I'm trying to widen the viewpoint here to set the activities and products of science in a larger, more dynamic context--hence, the comparison to a golfer working to master putting on a particular green.

The phrase "difference between fact and theory" is another way of saying that our actions do not or will not always work out. There are any of a number of reasons for why we fail. It is not always a simple case of us being in possession of a literal model that happens to not correspond to reality. The model/reality comparison is mostly a handy metaphor we use to illustrate our failures and potential for improvement.

Getting back to the OP a bit, the Wired article suggests that knowledge can be gained without discovering "underlying mechanism" and the like. Others counter that a real explanation and understanding requires an underlying mechanism. I disagree. I think that from the get-go, we learn to take advantage of the behavior we observe. We learn to drive a car by learning how the steering wheel, gas pedal, brake, etc. control the vehicle. Mastering such visible, on-the-surface behaviors of the car is genuine knowledge. When we do get around to opening the hood to observe the "underlying mechanism," we are not learning in a new way as if there was something special or closer-to-reality about the engine. We are observing and taking advantage of aspects of the car that we haven't attended to before.

I'm cautioning against this notion that the hidden or the underlying has a higher (or holier) epistemological status than the initially observable. This, I think, is easier to see when you set aside the notion that Newton produced a model and instead appreciate that he showed us a new way of dealing with the world. You make a few observations of a comet. You wait six months. You perform some calculations. You position your telescope according to the calculations. You look in the eyepiece and see the comet. That sort of mastery of the environment is what Newton produced. It is very real. True, it is not always perfect. Maybe you don't do as well with the planet Mercury. There is room for improvement in some circumstances.

As for your black hole model, yes, we say that science produces models and we will continue to say that it does. But strictly speaking, and such speaking matters only in philosophical discussions like these, science has nothing in its hands that behaves like a black hole (but is not a black hole--only a model of one). It has recipes, techniques, and procedures for making observations, talking about them, making use of them, and so on.

So, I don't think the Wired article is showing us a new way of doing science. I think science has always been done that way. Again, Newtonian and quantum mechanics succeed, yet they provide no underlying mechanism. That will always be the case because explanations have to stop somewhere.