- Feature Stories
- News Flash
- Making Noise
- The Fence
- Beyond the Cut
- Inspire Innovation
BMOC #1: Bill Gates« Back to Page 1
If we think of user activities in the home environment, that will also be very different. TV today is passive, channel-oriented type experience, and we've finally gotten about a million users connected to TV shows through the Internet with partners like AT&T, and about 20 of the other big phone companies around the world. Once it's connected up that way, the ability to make the show personalized, so on the news segment you see the subjects you care about more, the subjects you don't care about less. If you're watching the Olympics, the sports that intrigue you, you can ask for more information on, and skip over things that are not of interest to you. Game shows, sports shows, educational shows can all be a lot better. Advertising, which is the life's blood of TV by being personalized and interactive can also be more relevant to the user, and more beneficial to the advertiser as well. And so TV is on the cusp of a big change. We have the best content providers around the world now thinking through for each of these genres how they take our tool kit and take their shows and present them in this richer way. And we won't have a dichotomy between video on the PC and video on the TV, those two things will be brought together. So as we construct your personal guide, if we know that you have a relative who is doing some sports, even if it's very obscure, and a parent just happens to take an HD camera and posts it to the Web, we'll know that that belongs there on the list of things that you might be interested in.
Even education itself, which hasn't changed a lot, as we take and split education to the lecture piece, the study group piece, and the accreditation piece, the first and third piece, the lecture and the accreditation can be done on a scale way. Already the very best lectures in the world, the video of those are being put completely online, available to anyone, and it means students around the world are coming up and accessing those things. I recently was confused about solid state physics, so now MIT doesn't even know it, I'm taking a solid state physics course. And it's absolutely fantastic. I watch it when I want, I do the homework problems, I send e-mail to my friends when I get confused. It's a whole new way of getting access to information, and thinking about these different elements. And so I think something profound is going to happen there as well.
So software will change a lot of activities. Now software is becoming important not just for these direct uses, which you think of, okay, this is what the computer science department should be doing, programming data centers, natural user interface. Software is also important to the other branches of science. They're dealing with so much information that the databases, the queries against those databases, and the pattern, the need for machine learning to figure out what is going on with that data is absolutely essential. We were talking this afternoon about some work being done at this university where a lot of people are now focused on computational biology. What do those queries look like, what do those structures look like, how many levels of semantics are there in that? Without software, these problems are intractable.
Even a science that at least I used to think of as very simple, astronomy, is now a data-driven science. It's no longer about being the lucky person who is there at midnight and is looking at the telescope and something really cool happened up there, and you write it up, and you're a famous astronomer. Now there's all this data, various wavelengths, and various times, and whatever proposition you have about how the universe works needs to be tested against all that information. And so making it easy for a scientist anywhere in the world to take data from any of those telescopes, or satellites and various things and try out their ideas, that's an important foundation for that science to move forward.
Organizing that information might be the most difficult in biology, because the range and the amount is very tricky, and one piece of that is studying the brain. Microsoft Research is trying to reach out, not just to computer science departments, but to these other science areas as well, and have them tell us what they'd like software to do in modeling, in visualization, in data structuring, and use that to push the limits to help us understand where new inventions are necessary. The brain is a great example of that, so I want to show one collaboration just as an example, and this is one I hope it's not politically incorrect, that's with Harvard, and a brain scientist there named Jeff Lichtman. So let's hear a little bit about the challenge that he faces.
So I've got here on the screen now this software called HD View that takes that scan data that they have from the mouse brain I'm sure you're all looking forward to seeing the mouse brain and puts it in a form where people can navigate through it, and apply software algorithms to finding out what's going on. Dealing with images like this, you need high performance, we make it easy to get at zooming in, panning, looking at it in different ways. We can also take these different layers. So that's the different slices layered on top of each other. If I just focus on one at a time, you'll see as I go layer by layer I can see the different things that are going on there. Visually navigating and looking for features is part of it, but having software be able to efficiently get at the data is perhaps the most important thing.
Here we have software analysis that went in and found various boundaries in the different layers, and now it's trying to go layer by layer and see how those connect together, and then it's able to recognize where the neurons are, how they connect up to each other, and then start to form a model of what's going on in the mouse brain, how it changed over time, what the levels of functionality are.
Obviously, we're decades away from the total deep understanding there, but it's going to take software analysis, and software tools like this, connected to those domain experts, in order to make the kind of progress that we're interested in. That type of deep collaboration of the software experts, and these domain experts, is a real theme. And I know Carnegie Mellon has been a pioneer in thinking about these interdisciplinary problems, how do you make sure that you can expose people to enough computer science, and have in one person a lot of this knowledge, and in groups that are working together those capabilities.
It really is a fascinating problem, but a necessary problem to solve. It's almost like mathematics was, say, in the early 1900s, where the it became such an important tool for describing what was going on that people in the sciences, like it or not, needed to become familiar with those techniques. And the real breakthroughs in the 1900s, across many areas, were partly because of those mathematical representations.
I want to talk now about a special issue in terms of the impact that software might be able to have, and that is taking the perspective of the 6 billion people on the planet, and trying to look at how software will impact the top third, the richest 2 billion, the middle third, the next 2 billion, and perhaps most interestingly the bottom third, that 2 billion.
In some ways people think, well, can't we just make things cheaper, and scale them down, and that will have a very nice impact. In fact, in places where you don't have electricity, you don't have training, you don't have a schoolroom, or a teacher, it's not going to have much impact. The needs, the problems, the priorities are different than simply saying, hey, let's just take what we have and make it lower cost.
Making it lower cost is a fantastic thing, that's much more an issue, though, for that middle 2 billion, and the industry has done a great job bringing these things down. It's really the connectivity costs, and the training costs that dominate everything else. The hardware costs, software costs, and all of these things, the software is donated for these outreach type projects, the hardware has gotten fairly inexpensive. In fact, we have a cell phone now that you can connect up to a TV set and get a large screen display. So you're looking at cell phone type costs, even to do full-screen type applications. But, then you still need somewhat of a network. We can often piggyback on the cell phone networks, but there are cases where that's not even there.
We think what's going on with this the most challenged 2 billion, there are a variety of things that have to do with their farming production is very low, 64 percent of these people are rural, small farmers. The amount of nutrition they have is very, very low. The level of disease is very high. If we take the big, infectious diseases, 90 percent of the effect of these diseases is on that bottom 2 billion.
In fact, that's part of the reason why we have what I would refer to as a market failure. That is, the market directs itself to solve problems based on economic signals. And so the top 2 billion can send very strong economic signals. A good example of that is the top 2 billion, they don't like being bald, so billions are being spent on curing baldness. Nobody is dying, but, boy, they've got their money out on the table, and that's great. Some day I think they'll solve that. The bottom 2 billion, 1 million children a year die of malaria, and yet there's less than 10 percent as much put into malaria research as there is into baldness research. And you think, well, is that appropriate, or is that a market failure.
How do you solve a problem like that? I'm not someone who is suggesting that we mess up in some dramatic way the incentive system that has done so well. After all, if we look at the last 50 years the incentive system we have actually has done very well, not for everyone, and we need to be aware of that, but broadly the fact you can even talk about that middle 2 billion, that is an incredible success story, 50 years ago there were the rich countries and the poor countries. And now we have a complete distribution, even within countries we have a very interesting distribution.
Unfortunately, if we look at that bottom part, there are a few areas that are heavily represented, particularly sub-Saharan Africa. Parts of Asia, but most of those areas have seen rapid improvement. If you take out, say, Afghanistan and Yemen, and you're really looking at the toughest things, you're overwhelmingly talking about sub-Saharan Africa.
So we need to do something to try and get this innovation to apply to these problems. And I'm really pleased to see that there is a growing awareness of those problems, and a growing willingness to put energy into thinking about them, and thinking about where innovation can help.
For myself, I left the university without any awareness of the conditions of poorest 2 billion. If you had asked me I would have said, yes, I guess they have less food, less electricity, and things, but I didn't really understand the trap that exists there, and the nature of those problems. It was 15 years later that I was reading a newspaper article that talked about rotavirus and said it killed a half a million children a year. I thought, well, what's rotavirus, I've never heard of it. Is somebody working on it? It turned out at the time basically nobody was, because, again, it's a case where there's no signal that says that's something of importance, and it ought to be pulled together.
So I think if you go back 10 years ago it was very easy for somebody going to college, even somebody going to a medical school, where you think a lot of these problems would be highlighted in a fairly dramatic way, it was very possible not to know much about them. I think that is changing. I see the interest level, of course, the availability through the Internet to learn about these things, the desire to get out and see some of these conditions is much stronger today.
I was meeting with some of the students who are working on the thing called TechBridge here, which is aimed at thinking through what are those needs, and where they can be addressed. I've seen some wonderful projects that are actually getting out in large numbers. And sometimes they're not cool because of the technology, they're really only cool because of the impact. One that I'll site we call Digital Dream, and this was done by our India research lab that has a particular focus on the needs of they call it the bottom of the pyramid, those 2 billion.
They went out to these farmers and saw that a direct use of computing wouldn't be possible, no cell network, no electricity, and that the real problem you want to solve is the productivity, the farming productivity. If you have a drought, if you haven't grown enough the season before, then you literally face malnutrition that has a lifetime negative effect, or in the extreme case even starvation. So they saw that there were techniques that if these farmers applied they could more than double their output, but the current extension system just wasn't getting the message to them in the right way. So the adoption rates for typical advances among small holders was about 15 percent.
What they did was they did was they took the advanced technology of DVD and they went out and took films of the farmers doing it the right way, in local language, brought those back, edited them, took the best practices, and hired people who knew these farmers, were socially connected to them, to go out with a battery-powered DVD device and showed those videos. And what they got was a tripling in the adoption of these farming practices.
So a significant impact in terms of the nutrition, and the ability to live through bad seasonal effects, in fact, they even created this wonderful dynamic that the farmers wanted to be in the video. It's a lot like American Idol, but this is Farmer Idol, I've got this technique, you've got to get me on there. Competition in something like that is a wondrous thing to see.
Now, that's just one vignette, there's a lot in health, in micro-finance, education, so much to do and the failure rate of many of these projects will be very high. The ability of these tough conditions to defeat things that even you might think would work is daunting, and yet if we keep trying, if we have the awareness, we get out there and learn about it, I think we can do phenomenal things. So when we think about why software, and all these technology innovations are magic, I think we should feel the best about it if we're making sure it's magic for each group of 2 billion in the world.
So I hope you've gotten a sense of the incredible optimism I have about the advances that will be made. In fact, I expect a lot of them to be made by the people here. I think you're in this field at really the most amazing time ever. And so I'll be fascinated to see the wonderful things that you do.
Thank you. (Applause.)
MODERATOR: Now we'd like to do a question and answer session, and since professors are sometimes defined as people who can talk about anything, but it always takes them at least 90 minutes, what we'd like to do is have students, only students, ask the questions. The other thing I'd like to say is, we'd like the questions to really be questions, and to the point, so that Bill actually has a chance to answer them. So please. There are six microphones, and we'll hear them all. Please come up.
You're not a student. Please step on up.
QUESTION: Hi, this is actually a question about the Bill and Melinda Gates Foundation. I'm someone who really likes the kind of idea of the Foundation work, and I was looking online and when I was looking at potential jobs with the Foundation all of them you need 10 to 20 years of experience, yet you're telling us we should go out and make sure we do really good things. I was wondering, how do you expect to be able to go do good things if you can't have any starting point?
BILL GATES: Well, the Foundation today has about 400 people, and over the next three years that will grow to about 1,000 people, but clearly the main thing we do is grant to partners. The grant level is about $3 billion a year, and it's a lot of partners doing the research, doing the activities out the field. So I'll check that job thing, we should be more open to undergraduates, and summer-type activity, I'd expect I think maybe a third to a half of the openings should be getting young people in, getting that exposure, and so that they can be involved in these things. I'd say overwhelmingly, though, even as I look at that, it would be the partner organizations, there would be even a broader set of opportunities. So thanks for highlighting that.