "Singularity." Singularity is sort of a concept car operating system, it's a new system that's been built from the ground up with the specific goal of being more reliable, taking advantage of proof technology, of specification technology, of new kinds of system architectures that allow us to compose proved programs or proved components.
Singularity really is a new way of thinking about building systems and although you're not actually going to see Singularity, there's no booth for it. I'm not sure how we would talk about it in the booth because it's an operating system. As of, I think 11:00 this morning, we're making Singularity available to the academic community and the research community for free. So this is really going to be an exciting thing that we're doing for the broad community by making available this unique new kind of operating system.
We're also working with the community to think about ways of doing new kinds of computer architecture. I mean, one of the problems as it has become more and more difficult, more and more complex to build chips, is that within the academic community, it's much harder today to experiment with computer architecture. You know, that affects us because we like to do such experimentation, but it affects our academic colleagues as well.
Well Chuck Thacker, who's really the brains behind the original Xerox Alto designs, or one of the fathers of the personal computer, and John Davis, one of the researchers in our Silicon Valley research lab, have been working with the University of California at Berkeley on a new emulation engine called BEE3, that's Berkeley Emulation Engine 3, and really the idea behind it is to build a computer system that is configurable. Basically, this is a system that they've designed that has four large FTGAs, you know, commercial FTGAs, 64 gigabytes of fast memory per node, many kinds of high-speed input and output ports. And it really will give people in the research community the ability to now experiment with the creation of new computer architectures in ways they would not have been able to before. Effectively, what they can do is program this computer to be another computer or to do another kind of architecture and to experiment with new kinds of algorithms. You will see this—there's a great demonstration of this in the TechFest area.
Now, we're not just working on sort of core computer science areas, we're also looking at the ways human beings interact with computers. And I'd like to bring Merrie Morris on stage now to talk about some work we're doing in real-time collaboration with people using the Internet. Merrie?
MERRIE MORRIS: Thanks. Good morning. I'm excited to have this chance to talk with you about my research on supporting collaborative Web search. So currently, the tools that people use for searching for information online are primarily Web browsers, and the Web sites for search engine companies, but these tools are designed for a single person working alone by him or herself.
That's not always the way that we work. For example, business colleagues need to work together to jointly research projects that they're doing. Students in schools need to work together, again, on group reports and homework assignments. Friends look online together to find joint entertainment opportunities, and family members often need to search the Web in order to do activities like jointly planning vacation travel or searching for medical information for a loved one. So in my research, I'm interested in designing tools to make these sorts of collaborative search experiences possible and pleasurable.
We began this research by doing a survey of about 200 people to learn more about the ways that people currently are trying to collaborate when they search online, and this research helps us uncover some of the obstacles to collaborating with current tools. For example, there can be a lot of unnecessary, redundant efforts when people try to collaborate today. One of our survey respondents articulated this well, talking about a search he tried to do with a business colleague. He said, "It was difficult for us to do our search together because we each wanted to follow different paths through the information, so doing it together we less productive. But when we did it separately, we weren't sure how much redundant information we were gathering." That might sound familiar to you.
Another problem is the lack of shared context for collaborators. Another survey respondent described the time when he and a friend were trying to plan a vacation. They were in separate houses and they were each viewing a Web browser and talking to each other on the phone at the same time. So he said, "We were both searching for hotels online. The difficulty we had was in comparing the hotels. So I would say things into the phone like 'click the third result, see the link on the left that says more? Click that one, now look at the second hotel.' and that was very inefficient." So we're interested in designing tools to help make these types of collaborative experiences more effective.
The first tool I'm going to show you today is called SearchTogether. And SearchTogether is designed for scenarios where each collaborator has his or her own computer. So I'll show you a demo right now, actually. I'm going to open up Internet Explorer. So you can see, this is my computer closest to me. The other screen is showing my friend Matt's computer.
So SearchTogether is a sidebar that runs as a plug-in to Internet Explorer. I'm already logged in using my Windows Live ID, which means I can invite any of my IM chat buddies to join in the collaborative search with me. This list here shows all of the different search sessions that I'm a member of. So I can collaborate with a different set of people on different topics.
For example, my husband and I have a session about which hybrid car we should buy. My mother and I have a search session where we're searching for diabetes-friendly recipes that she can prepare. Right now, I'm logged into a session with my friend Matt where we're planning a ski trip we're going to take to celebrate our successful TechFest demonstration.
So the first thing you might notice, Matt and I are both logged in. And on my computer, you can see the group query history. So here you can see a list of all the search keywords that either Matt or I has entered on this topic. And this is meant to address the issue of wasted, redundant effort that appeared as a problem during our survey, because now I can see what keywords Matt has already searched for. He's searched for "Whistler discount lift tickets." I know that I don't need to do that search myself. If I want to see what Matt's done, I can click on any item from his history in order to see the search results that he saw.
Another way that we can collaborate and, again, be more efficient and avoid redundancies, is through a special search technique that we call Split Searching. So let's say I want to search for a restaurant near Whistler. When I do a split search, half of the top 50 search results for this topic are going to show up in my browser, and the other half are going to show up in Matt's browser. So this is a way that each of us can look at a non-overlapping portion of results to the same query so that we can explore this issue in depth in parallel.
This can also be a really interesting way for novices and experts to collaborate on a single search. For example, novices often have difficulty forming good queries or using advanced query syntax. So the more experienced searcher could type the query, but then the novice could still participate by helping look through and evaluate the search result.
Right now, I'm going to look at a particular search result. This isn't a particularly exciting restaurant, but pretend that it works, and I might type in, "Let's go eat here." And even give it a good rating, like a thumbs up. And so this comment that I wrote actually appears in a special summary that Matt's going to view on his computer right now. So this summary, in essence, represents the product of our collaboration. It shows a list of all of the Web sites that either Matt or I has found on this topic and commented on or given good ratings to. And the summary itself is interactive. If Matt clicks on one of the items, he can easily get back to that Web page. He can also easily e-mail this summary to anyone else as a way to further extend the collaboration outside of our little group.
One of the other difficulties that our survey respondents mentioned was trouble maintaining awareness of the context of their collaborators. So we've added special features that we call "peek and follow" that enable additional awareness of what your collaborators are doing.
So, for example, if you take a look at my computer, I'm going to click this special "peek" icon next to Matt's picture. And that lets me see whatever page he was currently viewing on his browser so I can tell what he's working on right now. And I'm able to do this because Matt has granted me special privacy permission because I'm a very trusted buddy of his and we're involved in this very close collaboration right now.
So I'm going to move on so that I can talk about a different prototype, but I hope you stop by our booth later and I can show you more of the features of SearchTogether.
So if we go back to the slide, SearchTogether was designed for a situation where all of the collaborators had their own computer, but this isn't always the case. It's very common, especially in settings like schools or even in the developing world where computing resources are scarce, for groups of people to gather around a single computer. This is especially common in schools where groups of students will work together to research a topic for a report that they're working on together. As many of you might know, sharing a computer with others can be frustrating. Only one person is controlling the keyboard and the mouse, and it's very difficult for the other group members to contribute.
So in response to that, we've developed Co-Search, which is a special system that is meant to enrich collaborative Web search around a single shared computer. So Co-Search consists of a special Web browser that runs on the shared computer, and that can be connected by either multiple mice or by students' cell phones can connect to the computer over Bluetooth so even the students who don't have access to the mouse and keyboard can interact with the shared browser through their phone.
And I'll show you here a couple of examples of the type of functionality Co-Search enables. So here, this student is using the joystick on his phone to operate a cursor in the shared browser. When he clicks on a browser tab, that page is loaded into this phone so he can scroll through and read at his own pace, because the student controlling the main display was scrolling too quickly for him.
In another example, the student is typing a query keyword into his phone, the same way he would type a text message, and now he's sending that to the shared browser where it's going to appear in a space that we call the Query Queue. So this is a place where members of the group can all send their queries and they queue up in a list and they can pick which query they're currently going to explore on the shared screen. So, again, this is a way for the students who aren't in control of the mouse and keyboard to still have a voice and participate in the search activity.
So in summary, I got to give you a brief introduction to our research on supporting collaborative Web search: SearchTogether for when everyone has their own computer, and Co-Search for when people must share. And I hope that you can stop by our booth to get more in-depth demos of both of these applications. Thank you.
RICK RASHID: Thanks, Merrie. (Applause.)
Now, another theme I think you'll see emerging from a number of the exhibits is this notion of wireless sensing, you know, really being able to -- we're now in a fundamental way able to think about collecting large amounts of information from our environment to be able to process that information and use it in really exciting ways.
One of the devices you're actually going to see if you walk around is this one right here. In fact, I've been told I'm going to be in a little trouble if I lose it because it's one of the prototypes. This is a Microsoft Research-designed wireless sensor that's being developed by Feng Zhou and his research team here in Microsoft Research in Redmond. And really, this is a device that can collect a lot of information about the environment—it can get information on temperature, motion sensing, a variety of different things, plus it has the ability to have other sensors plugged into it.
This is a wireless device, you can spread these out into an environment, they can communicate together and they can communicate with ordinary PCs using standard Web server Web browser kinds of technologies.
The key thing here is this notion that you can use devices like this to collect data, for example, in your home. If you want to be able to use energy more efficiently, we've been using technology like this to collect data and understand what happens in our very large data centers, which can save us an enormous amount of money and reduce our power footprint to make our data centers more green. So technologies like this are very exciting. We're seeing more and more of this happening.
Now, you can imagine distributing sensors like this into an environment as well, and we're working with a number of groups and a number of scientists to do just that. I mean, this is an example of us where we're working -- this is Catherine van Ingen in our research lab in the Bay Area. You know, she's been working with the University of California, Berkeley, and a number of other groups to help build out what's called the Berkeley Water Center. And the idea is to be able to collect huge amounts of hydrology information in Northern California, be able to understand what's really happening in that ecosystem, and be able to do modeling that can begin to predict how water resources change, what's happening to water, what's happened to water quality.
So this is exciting work. They've built a collaboration SharePoint Server that's currently serving over 900 site years of carbon climate data, and this is really helping environmental scientists understand more about the world around them.
Another example is some work you also see here at TechFest, which is we're working with Project Neptune at the University of Washington and a number of oceanography groups around the country to help build a platform for being able to process scientific workflows. In particular, these are oceanographic workflows. They're in the process of distributing thousands of kilometers of sensors on the floor of the ocean off the Strait of Juan de Fuca to really be able to measure what's happening in the deep ocean at a level of granularity that no one could ever really do before. And we need the tools to be able to do that, to be able to process the information. You'll have a chance to see that at TechFest.
Now, the last thing I'm going to talk about is something I think is really exciting. I know some of you -- some of the reporters who were at the reception last night had a chance to see this. This is very exciting work that is really a follow-on historically to the work of Jim Gray in the Terra Server and the SkyServer. There's work by Curtis Wong and Jonathan Fay where they've really been able to build something called the WorldWide Telescope, the new way of thinking about looking at the sky, being able to collect all the great data from the great telescopes around the world, be able to pull that into a single view that looks like the sky, that gives you a sense of being able to look out into the sky. So you're not confined to just using telescopes, you're not confined to data from only one data source. What Curtis and Jonathan have done is really pull all of these things together, working with top astronomers and top astronomy laboratories around the world.
So with that, I'm going to let them run this demo. By the way, this demo was actually recorded. It's running within the application itself because the application supports the notion of tours, it was recorded just a week or so again, and the voice you'll be hearing is Roy Gould, one of the top astrophysicists in the world at Harvard. Go ahead and play it.
RICK RASHID: Okay, I'll admit it, my kids have had access to this now for a month or so. I've got a nine-year-old and a seven-year-old, they're just excited. The educational value of this I think is going to be tremendous. Curtis Wong is actually in the room. Curtis, do you want to stand up? You know, Curtis is the one who's really led this project, and I think I saw Jonathan here earlier, but I think he may have gone back to the demo room. This has been an extraordinary effort, I think you'll really be excited when you have a chance to see it here, and it'll be made available to the public towards the end of the spring, sort of late springish time frame.
Now, as you walk around, you're going to see projects in all stages of development, you know, early stage, middle, some of them are closer to potential productization. People always ask me, like, well, what's going to be the next big thing? Is this really going to be great? What's really going to happen to it? What you're seeing when you're at TechFest, whenever you visit a research lab, you're seeing raw materials.
Scientists don't really create the future. We don't decide what the next big thing is, we don't even usually know what it is. You know, what we do is we create the raw materials. People really create the future from that. We generate artifacts, we try out ideas, but it is the product creators, the developers, the program managers, it's the individuals in the way they use the software, it's the societies in which we live, they really decide what happens, they decide what's going to be important. And so as you walk around, you should keep that in mind.
Thank you very much. (Applause.)