I don't think the fact that others have tried with limited success is an indictment of prismatic at all. If anything, that strengthens that argument that this project required a lot of NLP skill.
I do find the whole field boring. Making a news aggregator that's X% better than existing ones doesn't improve the world a whole lot, or even offer a compelling consumer value proposition.
I don't want to sound like I'm against all consumer entertainment tech. It's clearly something that people enjoy, and it's exciting for some people to make. But this particular application is one whose value I am very skeptical of.
But Prismatic has told me about articles that make me better at what I do. Depending how widespread that experience is, they may be having a larger impact on our 10 most pressing needs than most teams that attack those needs directly.