Now any analysis applied to a data set is now "Big Data." While easy to just resolve it to being how the media treats most science stories, I had hard time understanding the demarcation. I now go by DevOps Borat's definition - "any thing which crash Excel" [1]
It's worse than that. People will run Hadoop on 20 nodes to do stuff Excel could handle easily, so they can join the "big data" fun. It's just CV-padding really, ironic given the discussion.
I has someone contact me about a "big data" start-up recently, and did I know anyone that would be interested.
I replied saying "I may be" and what size was their "big data", as usually I don't see the point in taking it out of a traditional relational database in most cases. I don't think they like my attitude, as I never got a reply.
[1] https://twitter.com/DEVOPS_BORAT/status/288698056470315008