Had a similar experience with another (also Quanta?) article last year about some optical physics breakthrough made at my not-very-well-known alma mater, but the article was written as if the work was all done at MIT, which one of the collaborators was affiliated with. You had to scroll to the second page to see the name of the institution that the PI was actually at.
Deep learning is a form of optimization. Optimization involves moving along a high-dimensional surface to find the lowest point. In principle this can be nearly impossible because the surface might be covered in dramatic peaks/valleys/saddles obscuring the route to the lowest point. Some simulations have implied that this is not what the surfaces corresponding to deep networks look like, and that they instead look like a big gentle slope down to the minimum, with only small bumps along the way.