I think you're certainly right in part.  I had in mind when I don't know what I'm searching for, and the results come back heavily top-loaded with sites which aren't the real thing -- advertising sites, sites which are using the name of what I'm looking for.  I'm sure there's a better way of putting it.  Google tends to have the valid result riding high.

I grant you, Google will get you to where you want to be, if you already know where that is.

So, in some sense, you are really doing the "searching" itself.

You know what you are looking for.

You create the search logic ("this AND that BUT NOT that other thing")

You then use your eyeballs to scan the top search results, weed out the missed syntax, homophones and so on.

And then "Google" lets you click on the result you need.

So this is the equivalent of going to file cabinet and walking to where you think the stuff you want is, opening a drawer, rifling through the manilla folders and then going, oh, yeah, this one.

You are the inference engine.

And by the way...I don't object to that...in fact, obviously...it's quite successful!

Jabailo, I also find Google's ranking of search results to be satisfactory, especially compared with Bing or Roadrunner.  If I search for something specific, Google doesn't usually give me a lot of garbage at the top of my search results.

mtechie, you're right.  It's irresistible to talk and speculate, but we need to see how it works in practice.  And the examples Facebook have been giving aren't very enlightening.

I think it will mean more, at least initially, to those sites that already use friends' comments and feedback as the basis for their basis, sites like Yelp and Angie's List and Foursquare. Just read something in the Washington Post that supports my opinion, for whatever that's worth!

One of my pet projects way in the past was CBR, or case based reasoning.  That is a combination of a logic trees and text search, but it changes the basic search from a set of logically connected terms and replaces it with a pool of questions that focus and direct the search.

The thing about Google that is hard to beat is the simplicity.  You put in a term, you get 10 million answers.  It's a bit like going to a slot machine and getting a jackpot every time.  However, people don't realize that the payoff contains a lot of wooden nickles.

Google's "expert knowledge" is really also crowdsourcing.  They take the search results and track the links clicked on for hits.  They also look at how many people link from web pages to articles, making them more relevent.  So their "expertise" is built into usage.  As I think you allude to, that's a dangerous kind of self-validating snowball where error can be compounded upon error!

Yes, knowledgebased systems are based on expert knowledge, even onces built with CBR.  The problem there is they are domain specific.  So these systems are good at predicting oil deposits, but probably bad at telling you the best gift to buy for your uncle.

Yet, things have been changing...certainly the large scale availabilty of data warehouses, the exponentially growing social media content, are all in some sense expertise when properly managed and sifted.

But unfortunately there is that bit of the elitist or should I say aristocratic bent in me (and by aristo- I mean rule by the best, not by birth) and that some people, who take the time to develop expertise can be faster, more precise and more relevent than a flash mob.

Directed search, augmented by real people, sounds a bit more like Google than Facebook right now.

I'm afraid my prejudice in favor of experts, and against crowds, derives especially from my experiences with Wikipedia.  Experts can be glib, but I think most questions don't really require lateral thinking.

Actually many AI techniques ultimately gave way to "swarm" methodologies, where rather than having a single monolithic logic tree, many independent agents go off to attack a problem.

That is, in fact, how Watson is designed.  At any point, several..many...threads spin off to tackle the problem, or answer the question, using different techniques.   First one "wins".

So, if correctly done, the collective intelligence of many people added into search can guide people to the right answer.  Well, yes, that's how its supposed to work, but as you allude to, crowds can be dumb.   100 people with the wrong assumption will be no better than 1 person who has really studied the problem.  

On the other hand, solitary experts can make big mistakes as well because they tend to focus too fast before thinking laterally of all the possibilities (sorry for sounding philisophical, I'm in the process of reading the classic Zen and The Art of Motorcycle Maintenance).

Still, I really like the idea of directed search augmented by real people, even if at first it amounts to bunch of Facebook Likes!   Eventually people can justify their opinions and become SMEs in topics like Best Taco Bar.   However, I still believe in the swarm design, so long as its collective intelligence, not collective stupidity!

Quite right, mtechie.  You're really only searching the tip of the social graph which shows above the water.  The iceberg of knowledge which lays beneath...oh, it's too early for this metaphor!