Groove's revenue exceeded expenditures, though, right? Or could have? I'm not sure it was reasonable to expect success at that scale for an entirely novel class of application in an era where startup investments were still expected to 'mature' only after about a decade or more. Groove was only ~8 years old, and 1.0 had been released ~6 years prior.

> But the enterprise and consumers chose centralization as the winning architecture, and the rest is history.

As I recall, around 2.x you were more-or-less switching from a distributed architecture to a decentralized one, but... your point is taken. I suppose I should admire that you saw the writing on the wall as soon as you did.

The wheel keeps turning though, and I wonder if another stab at the idea could succeed today. Certainly the various risks of centralization are much more apparent, and to a broader audience, now.