Urbit

Urbit's design directly comes from the founder's neo-monarchical political theories. Therefore, I think his political theories, which include sympathies for neo-nazism ('Yarvin wrote that, although he was not himself a white nationalist, he felt “the urge to defend” the ideology'), a hatred of democracy, and the desire to replace it with a model that grants more rights to wealthier individuals are relevant to this discussion.

So I submitted an article on those beliefs to HN. I'd like to see what other readers think of their relevancy to the product.

https://news.ycombinator.com/item?id=21673760

> In one design sketch for Urbit, Yarvin made the link between monarchies and the platform more explicit, classifying users as “Lords,” “Dukes,” and “Earls.” The design behind the titles, he writes, “is standard Lockean libertarian homesteading theory.” At the end of the sketch, Yarvin indicates that he’s reserved a special title for himself: “The prince (because he spent 8 years working on this project, without being paid), has reserved 32 duchies for his exclusive personal benefit."

Note: Curtis Yarvin resigned from his startup. I still find his ideas relevant, because he designed and started the startup in question.

I've never come across this project before, but I've been playing around with a similar idea...

Essentially I want to scrape data from my own accounts and anything I'm interested in following, on a home server and push that data to a sort of "dashboard" interface. To watch either on my phone or laptop or, in theory, anywhere.

I've always hated current phone interface paradigm. I see a grid of app icons like having a wall full of TV's, one for each channel. I don't care about apps, I care about intent. If I'm hungry I want my server to tell me what places I already like are nearby, whats in my budget by pulling data from bank account, and whats within my daily macros by hooking into MyFitnessPal. Or if I want to catch up on news, have my server scrape full articles from RSS feed links... Default to an auto summary view, and highlight terms using named entity recognition to make things even more quickly skimmable. If I want to catch up on sports, pull in football scores, my fantasy points, top sports news, and my sports groupchat in a single view. If the stock market drops over 1% in an hour, take over my homescreen with data on how my current positions are doing, a muted Livestream of Bloomberg, and top financial headlines that are currently trending. Not the most flushed out examples but one gets the idea. Server side processing of all the data I generate or consume, and present it in custom views based on intent. I feel like a lot of the smartphone anxiety ppl have comes down to fear of missing out. You look at your grid off apps, each one a portal into its own separate world and you don't know what's going on in each one. What you're missing out in each world, so you go into each one trying to stay up to date with each. Vs having a system that pulls all that data in and you can trust will catch you up on anything important that happened on your own time.

Breaking the current app paradigm would also make foldable devices a much more practical class of device. Instead of Samsungs sad multitasking demo of trying to stuff 3 app ui's on a galaxy fold, one could have a much more dynamic single view if the data from the apps if the data from them could cross talk.

Right now I'm playing on an old desktop computer (hard drive died so this whole thing is on pause rn), but it would be cool to get something like this running on raspberry pis using elixir (just started learning, my current scraping experience is with Python and zmq but elixir seems like a much cleaner and resilient base to manage scraping and organizing data from various sources in a more hands off way). Power users could easily scale by adding more pis or even a specialized processor if you're doing something more unique like parsing video.

If every user could have their own cluster that pulls in their own data, it would be much easier to take control and sell that data, instead of current system where it gets gathered regardless. Or not sell it if that was their preference.

For instance, it would be cool if my "cluster" knew my route to work, and that there was a certain delay, and a more optimal route was available.... But beyond that, know what coffee places are on my way, and in an ad exchange kind of way, allow these locations to have agents that bid on my business by offering dynamic discounts. Like if my reroute traffic path brings me closer to their location, and I'm running late, maybe they offer less of a discount because they provide value thru more convenience. Maybe another location sees they're now more out of the way and decides to bid lower, or maybe they see that they're located next to my favorite donut place, and the automated agents of those two business negotiate something between each other in order to provide me with some sort of combo deal. >> All this happening behind the scenes and I see a unified view of my new map route and a pop-up pins on the map of the offers if I want to accept them.

Lots of other interesting scenarios could be thought up, especially with smart homes. But in order to make that leap, I feel like the first step is to aggregate all your data into one place in order to take back some control of it and open up new possibilities of it's use by combining it together. Maybe I'm misunderstanding what Urbit is, but it sounds like a cloud hosted version of this core first step.

I don’t know whether this matters, and I guess it probably doesn’t, but he was kicked off the Popehat blog not long after writing this, for virulent (and, since then, increasingly theatrical) racism.

I don't see a problem. If things get repetitive people will start complaining, but in the meantime these posts have been upvoted out of interest.

Yarvin sounds like a loon.