I’m attending WWW2008 in Beijing this week. It’s turned into a big of a monster conference… nine simultaneous tracks over three days, not to mention a day of workshops and tutorials! Yow! And I’m seeing a number of colleagues from the usual haunts here as well. Both Kai-Fu Lee, head of Google China, and Harry Shum, head of Microsoft’s Live Search development, each gave keynotes, and I thought the themes on them was quite interesting and contrasting.

Kai-Fu Lee’s theme was Cloud Computing, or moving to a world where data and computation was handled on remote anonymous servers and applications then ran. He gave an overview of a number of Google applications that ran on this - Search, Mail, etc. I was struck by one comment he made, which is that cloud computing frees people from the monopoly of a single company controlling everything. Except, of course, the company that runs everything in the cloud for you…. Meet the New Boss…. Same as the Old Boss! But digs at Microsoft aside, the path outlined was clearly focused on Web applications built out on cloud computing, with those applications all leveraging large scale, reliability, and naturally massive amounts of data to handle things.

Harry’s talk was more of a Company Meeting talk, in which he handed the microphone to Graham Sheldon to show off some demos, in particular highlighting some of the cool things MSRA is doing as well as some of the latest on the Live Search release. They led off with what I thought was the best, which is some work from MSRA’s speech group that extracts speech from video and then enables you to see related videos while watching them. It was put together well, so it isn’t so much a “watch while on the Web” demo but “imagine you’re watching TV” video. I’ll see if I can’t find a link, but good stuff. Also shown was Guanxi, which tries to do a people / relationship search… in this case, it showed who was related to Bill Gates. They also showed a demo where you could do query-by-image, which would show images related to a target image. I need to ask some of my former UW colleagues who did things like QBIC (Query By Image Content). The demos of released Live Search features were focused on new features in the News and Local Verticals, including some cool stuff from the Maps team (which continuously produces some great stuff). Oh, and they have a few things on health they’re experimenting with, and trying to get things hooked up with the HealthVault.

OK… so we have two “My company is doing cool stuff, come work for us!” keynotes. But do we have any insight here?

Yes. Google, as widely reported often and everywhere, is busy making an operating system platform of cloud computing that they then build their services on. They’re not actually selling or providing a cloud - Amazon is, with EC2 and S3. But they’re creating the applications that depend on the cloud.

Microsoft, on the other hand, isn’t really pushing the cloud platform. They have a number of components for that, but the demos shown are all slices on search. But they’re certainly not talking about the power of their platform; they’re talking about cool features. But I worry along that line. The problem they have, which they and Google are trying to address, is user flow. Users don’t go to a vertical, they go to search. So now the problem is to discover intent on when it’s appropriate to show essentially a house ad for a vertical with some content, and then create a compelling, and consistent, experience as a user moves from “search” into “news” or “health exploration” or whatever they’re doing.

What I can’t help but wonder is why neither appears to be really pursuing differentiated domains and brands. For example, I still don’t think of Google, Yahoo, nor Microsoft when I think “news.” I think CNN. And really, I don’t think “news search” so much, I want more of a news paper. Archival search is great, but should be from within the news portal. To that degree, I wonder why “Live News” isn’t more MSNBC, or even just a different URL, such as www.livenews.com (it’s some random news site… probably buyable!). Certainly there’s lots of direct visitation to www.youtube.com, and I’m still more familiar with www.mapquest.com than the URLs for Google, Yahoo, or Live maps.

Anyway, food for thought… as always, I’ll lie about updating this later as the conference progresses.

Update 4/25: We (a number of anonymous conference delegates, and yours truly) now have short synopses on all the keynotes. In order:

• Kai-Fu Lee, Google: Use our stuff!
• Harry Shum, Microsoft: We have stuff!
• Sir Tim Berners-Lee, W3C: I invented stuff!
• Robin Li, Baidu: I paid for this stuff!
• David Belanger, AT&T Labs: We route stuff!

In fairness, we’re sort of making up Robin Li’s synopsis. Sir Tim’s keynote was somewhat, uh, long and rambly, and after about 30 minutes of it the audience in the Great Hall of the People got restless and started heading to the drink counters for more beer and wine. Sadly, by the time Robin got to the stage, the audience was in no mood to listen and was already engaged in conversation, so we’re not really sure what he said. But Baidu did sponsor the banquet, which rocked, so we thanked him for that.

David Belanger’s keynote was the best in my opinion… and not just because he didn’t do either a passive-aggressive product placement speech or an aggressive-aggressive product demo speech. He just talked about content, experience and devices, and networking to them and a lot of the challenges. For example, apparently as of 10 years ago when AT&T licensed out its rotary phone service, that was still upwards of a BILLION dollar business. For rotary phones. When a new touch-tone costs $10, or is often free. The main takeaways were that (a) there are loads of devices and enpoints, and it’s all increasing, and (b) the observation and re-iteration that old devices don’t go away slowly. The last is ignored at people’s peril… people hold on to things a lot longer than nearly everyone else would like.