Esther had a great interview with Sergey Brin, co-founder and President of technology for Google. First I should say I was generally very impressed with Sergey -- his demeanor, energy and clarity of thinking was refreshing. Details on the interview follow.
Esther wants to know whether Google plans to leverage user identity (e.g. search history) to do more personalized search results.
He points out that philosophically it cuts against the grain of Google -- a search on a given string should always return the same results for any user. He gives the example of performing a search and sending the search URL via email. If it was profile/history driven, this would break. There are some benefits that could come from this, but he thinks the user interface and user education challanges would be substantial in making such a change in how Google works (or how users expect it to work).
Esther wants to know if they're considering other verticals for their search technology, for example job search?
He says they'll go after any opportunity where it's clear that their search technology can ad value. The best example of this is clearly their 'Ad Words' technology, which to some degree has radically changed the quality and model for inline advertising. They've also done other things, like acquiring Google Groups, which is a small vertical, and Google News, which leverages their knowledge of news content and sources all over the world. It's quite amazing the quality and richness that comes from Google News, and certainly exemplifies the power of their platform.
He says they're experimenting with 'Contextual Advertising' for web pages themselves, where any webpage, on other websites can leverage the google search and 'Ad Words' engine to to much more targetted contextual advertising based on what content a user is browsing.
What will happen with Blogger?
As we've heard, he says there is no master plan. They're getting tons of feedback on ways to leverage Google into Blogger, and the team will probably experiment with a few of these. The major focus now is getting Blogger into their infrastructure, including their ad infrastructure, which can really improve both the user experience of ads in Blogger as well as the contextual linking of blog content to ad content.
When or will you go IPO?
I was really impressed with his answer on this. He understands the incredible distraction that an IPO can be both for the management team and employees. He says it comes down to the fact that he's just too lazy (which is clearly not the case!). The essence is a moral response --- public markets are too short-term focused and he wants his team focused on the long-run. Whatever the case, he admits that ultimately their employees and investors want liquidity and that will have to happen through an IPO or acquisition of Google.
A person in the audience asked a very interesting question -- idea -- about Google playing a leading role in enabling the semantic web. The basic idea was that Google should process pages and classify them into ontologies, and that they can derive lots of linked ontologies from all the content they already index.
He basically said he doesn't believe in the semantic web as a set of linked RDF data-structures. His basic argument is that the structure of natural language and what it presents is much much richer than meta-data tagging schemes. Clearly, Google's understanding of natural language is unique, but there still is a need for machine readable APIs for data on the Internet.
Will Google provide a desktop utility -- e.g. index/search my drive?
He says no, it's not a core competency and no one has ever made it work that well.
What's the role of advertising?
He says if it has utility, users will like it. That's the essence of 'Ad Words', which really provide utility to both the consumer (highly relevant commercial content) and merchants (hits with conumsers on topics they're clearly interested in).
Another great question: Google search clearly embeds values -- e.g. Page Rank is a normative judgement on the quality of a web page. Do they design their algorithms with values in mind?
He says they don't design them, but offer discover them once they've been tested, and that this does play a role in choosing what they launch into their engine.
