Most of the time since mid 2012 (me for about a year, me and a few other for two for two more) has gone into writing Arvo, a purely functional operating system. Arvo is not enough like anything else to be seen as anything else but research.
This is a fairly typical timeline for CS research. What's unusual is just the depth of the stack. I would also point out that if you look at the normal cost of developing any sort of operating system, even to the alpha level, the metrics are pretty good.
Where the project looks really unproductive is in objective measures of content production. For instance, Nock took me roughly from 2002 to 2008, which is something like a bit and a half of output per day.
This obviously makes for slower going, especially near the beginning.
My address is my first name, followed by the first initial of my last name, then the letter Q.
That confuses the hell out of me, because I think your plan is brilliant. What gives, broseph?
My only formal CS-education was at summer camp learning Logo when I was nine. In my early 20s, I taught myself LISP from PG's book purely so I could start wasting a few weekends a year coding an urGAI. The first practical coding I've ever done was this year, when I finally bit the bullet and learned Python and Java.
Is your vision convincing to me because I'm clueless about the harsh realities of programming useful things? Or am I able to discern its splendor because I avoided exposure to the false conventional wisdom that blinds other coders to your insights?
If you put a gun to my head, I'd have to pick door number one. Actually, if you put a gun to my head I'd say Urbit was the best thing since Filmer and then apologize for that over-the-top pastiche. But you see my dilemma.
[Edit: Serious question. You came up with Watt/Hoon 5+ years ago, and you claim it's not much harder to use than Lisp. If that's true, where is all the useful Hoon-coded software?
If I gave you five years to write things in Lisp, you could produce an incredible library of offerings... No?]