Bill Gates Conversation Transcript

Well, I have the full transcript of our hour with Bill Gates. It’s pretty long, so I figured I’d post it in sections. As I said earlier, Bill came in and talked to us first for 10 minutes or so. Here’s what he said:

(Note: The transcription was done by someone in a different location listening in on a conference phone, so there may be some things he didn’t get or heard wrong. But for the most part it looks pretty good.)

Okay. Well, great. Well, I saw your agenda, so I’ve already learned a lot, and I’ll just say a few things and then let people ask whatever questions they want.

You know, we started believing in the magic of software back when it wasn’t a very mainstream thing, computers weren’t used very broadly. And even then the idea of revolutionizing how you would get at information and how you’d interact with things was part of the dream.

We’ve come obviously a long way as an industry. We’ve got a good sized software industry, and we’re revolutionizing most everything with digital approaches broadly.

Most recently, the idea of how you match buyers and sellers, and how you do deep collaboration, obviously that’s become a defining application for the Web.

In the next few years, things should improve pretty substantially. I mean, we still don’t think of TV as being a Web application, but the right pieces have been put into place to change that. The phone companies, AT&T and Verizon, actually do use Web-based delivery of video, non-broadcast, and, in fact, they’re building infrastructures that allow you to do personalized video, so you can interact with the context, interact with the ads, things that we’ve targeted, high-definition, and they use a software platform called Media Room that we created to do that. That’s one of the things we got into way before its time, have been working on it for a long time, and we’re just at about a million people using that right now. Over time, we expect the cable cos will want to go with a broad software platform as well.

So, the arrival of video on the Web, until you get TV sort of connected up, you’re not getting the kind of scale that you’d really like. I mean, yes, it’s happening, it’s a great thing, there’s some cross-fertilization in terms of how things are developed, but it really hasn’t been there.

There’s a lot in terms of the interaction style, where if you have to interact with the keyboard and mouse, there’s just a lot of things that aren’t simple. You have a lot of companies trying to make breakthroughs on that. For us really an iconic thing is something like Microsoft Surface where you just take information, take how you interact with a map when you have a keyboard and a mouse, and then take when you have Surface where you can zoom in, pan around, try out different things. Take how you would sit down with a friend and look over some photos. Your computer is set up so that it’s not that easy for multiple people to sit there, not that easy for multiple people to navigate around, it’s just not a shared experience; whereas having the tabletop type environment with the natural interface makes that very simple.

We actually have a product that’s not a consumer product, so it’s not as visible as a lot of consumer things, that are called RoundTable, that show some of the potential on this, where you take and you have just digital cameras and digital audio pickups, and so if somebody is remote but wants to have a meeting, a virtual meeting with a group of people, they get to see all the participants, and it automatically directs the meeting in terms of knowing who’s talking, telling them who’s talking, showing that in the main display, but still showing the 360-degree view of everybody in that room.

Our view is that these natural interaction techniques are very complementary to each other; that is touch screens, surface touch, touch whiteboard, pen type interfaces, which you’ve seen on the tablet, natural 3D input, which you see on things like the Nintendo Wii at this point, but you’ll see very broadly as a standard input peripheral over time, not just for videogames but for PCs as well, and that’s one of those push/pull things where the Web is not 3D today, but that’s partly because the peripherals aren’t there, and the peripherals aren’t there partly because the Web is not 3D. Well, why isn’t it 3D? Well, there’s been at least five startups a year that have said now we’re going to make the Web 3D, and then they go and fail. The tools, the performance, the richness, the environments. You know, Second Life has gotten some degree of critical mass in terms of what they do, but still people won’t think, okay, I go to a bookstore on the Web, it’s going to be a 3D experience; I go to a site to navigate what my house is going to look like, that’s a 3D experience, but that will change. We and many others are investing super heavily in the input devices, the runtimes.

And, of course, our friends in the silicon industry are making it reasonable. In fact, one of the reasons 3D failed is that if the refresh rate is not just incredibly high, and the quality is not incredibly good, 2D is better, and you can just read the information and create the illusion of whatever you want in your brain, as opposed to something that literally can make you feel poorly because it’s just jittery and out of date. So, a variety of things will come together there.

So, I’d say TV and the Web, 3D and the Web, the Web being user-centric, so as you’re going up onto various sites, all the things that you do just get stored in the Web instead of stored on the device itself. Why don’t people license more music? Well, there’s a lot of reasons, but one of them is that if your music licenses don’t sort of exist forever up in the cloud, if they’re down on some crazy client, what if you lose that client or you decide to switch to some other operating system, or you want to change something, you might lose it, whereas physical CDs at least you can always go and re-rip those things. That’s sort of eternally available rights for you.

Just moving around between devices, the fact there’s any state that just gets isolated on that device really makes no sense. Web storage is cheap enough that having whatever your favorites are on one device show up on another one, even cross-device where it would be a ticker when you’re watching the TV set, it would be a homepage when you’re on a PC, it would be a set of channels when you’re on a phone-type device, that stuff should be user-centric instead of device-centric, and we’re building a set of services there.

Also in the Internet today, if you want to build an application that’s going to be very high scale and very reliable, you’re basically having to reinvent everything to do that. The vision is that people should be able to just subscribe to a service that takes care of that for them.

Now, no one is offering that today. Amazon offers raw computing with EC2, they offer raw storage with S3, but they don’t offer a scalable model where you just basically write the app and then it scales infinitely. You have to do all the technical work still, because it’s basically a UNIX machine is the paradigm.

The idea of cloud services that take care of fault tolerance, load balancing, and then let any kind of startup just have it be auto-hosted, and then, fine, if they’re popular they pay a little bit for the capacity that is used, but they don’t have to do some brilliant engineering design.

You could even say, you know, why did MySpace over a year ago, why did MySpace beats Orchid? Well, if Orchid or Friendster had performed so that you didn’t have to wait, so you could log in, maybe they would have been the one to accrete the kind of positive momentum. Or if they’d been able to connect up to a general platform, then this idea of extensibility, they wouldn’t have had to generate it on their own, it would just come inherent with the platform that came along.

So, it’s still too hard to write scaled Web sites, and there’s a way of taking this cloud-based computing offering and making that available, so it’s quite easy for people to do things.

This gets very interesting when you get beyond classic Web sites, and you look at certain types of scientific computation problems where the way that the data flows, it’s not as simple. Most Web sites what you do is a front-end/back-end architecture, and you just do state management where the front-end is state-less and you have some criteria for responsiveness, but a lot of these other Web things, particularly what we have to do to do Web analysis, to build the algorithms that actually drive the search engine, these are state of the art scientific computing problems, and so we’ve created the general fabric that will let people host those things as well, which is a big breakthrough in how you think about how computing gets done.

Down on the clients Silverlight is a big deal for us, and MIX will be a huge milestone in terms of the feedback we’ve got, and the new things we’re doing there, that we’re pretty excited about that and making that both richer and broadening the set of devices that that connects to, and ourselves having our key applications take advantage of that in a pretty rich way.

And there are some high level services. Still today you have too many passwords, too many accounts. Any preferences you state just kind of stay in one place. Moving things cross-Web is still way too difficult. So, we have some new ideas that we can expose about that that will hopefully move the Web — help it move at the incredible pace that it’s been going along.

You know, MIX has become a great event for us, and I’m certainly in a lot of meetings where we’re developing neat new stuff, and people say, well, will it be done by MIX or why don’t you hold that till MIX. So, along with the PDC, which is more just the basic platform type things, I’d say MIX is our most important event for what we’re doing in terms of developer technology.

With that, I’m sure there’s lots of interesting areas I didn’t touch on, but let me go ahead and open it up for whatever people are curious about.

At this point, the questions began. I’ll probably spread out the questions and answers over the next few days rather than just dump it all on the blog at once.