What he described in the beginning of that essay is how Microsoft Research worked just 15 years ago. Maybe it still does, I wouldn't know. Yes, you are judged b

What he described in the beginning of that essay is how Microsoft Research worked just 15 years ago. Maybe it still does, I wouldn't know. Yes, you are judged by your results. But beyond that and beyond some flimsy constraints dictated by what "team" you were on, there were no limitations whatsoever. This was the most productive and enjoyable time of my entire 25+ year career, and a staggering contrast to how the rest of Microsoft works. And the people were by far the smartest I have ever met, and this was not the only lab I worked at. Most of them weren't genius level (although they were still unusually smart), but some very clearly were, to the point where I'd sit in some meeting and lecture and think "what the fuck am I even doing here". Too bad most of the stuff they do never ends up in products - the rest of Microsoft can't tell a gradient from a hole in the ground, with very few exceptions (basically just Bing and some parts of Ads).