Good faith assumed and accepted, thanks. Hi! (:
I've been working on Urbit full-time for ~5 years now, have been up and down the stack, and consider myself an "urbit maximalist". I'll give you my personal perspective and thoughts here. Note that not everyone working on the project (let alone hanging out on/around it) shares these exact views. Urbit is big enough to mean different things to different people.
I'd encourage others to share their own perspective if they deem it useful, or just pedantically correct the generalizations and simplifications I'm about to make below!
You do certainly sound like like you'd easily slot into the O.G. Urbit target audience. Some have described Urbit as "contrarian computing", because it essentially says "computers in general are garbage, we need to do away with *everything*."
The tldr answer to your question of "what are you working towards" is, uh, somewhat necessarily vague. I'd say "Urbit is a computer" and mean "computer" in the broadest sense of the word, encompassing the entire (hardware and) software stack. I assume you've already read about this. I'm personally working towards a world in which my entire computing experience lives exclusively on Urbit. I don't want to see or touch any Linux, any C, any Webkit. Urbit-native all the time. Why? Urbit, from a technical perspective, is smaller and easier to understand, more _actually mine_ than anything that exists today, and this ripples out to my user experience.
Now, you're correct in seeing multiple different Urbits. I'll use the ones you listed as a framework for my explanation, giving one-line summaries and then expanding a little bit. I won't skip over 1 and 2, just in case we aren't already on the same page.
> 1. the urbit yarvin wrote (and/or intended?)
The Urbit Yarvin wrote on his own was a proof of concept, showing a new way of computing.
The Urbit Yarvin wrote with the company (Tlon) was a usable alpha for a technical crowd.
> 2. the one that currently exists
The Urbit that currently exists is a young yet robust from-scratch computing stack that desperately wants a software distribution story so it can stop being confused for its one big userspace product.
The Urbit that gets talked about in marketing materials is a blurring between Urbit proper, Urbit's userspace experience, the things being worked on, and the vision for the future.
> 4. the urbit of the future that has aspirations beyond the one that exists now
The Urbit of the future is the only legal form of computing.
Let's talk about those Urbits (rather, phases of Urbit) in a bit more detail.
1. Yarvin's Urbit
You've probably read about the fundamental pitch by now: the internet is fscked because the software it's built on is fscked. If you want to fix this, you need to do away with the whole thing, learn what you can from the smoking ruins, and build anew from the bottom up.
Yarvin was bad enough a dude to actually attempt that. Many years later, he emerged from his garage with Nock ("assembly replacement"), Hoon ("C replacement"), and a prototype Arvo ("kernel replacement"). The craziest thing: it actually worked. He had built a peer-to-peer operating system from nothing.
That prototype contained a lot of interesting ideas, and many of those we're still moving forward with today. Of course, the thing wasn't perfect, nearly all components of it have been written many times over since the olden days. Some ideas abandoned, some transformed, some put in the freezer for later. Rome wasn't built in a day, and an as-perfect-as-possible software stack isn't built in a single iteration.
The minimal prototype was interesting enough to get funding, letting Yarvin start a company that he'd leave and entrust the fate of the project to not many years later.
2. Tlon's Urbit
Tlon was originally established for the purpose of developing Urbit, plain and simple. But Urbit is a research project, how do you even begin to make money off of that?
I assume you've read about the address space, and can conceive of that as a limited source of income. Nobody will want to buy address space for a network that's all abstract and technical though. You can get by on that for a couple of years, but eventually people will ask you what you have to show for all the work. "Look at this kernel!" you say, but the world sees some nerds in a basement.
Luckily, you have this peer-to-peer operating system, at a time where BTC-USD is still in the lower triple digits. This world is about to become booming, and you already have all the tools you need!
So you build a web interface for your cli chat app. You realize you want a forum, and build that too. Maybe you make these things accessible over the public internet somehow. Now it's *real*. And you do it all in a matter of months. $2 billion company for a chat app? No thanks, just have the intern do it.
This continues and repeats for a little while. You get good at figuring out what to build, how to play to your strengths. Sometimes weaknesses in lower layers of the system come to light. You spend half a year redesigning the thing, and it's worth it.
After a while, Tlon ended up with an entire suite of userspace applications, and a fancy web client for interacting with them. They're good, but they're also "the Urbit experience" for many people. When you use Urbit today, you rarely see the OS. In some sense, you never really did. You're seeing the applications.
This isn't bad in and of itself. But Urbit doesn't have a software distribution story yet, meaning you're kind of locked into whatever has been deemed acceptable to ship with the distribution that you're subscribed to.
At this point, rectifying that is on the top of the list. Relatedly, you may or may not have noticed some distinction between "Tlon" and "Urbit.org" (aka the Urbit Foundation). The former is in the business of staying alive as a company of developers and making products/services on top of & around Urbit. The latter is in the business of making Urbit itself and the ecosystem around it flourish.
This separation is very young and still incomplete in many ways. It's a gradual process, but a necessary one. Landscape has S3 integration, but Landscape is not Urbit. We're working on making it not feel that way.
3. The news' Urbit
It's easy to overpromise. We do that sometimes, or even more often we just don't communicate clearly enough. Talking is hard!
More seriously, it's generally just the case that there's overlap between the way we talk about Urbit, and the vision we have for it. Yes, indeed you can choose exactly who you get your OS updates from! But it's also the case that in the current moment, there's only one party pushing them out, and you need theirs to stay connected to the rest of the network that gets theirs.
4. The Martians' Urbit
In the future, Urbit will be a vast and diverse network of peer-to-peer personal servers. Various forks of the operating systems will run side-by-side. Many applications, clients and runtimes will be there for you to choose from. Writing peer-to-peer software will be simple (actually, I think we're already 80% on this), your data will be durable, and your computing experience will be fully yours.
Nearly all of the things we promise, we want to deliver on. As others have already mentioned, staying alive in the present demands sacrifice in the short-term. But I certainly refuse to lose sight of the long-term vision, and I'm sure most everyone else feels the same way.
...Was that too much? Maybe that's a bit too much prose and too little substance. Let me address some of your stated concerns (that haven't been included above) to close this out:
> alright. my iphone basically does this, who cares: except for the stressed concepts of ownership and sovereignty(?) that this seems to negate
Does it negate ownership and sovereignty if you choose who to trust?
Urbit is a high-trust network. You've probably read about how identities are scarce, and how this incentivizes people to behave well with it. There's of course all kinds of game-theoretic rabbit holes you can dive down here, and ultimately people aren't rational actors after all... but you're always in control.
Right now, your *practical* range of control is choosing between "I pick someone to spoon-feed me the same updates as everyone else" and "I don't get updates at all". As previously stated, this will improve in many different ways.
(Actually, that's not entirely right. There's already a small amount of people pushing out lightly-modified distributions with new applications in them!)
> the announced goal of getting people's cloud-independent personal servers back into the cloud
I definitely agree this flies in the face of the old-school marketing copy that said Urbit should be simple enough for your mother to administer.
Is it simpler to run than a full Linux server? Yes. Do you still need *some* knowledge of Linux servers to run Urbit? ...Also yes.
You can imagine a world wherein Urbit is a .app or .exe or some such thing that anyone can trivially install on their laptop or phone. I believe a community member is working on just such a thing!
But that doesn't invalidate the *option* for someone to offload the self-hosting risk to some other party. For example, for my work machine, I keep backups on an external drive. The external drive is in my house, where my work machine lives. If my house burns down, I'm screwed. If I was a smarter man, I'd pay someone to keep a copy of my backups off-site.
> or possibly a pyramid scheme? (this is not moral judgement, if it's a scam to do fun computer experimentation, i find that admirable. i'm just trying to figure it out.)
The detractors are right. Urbit is literally a pyramid scheme. It's also performance art. And all of that is actually good. (;
(You might also find some useful counter-arguments on the "common objections" page, in case you haven't seen that yet.)
https://urbit.org/blog/common-objections-to-urbit/Hope this is a useful perspective to you. Again, from your background, Urbit seems right up your alley. But it's also initially opaque, and heavily in development, and dealing with mulitple-personality disorder.
If you ever feel like starting your comet back up, don't hesitate to DM me (~palfun-foslup) if you want to chat. Thanks again for enquiring in good faith, that's the Urbit way. (:
—
~palfun-foslup
https://urbit.orgfrom Hacker News https://ift.tt/3sLmPcS
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.