Bootstrapping Urbit from Ethereum

So identity, IoT, and literally everything in between.

Urbit has been around for a few years and has done nothing to clear up the confusion around it. I think at this point it is safe to say that any claims of Urbit being a revolutionary form of computing are hogwash.

This is probably similar to how Scientology got started. Nobody really believes in it, but are too afraid to strongly denounce it because they don't want to be the one who "doesn't get it." So they all just go along and get sucked further along down the rabbit hole. The difference between Scientology and Urbit is that Scientology managed to rope in a few celebrities and gained some sustaining mass. If Urbit wants to make it big, they should pay Zuckerberg and Musk to evangelize it, and then everyone will be falling all over themselves to be part of it.

Others, including myself, have noticed this too and it reminds us of something else a little closer to home: namely, what happens to Scientologists and the like after they've been in the cult a while. Even the Scientologists who later leave still think in terms of engrams and body thetans.

I think Curtis Yarvin is a master huckster and wordsmith, but near as I can tell what's good about Urbit isn't original and what's original isn't very good. The central concept is basically Java's "write once, run anywhere" concept from the 90s. Remember that? It wasn't just about applets; applets were only the beginning. With "mobile code" that could run on any device, Java was going to make existing operating systems obsolete and finally bury Microsoft.

Sound familiar?

Curtis adds sizzle to the steak by mixing in ideas Hackernews people will find appealing -- functional programming, blockchain, distributed computing -- and finally he plays the same semantic shellgame with his words he plays with his political essays, to convince you that the horrible political views he has aren't so horrible. Except this time it's to convince you that Urbit is more than a reimplementation of interesting ideas from the past, some of which failed massively.

Why? Well, the fact that he is selling "virtual real estate" in his new internet that's totally going to change everything real soon now might be a clue.

I'm one of those people that get it, so I'll try to explain it without going off the deep end and waxing poetically about planets or whatever.

Urbit is a platform for building decentralized apps. To that end, it's a tightly integrated set of different features that play into that: an identity system so that all the apps can refer to the same people by the same handle, a typed RPC network for easy message sending, and an append-only log of all events that platform handles being the most important parts.

Right now you can build decentralized apps like Mastodon, except they 1) take hours to setup a node, along with having to know arcane Linuz sysadmining 2) aren't actually decentralized, but federated. Urbit wants to make it easy to setup your own server, which runs as a node for all these decentralized apps (instant messaging, Twitter, etc.), along with be useful for server-y things like aggregate APIs (email, Facebook).

There's not actually anything that amazing about Urbit once you figure it out. It just wants everyone to be able to run their own server, and make it easy to build decentralized apps that talk to other Urbit servers. The important part is that somehow, they saw how bad trying to do that currently is, said "let's rewrite everything", /and then did/. It's like reading about Oberon or Plan9/Inferno.

Edit: This post is probably the worst introduction to Urbit imaginable. It basically is a technical spec for bootstrapping a PKI over Ethereum, and you should expect about as much as if you got linked something from BitTorrent about that. It assumes domain knowledge from both Urbit and Ethereum (both of which are terrible to explain), and /doesn't actually matter/ for most people interested in Urbit. Please don't use this as the benchmark for "babies first urbit intro".

I never heard of this. Like, never even in passing have I seen that bounty mentioned. One of the trivial apps that was demoed literally years ago was a simple tic-tac-toe app, and there is even a MUD written for Urbit, so saying it's impossible to write apps for it isn't true. The web server is hosted in Urbit, along with the developer chat room. And really, the problem isn't that apps are impossible to write but that they would be easier to write using a different platform, which is probably true (for strip poker at least), since you don't need anything more complex than a couple hundred lines of C.