[hyperpallium2]

The opposite of "Deep Work". Great for getting stuff done; but a poor substitute for understanding. Good for practitioners not researchers.

Because a working but incorrect mental model is a terrifyingly insidious, almost impenetrable barrier to actual understanding.

[ineedasername]

Yes, this is, at best, a recipe for understanding how to properly use a complex tool (say LTSM networks) but by no means a mechanism by which you would gain actual understanding. And if you mistake the two, you run into the Dunning-Kruger effect, believe you know more than you do, tweak things in ways you don't completely understand, and end up with utter crap that appears to be gold.