npwsol - 6:58 pm on Sep 19, 2008 (gmt 0)

Why are people so skeptical of Google's technological ability?

Touching on Whitenight's rant specifically: What you're failing to recognize is the foundation and architecture Google has been amassing since its inception. No one but Google knows for sure, but I've heard rumors of over a half million servers throughout the world, and that number, I imagine, is only growing day after day.

So, what does this have to do anything? These half-million servers, apart from managing Google's website, are responsible for crawling all over this great web of information, categorizing everything it sees. From words to images to flash animations, anything can be read by Google's bot.

Words are easy to categorize, because they're already in a representative form; in PHP, words are the key in your associative array. What google is missing from the equation is the value, and that's where I think we're going to see some massive shifts in the coming years. If what this post implies is true, Google will begin to attempt pattern recognition in images and movies.

While I don't know how exactly this would work for them, I have my suspicions. Google could begin to look for consistent or similar shapes in all the image results for "apple" and eventually come to the conclusion of what an apple looks like. The result? Instead of basing image search results around the alt text, filename, or surrounding content, Google would have the ability to search its image index for the shape of an apple, leading to much more accurate results.

The same can be applied to language. Presently Google's results are based on a number of factors but are based largely on Keywords (links get a lot of weight, but the words are what's important here). There is no meaning associated with the keywords, they are simply words that are common throughout the document. But suppose Google starts being able to recognize the pattern of a sentence. Subject Verb Object. Subject Object Verb. Languages follow these patterns to a T; it's the fluff and exceptions (Sentences like "Sit down." have implied subjects) that's throwing Google off now.

With a large catalog of objects you can recognize by name, though (thanks to the image pattern recognition I was referring to), you now have two key parts of a sentence. Throw in YouTube and you can start to get a bigger piece of the puzzle: actions. If Google has the ability to track a pattern over a series of images instead of just an individual image, Google may begin to recognize actions (verbs).

In mathematics, you're given known quantities and unknown quantities (variables). Algebra is the art of determining the value of an unknown quantity based on the known quanities surrounding it. In my scenario, the known quantities at this point are the subject, the (direct) object, and the verb. The unknown quantities are abstracts, articles, metaphors, and probably some other things I haven't thought of. Those would be tough things to break down...

Thing is, though, Google has that thing it likes to read called the internets, and that has a lot of information, mostly in the form of words and other language constructs, and those unknown quantities are being used all over the place. With most of the parts of a sentence already being defined, it would only be a matter of time before Google understood the context of those as well.

Regardless of whether or not this system is sentient (I think it ultimately will be; Singularity here we come!), you now have a system which not only has a ton of information, and not only a system which can cycle through that information a million times in a second, but a system which is analyzing that information in the same way a human would: based on context and a pre-existing knowledge of what it represents (concretely at first, abstractly later).

This may not be the monetized success that you think it might, but then, none of Google's products really are. What it would be is a technological innovation, one which could give a severe bump to artificial intelligence as we know it (and, if the Singularity holds true, another severe bump after that, followed by another, and another, and another...).

Maybe that's part of the appeal of Google, though; they hand out their technologies without concern for the potential monetary gain. They continue to improve their search algorithms even though I get what I'm looking for most of the time. Sure, they work to appease their shareholders, and maybe low adsense payouts and high adwords prices are a result of that. Maybe, just maybe, Google is pushing webmasters around a bit when they decide your site isn't good enough, so you have to pay more to advertise.

For me, though, this is a good thing. The internet has become far too cluttered with people who are trying to take the easy road into your pockets. Too many sites are the same. When you and your competitors offer the same service with a different layout, what should drive me to one or the other? The only thing left to compete over is your position in the paid advertisements; who knows, get enough competition in there and you could be paying $20 for that keyword... Could this be a means of battling inflation?

I digress...

On another theme in this thread, claiming Google has only one successful product is naive. GMail has been a success; it may not be terribly innovative (well, their spam recognition is well above the fold; I've never had a false positive from Google, nor has any spam ever gotten through to my inbox), but it has brought people even closer to Google. Blogger may not be wildly successful, but it carries the Google brand.

Google is in the dictionary. Google is a household name. Millions of people use Google, and they don't just search the web; they do search, however, and when they want to search for something they have more than enough reminders about where they should go.

I'm of the thinking that Google measures its success primarily on customer retention, and what better way to retain customers than to give them everything they could want for free?

I hope I'm right; it would be so refreshing to see a company after my own heart. Success is not measured in dollar bills, it's measured in smiles, sunshine and rainbow farts.