I think that if you filter-out the exaggerations in the article, the general message is still worth to reflect on. Some process doesn't get more efficient just by virtue of using a computer, and many tools we have now are more about good-looking interfaces than actual productivity.
Not only that, but a lot of the efficiency gains are directed towards tasks that, otherwise, would not be done because it wasn’t worth the cost. Now that the cost is low enough, their marginal gains absorb the resources that’s be freed by automation.
