Whoever wrote that piece did not fully get the Song, Price, Guvenen, Bloom, von Wachter (SPGBW) paper. The whole point of the Stiglitz "imperfect information" papers in the seventies was to show how usual conceptions of a market fall apart when you add information asymmetries. The whole point was that models of imperfect information made better sense of the world than the stuff you saw in undergraduate textbooks.
It is total revisionism to say something like
Since the whole point was that separating equilibria were the more natural state of affairs (the more intuitive example being job signalling by investing in education). You would expect a libertarian to call out the unions for compressing the wage distribution, but I guess we're all too young for that now.
The theory of sorting in the labour market is not new: it's been around since Becker in the 60s. But it is also subtler than just saying "it's all education's fault," which has been around since the productivity slump in the 70s. It certainly isn't totally explained by "the end of asymmetric information" - if that were the case, why do smart undergrads spend so many hours trying to beat Google's interview questions?
What the new paper's results do show is that firms, particularly firm technologies, matter. It's no longer a story of "if you're smart you can make it anywhere" - every job you get is a red circle on your ability and potential, or provides you with necessary social capital to succeed.
If you were a real radical about this, you would think the ideal policy is to literally shift funding away from higher education, and use it to subsidize the most profitable firms developing their own curriculum teaching what their business needs. Worst case, a student comes out of the curriculum and wants to be a startup entrepreneur.
That is why I am so militant about the quality of secondary schools in poor neighborhoods
Why secondary school? Why not primary school? James Heckman has been
arguing the former for decades.