Chris Anderson makes some incredibly dubious claims in that article. For example: "In short, the more we learn about biology, the further we find ourselves from a model that can explain it." for example...we only need to look at things like Mitchell's Chemostatic Hypothesis
for how important and I would say, more central modeling has become for things like understanding mutations and disease (in fact modeling using a program called Rosetta has made great strides in the past couple of years in the search for a cure for influenza...I'll try to dig up the article). The more we learn about biology, the more detailed and elaborate our models have become...pretty much like all other sciences of physically real things...we learn more and more about causes and effects
I, frankly, don't care if our machines learn all the correlations.
need to learn, and when we do, we call that knowledge a model (ore if we are fancy "theory")...even if it is a cartoonish, qualitatitive, and poorly described one. Modeling has been around since we could perceive--it is part of who we are...perhaps it has only been a disciplined process since science began, but unless we are gone, models are going to be around.