A Call to Rethink Internet Search - NY Times

A Call to Rethink Internet Search

By STEVE LOHR

“We could soon view today’s keyword searching with the same nostalgia and amusement reserved for bygone technologies such as electric typewriters and vinyl records.”

So declares Oren Etzioni, a computer scientist at the University of Washington, in an essay published Thursday in the science journal Nature. (Available online to subscribers or for a single copy purchase of $32.)

The missing ingredients, he writes, are mainly the necessary investments in money and science by leading technology companies and universities. The better world of search, according to Mr. Etzioni, will be services that field spoken or typed questions and generate useful answers. Or, as he writes, “natural-language searching and answering, rather than providing the electronic equivalent of the index at the back of a reference book.”

Many people have lamented the shortcomings of Internet search, but Mr. Etzioni’s critique is provocative and informed by his own research, and he describes the way ahead and the technologies needed to get there.

One threat to progress, Mr. Etzioni writes, is the keyword search box, an innovation-retarding trap that “exerts a powerful gravitational pull.”

That observation echoes one heard for years in another field of computer research — user-interface experts who have long bemoaned the “tyranny of the desktop metaphor” (document and file icons strewn across a “desktop” screen) as locking in the past in personal computing.

The impetus for change, Mr. Etzioni said in an interview, comes from two sources. On one side, he says, is the rise of the smartphone, which people increasingly use to search for and find information. “The device of the future is the smartphone, and the 10 blue links of traditional search don’t cut it anymore,” he said.

On the other side, Mr. Etzioni said, is the rapid progress in software that can essentially read text and infer its meaning.

The big companies engaged in search — Google and Microsoft — have taken steps toward delivering smarter search, Mr. Etzioni concedes. But, he adds, “most of what Google and Microsoft have done so far is to take traditional search and play with it around the edges.”

I.B.M.’s “Jeopardy!”-playing Watson computer, he said, is a signpost of what is possible. It was an impressive question-answering machine, though it is uncertain how far it could go beyond playing “Jeopardy!” And, by the way, there were three players, including Watson, in the “Jeopardy!” games, while Google fields billions of queries a day.

“I’m not claiming that Google should turn on a dime,” Mr. Etzioni said. “My pitch is that we should put more energy into this research, that we’re under-invested in this technology of the future.”

Mr. Etzioni’s pitch, it should be noted, is not disinterested. He is an artificial intelligence researcher working on a key technology needed for the kind of smart search he describes — software to read and decipher the meaning of phrases in Web pages. Other universities and corporate labs are also pursuing that research.

Supporters of the University of Washington research team include the National Science Foundation, the Defense Advanced Research Projects Agency, the Office of Naval Research and, yes, Google.