Kevin Rose is the horse of choice for the Riders of the Startupocalypse proclaiming the end is near.
I'd say there is an argument to be made in both directions. Clearly there are a lot of startups that didn't arise from market analysis, but there's also a lot to be said for not panhandling for VC money without understanding your addressable market.
I'd rather not see those arguments unintentionally strawmanned by somebody who is apparently unaware of the current state of Facebook's stock or what exactly became of Digg and why. Or how Rose is widely reviled for throwing his engineers under the bus for what were ultimately product decisions made high up.
The author outlines a lower risk approach to customer development that I think works better for most people, unless your goal is to be the next Zuck at all costs.
"Research the market" approach is good if you want to build a regular bootstrapped business. If you're building a startup, you should actually build an MVP and "Test the market".
Please don't get me wrong, but I still believe that building a couple of duds and failing is the better way to go. If you're building something truly innovative and disruptive, most people you ask during "research" won't even know that they need it.