Ask HN: Detecting popularity stress
There probably should exist an algorithm for detecting some "popularity stress"(tension), i.e. a discrepancy between actual popularity of something and the "should be" popularity. Think of Google Trends or Flu prediction, which is somewhat obvious.. but also think of the stock market - there probably exists an algorithm that can pick some signals about when some stock is going to become much-much more pricey (popular).
Well, Google Trends is obviously detecting post-factum popularity - if something is popular - it will find out that X queries today is much more than Y queries yesterday...
...but I'd like to talk about predictive algorithms.
How can we detect something (a website, a blog, a technology, programming language, scientific paper, direction in which humanity is going to be developing) that has the potential to be popular, but yet lacks the momentum , therefore is under "popularity stress" (maybe there's a better phrase for what I mean).
One signal that I can think of - is whether this "something" is going to make our life much easier, without requiring any/much effort/resources on our part. I.e. C is making our lives tough, but ooc makes it easier. But can we make computers find out what's going to make our life easier? Or maybe some crowd-sourcing effort about what makes people's life miserable and tough?
What differs something that is born dead from something that is going to be huge? (Aside from billions of $$$ in marketing?)
Do you have any ideas for detecting this kind of "discrepancy"? Algorithms? Signals?