This article was fascinating.

I was just talking to a couple of software engineer friends the other day about how engineering research like this doesn't really happen anymore outside of the massive companies, and even within those it's greatly reduced.

Now it's all about applied engineering (app development using established technologies and techniques), with research limited to incremental gains with new technologies, augmented by published research. But it wasn't always like this; there was a gradual erosion. Just prior to this latest era, a company could at least plausibly start a project to use published research with no public implementation and build an implementation. Our careers started in the 2000s and we remember a better time...

Two of us work in a large company currently and were recently closely involved with some of the most "speculative" research at the company, and it was almost entirely incremental. The third person is a literal research engineer at an engineering research firm who says real research described in articles like these is dead.

I can't imagine having two years to produce something so ex nihilo these days, and the fact that they were able to achieve so much in such a short amount of time is truly incredible, and a testament to the quality of the engineers.