Urbit is an attempt to fix all of the major issues of networked computing in one single system. Security, spam, maintenance, up-time, software complexity, iden

Urbit is an attempt to fix all of the major issues of networked computing in one single system. Security, spam, maintenance, up-time, software complexity, identity, network governance, etc. The system is now reasonably complete and achieves this goal in theory—all of these problems magically disappear. In order to truly succeed, they will next need to make it performant (it is very slow) and truly secure. Urbit is a better realization of "web3". It allows individuals to completely control their computing experience and makes it possible to build a web with rich social experiences, but that feels more like owning a car than like working at a multi-national corporation. People often wonder whether Urbit is an embodiment of Curtis' "radical right-wing" ideas. This is sort-of true. The structure built into Urbit's identity systems are designed to support a certain network governance scheme, which looks a bit like a republic. It allows for relative autonomy of groups and individuals and supports collective decision making at different scales. It's not direct democracy, like some web3 projects aim for, but it strikes a balance between that and the way network governance works on the current web, which is a partly-corporate network of oligarchies. It's also, crucially, not a dogmatic design. There are some cryptographic decisions that suggest certain governance structures, but there's nothing stopping anyone from governing pieces of the network however they like. On the current web, large corporations like Google and Mozilla coordinate with de facto absolute powers like ICAAN, IETF, W3C, etc. to manage the highest layer, while corporations like Google, Facebook, Apple, and Microsoft entirely dictate the applications layer for the most part. There are of course exceptions, such as this website, but they are responsible for less and less of the average user's experience. The barrier to entry to compete with established players is very high, or in the case of governing bodies, impossible. Since Urbit allows for secure communications, it shares the same implicit politics as other products of the cypherpunk ideology. Same as the ideology of the EFF, Signal or any web3 project, Urbit can be seen to support uncensorable communication and coordination between people. This is a double-edged sword. It supports "free speech," makes political oppression infeasible, and hinders certain forms of lawn enforcement. This can advance the agendas of both the left and the right. So, fundamentally, this type of project takes aim at some of the points of control wielded by the nation state. It creates better conditions for individual sovereignty vs state sovereignty. This supports Curtis' vision, which is mainly about replacing the bureaucratic administration state with more functional and decentralized governance. It also supports other political agenda like libertarianism and anarcho-syndicalism. In my experience, people who work and live on Urbit are all over the political spectrum. It's an environment of tolerance. People often criticize Urbit over its overly poetic and, arguably, obfuscated programming languages (which are designed entirely from scratch). The system is built from abstractions that are similar to, but often meaningfully different from the structures found in other languages. Instead of naming them according to their closest relatives, they are named de novo. This was done out of two motivations as far as I can tell: (1) the system was designed via a thought experiment about what kind of system could last for 10,000 years. From this perspective, today's fashions are unimportant. Will an array be an "array" in 1,000 years? (2) We can limit our creativity by using familiar words, which encourage thinking in tropes. Urbit's naming forces you to re-learn and this reconsider every part of the system. I have seen many people learn Hoon. I don't believe it's any harder than learning C++, but it is a frustrating experience for someone who has learned C++ and needs to go through the hard parts of that experience again (memorization). This is a limitation and possibly an advantage of the system. Ultimately, one should understand Urbit as a very long-term project to build an operating system and network architecture to last 1,000 years. It's not easy to learn and it's not designed to work tomorrow, but it's surprisingly functional already and I find that pretty exciting.