[TextRunner]. Basically Google or other search engines create a big index of the Web. So you type in “Rachel Tompa” and you get all the pages that contain those words. That’s pretty much what it’s been doing for the last four years, with some updates. What if I could take those pages and look at a sentence on each page and extract the basic facts from those sentences? If it says, “Rachel Tompa lives in Seattle,” I could say OK, there’s “Rachel Tompa” and there’s “Seattle,” and there’s a relationship between those words, and the relationship is “lives in.” And then I could ask, who are the people who live in Seattle? And I could get that information not just based on strings, but based on the actual relationships. So it’s more like a database than just an index. That’s a very compelling notion. We’re starting to do research on that, and so are some other companies, like [Seattle-based] Evri.
X: What are the business applications of information extraction?
OE: Let’s say you were interested in making a profile of someone. Say you’re a salesperson and you’re calling several contacts. You want to have key information at your fingertips. You say, Oren Etzioni, who is he? You can go to my homepage, but that just tells you what I want you to see. You can browse through a few Google results, but that might only give you a few news stories, and it would get pretty tedious. Instead, you could have a profile that would fill in the basic facts about a person and it would give you a much more comprehensive view based on hundreds of pages.
Or, what if I want to know what biotech companies are hiring in Seattle? Right now it’s a pretty tedious process, or I could go to Monster but that’s incomplete. Instead, we could have a system that went to all the companies’ job pages and checked a) are they located in Seattle, b) are they hiring, and c) what kind of job? And did that automatically. So it’s very practical.
X: What does it take to make a successful startup, and is it different in today’s economy versus a few years ago?
OE: The elements of a successful startup are still, first and foremost, great people. You need great people on the business side and great people on the technical side. What is challenging now in this economy is that some of the very grand ideas, the “if you build it they will come” ideas, are now less popular. But people are continuing to invest in good ideas and good teams.
One thing that’s changed for the positive is the cost of putting something together. Because of open source, because of Amazon’s EC2, those things have made it a lot easier to put together a simple website or a simple prototype of your idea for less money. And that’s a good thing. I think people do want to see revenues sooner than they did in the past, but still, if you’ve got a great idea and a great team, it will get funded.
X: Are investors behaving differently now?
OE: I would say that they’re just paying extra attention to the cash needs of the company, when will it start making money, when will it hit the milestones. There’s a bit less slack in the system. But good investors, particularly those that invest in early stage companies, know that by the time the company comes to a potential exit four or five years from now, things will look quite different. They want to make sure they hit their milestones for subsequent financing rounds. So it’s not that different, I would say.