This post originated from an RSS feed registered with Java Buzz by Elliotte Rusty Harold. Original Post: Google for Christmas Feed Title: The Cafes Feed URL: http://cafe.elharo.com/feed/atom/? Feed Description: Longer than a blog; shorter than a book Latest Java Buzz Posts Latest Java Buzz Posts by Elliotte Rusty Harold Latest Posts From The Cafes

This morning I’m sitting at the computer pruning my address book for Christmas cards. (Yes. I’m running late.) As usual I discover a few incomplete addresses. Missing zip codes are especially common. I know there’s a zip code lookup finder on the USPS web site somewhere, but rather than Googling for it on a whim I paste an address into the Firefox location bar.

Bam! Up pops the complete address including zip code! Five or six more and I’m done. Easiest data cleaning I’ve ever done, and maybe a few more people can get their cards before New Years.

Search is continuing to improve. We’re gradually moving beyond searching for pages and moving into the realm of searching for answers. These days there are three sites I go to, depending on what I want to find:

Wikipedia

For basic information on any major or minor topic

Amazon

For anything I expect to find in a book, as well as many products I want to buy

Google

Maps, pictures, web pages, and products

However even when I’m searching for web pages, it’s shocking how often Google’s little snippet of the page gives me the one piece of information I need from the page. I don’t even have to go to the page they’re linking to.

I don’t think computers have ever been this good at answering questions before. It still seems to be improving. I’m not sure how much further this can go. Perhaps the artificial intelligence we actually need is just a really good search engine with a lot of indexed data to churn through and the right ranking algorithm.

Of course this is still just half of AI. Google, Amazon, and Wikipedia all work by indexing large amounts of data Only some of the map data (and not even the most useful part of that) was created mechanically by machines without human intervention. Everything else these sites work with was created or discovered by thinking, breathing humans. Even if a computer can answer any question we pose of it, that doesn’t mean it can gather new knowledge or decide what questions to ask. Perhaps that next step will require computers that can actually think.