Could IBM's Watson Put Google In Jeopardy?

theodp writes "Over at Wired, Vashant Dhar poses a provocative question: What If IBM's Watson Dethroned the King of Search? 'If IBM did search,' Dhar writes, 'Watson would do much better than Google on the tough problems and they could still resort to a simple PageRank-like algorithm as a last resort. Which means there would be no reason for anyone to start their searches on Google. All the search traffic that makes Google seemingly invincible now could begin to shrink over time.' Mixing supercomputers with a scalable architecture of massive amounts of simple processors and storage, Dhar surmises, would provide a formidable combination of a machine that can remember, know, and think. And because the costs of switching from Google search would not be prohibitive for most, the company is much more vulnerable to disruption. 'The only question,' Dhar concludes, 'is whether it [IBM] wants to try and dethrone Google from its perch. That's one answer Watson can't provide.'"

Comments

Re:Better searches no good if they're too slow

[Trepidity]

It's also a completely different problem from information retrieval in a messy domain like "all documents on the internet". Watson is built mainly out of more structured data: dictionaries, almanacs, atlases, Wikipedia infoboxes, etc. It turns this into a huge database of knowledge, and then does inference on that database to try to answer Jeopardy-style questions posed in natural language. But this doesn't even try to tackle the other side of the natural language problem, which is parsing not only a natural-language query, but the entire contents of the internet.

In short, Watson might compete in the Wolfram Alpha space, of retrieving structured knowledge from databases, but not, at least not without a major overhaul, in the general document search space.

Sick of 'smart' searches

[RandomUsername99]

I'm getting a bit sick of 'smart' searches... or, rather, not being able to disable the 'smartness.' More often than not, I really don't want a search engine making assumptions about what I meant, rather than just taking what I enter completely literally, and I *never* want it to insert results that don't contain all of my search terms because it scored exponentially better with the other items in the query. Chances are, I added in the term they were ignoring, specifically, to drastically reduce the number of results I got, because I wanted to *narrow it down*.

Maybe I'm a curmudgeon, but I would rather tweak the search to narrow down crap results than try to outsmart the 'smartness' any day of the week. I understand that this isn't necessarily what John Q. Internetuser is looking for in search, but at least having the option there would be a big help. Google used to have a very straightforward syntax to help you modify your search results in specific, predictable ways... while much of that syntax is still valid in google searches, now it seems like everything can be arbitrarily overridden by what google thinks you 'should' have meant, rather than what you told it you meant. Very frustrating.

Re:Google, really?

[davester666]

Of course Watson will be able to do this.

Obviously, Dhar did a bunch of research, and determined that even though it took a massively powerful computer to answer one question at a time that has a predetermined single answer, over several seconds, it can trivially scale to support millions of simultaneous queries, which may have zero, one or multiple answers.