Live Blog: Qi Lu Speaking At Web 2.0
Microsoft’s Qi Lu announced two new partnerships today at the Web 2.0 Summit in San Francisco. Here’s a transcript of his chat with Tim O’Reilly. Qi Lu I was at Yahoo for 10 years, and I always told Jerry, after 10 years I’m going to try something new. I went to Microsoft because I wanted […]
Microsoft’s Qi Lu announced two new partnerships today at the Web 2.0 Summit in San Francisco. Here’s a transcript of his chat with Tim O’Reilly.
I was at Yahoo for 10 years, and I always told Jerry, after 10 years I’m going to try something new. I went to Microsoft because I wanted to have an impact.
Search is about computationally understanding user intent. Understanding their interests and needs. That’s “absolutely” our long-term goal – to “build a mind-reader.” Once you understand user intent, you can figure out different solutions. There are many queries where images are best at fulfilling intent. Sometimes videos are best at fulfilling intent.
Ahead, you have Facebook and Twitter – particularly Twitter. There’s a strong velocity of things flowing through that will enable people to get answers.
Yusuf Mehdi to do a demo now.
Bing launched about 100 days ago and off to a pretty good start.
People will start to ask more complicated questions. To answer that, you need better access to data. Starts demo of Bing Wave 2. Shows recent photo search tool. “It changes the way you do search.” Shows example of search for “top iPhone apps” — this has “really resonated with consumers.”
Second area is about real-time, hot information. Two announcements:
Partnership with Twitter. Been working on it a long time. We will get access to all public Twitter information in real-time. “We’re gonna do some really exciting things.”
Other partnership is with Facebook to get access to all of their public data, and that will come at a later date.
Now showing example of Twitter integration. Shows full, real-time feed. Looks like existing Twitter search, just wrapped in Bing interface. Will update automatically.
Bing applies “best Match” technology to tweets: de-duping, then focus on
quality – look at who’s tweeting and assign soicla relevance score
popularity – look at the caption – length of comment, links, etc., affect quality and relevance
usefulness – number of retweets affects results
Then we apply spam filtering.
Bing also shows hottest topics on Twitter – a tag cloud. Looking at tweets related to these items – Bing takes most popular links and assigns its relevance algorithm, then shows the hottest links, too.
Pick a topic like “Yankees” – Bing shows most recent tweets, then shows most popular links. Bing pulls out landing domain from shortened URLs.
This is beta code. Sean Suchter working on it at headquarters. Should be available shortly at bing.com/twitter.
Demo over. Qi Lu speaking again.
We don’t disclose financial terms with Twitter. “I don’t know specific length of agreement. This is a start.”
The key thing is to have great teams, great technology, and keep the pace of innovation.
Audience questions now.
Qi: In a real-time corpus like Twitter, there’s a lot of signal velocity. The Twitter corpus is still evolving. But based on what we see, you have enough signals to have meaningful content. A lot is meta content that refers to other content on the web. But this is just the beginning. The future is very compelling.
Are you keeping the Twitter stream? Are you archiving the firehose?
Qi Lu: Don’t want to answer because I may not be accurate.
The Twitter deal is non-exclusive.
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.
New on Search Engine Land