This is exactly the point of the article. From the conclusion:

> Hopefully this has illustrated some points about using and abusing tools like Hadoop for data processing tasks that can better be accomplished on a single machine with simple shell commands and tools.