[willwagner]

A few years ago, I worked at a startup peripherally related to the search space and we did some usability studies on the google search page. What we found then, more or less, is that people generally feel like there is no problem with google's search results, and if you pressed them to try to find something that google could not find easily, instead of thinking google let them down, they essentially blamed themselves for not being able to find it. It's hard to sell a solution to a problem people don't think they have.

With that kind of thinking, it's really hard to build a product to go head to head with google that is so much better that can change people's habits directly. From my perspective, you either need to somehow change the equation for some sort of query attribute (e.g. google may not be as timely as twitter so I'll switch to it when something like an earthquake happens) and eat slowly away at the google advantage, or you need to build a product that first targets the thought leaders (something like building a better code search or social search) and then make google un-cool to use.

Either way, it's a tough row to hoe.

[wololo]

And if you are successful enough: being copied by a company with orders of magnitude more engineers, money, etc.

So I would bet that differentiating on any "pure technology" query attribute is out. Otoh there is so much potential in twitter and Ark and friends.