This is really sad. I'm sorry to hear things didn't work out, but I'm still left wondering why not.
I guess I was really hoping for a couple thousand+ word post-mortem, describing the history of the project, and which hypotheses you tested, with a thorough explanation of the results.
If you weren't getting enough math input, why do you think that throwing more people at the problem wouldn't generate better content? Just having a bunch of links to the most intuitive and elegant explanations, gathered in one place, would be a huge help to both readers and writers. Students trying to learn are already doing this through blind googling, so the marginal work to drop the links is low.
Pulling all the info together into a good explanation still requires one dedicated person, but perhaps that task can be broken down into chunks too. Like, once one version is written, translating it for non-mathy people should be relatively easy. Same for condensing things for mathy people.
But, why wouldn't adding more mathematicians mean a few would be good at and interested in writing new articles? Where did you do outreach? What did you do? There are entire communities, scattered across the web, who exist ...
I feel sad that the project is gone before I even understood how it was supposed to work.
I was like: "I have absolutely no idea what this is supposed to be or to do, but smart people seem enthusiastic about it, so it's probably a smart thing, and maybe later when I have more time, I will examine it more closely."
Now, my question is... how much should I use this as an outside view for other activities of MIRI?
Debates didn't work because... well, it's a very complicated problem.
I'd love to hear about this in more detail. What have you learned about the problem? Do you know what good solutions would look like, but they're too hard or expensive to implement? Or have you learned that it isn't feasible?
Here is my person take on why it's complicated:
When you ask someone if they would like a debate platform and describe all the features and content it'll have, they go: "Hell yeah I'd love that!" And it took me a while to realize that what they are imagining is someone else writing all the content and doing all the heavy lifting. Then they would come along, read some of it, and may be leave a comment or two. And basically everyone is like that: they want it, but they are not willing to put in the work. And I don't blame them, because I'm not willing to put in the work (of writing) either. There are just a handful of people who are.
So the problem is definitely not on the technical side. It's a problem with the community / society in general. Except I'm hesitant to even call it a "problem," because that feels like calling gravity a "problem." This is just the way humans are. They want to do things they want to do.
It seems to me like you never really tried to seriously invite people to participate and write content. The fact that people had to follow Arbitals process through "whispers, rumors, and clairvoyant divination" gave the impression that it was more in a closed state than that it was inviting people to participate.
How about writing a "top 10 posts on Arbital" post for LW discussion? That way it's easier for people to see discussions to which they might want to contribute.
When we were doing debates, we reached out to a few people, because we didn't quite know what shape we wanted to the debate to take. So we didn't need that many people.
That's a valid strategy but if that's what you did, why do you think your experience proves that it's hard to get people to contribute to the discussion?
I agree with Christian. Did Arbital ever even come out of closed beta? My impression was that it did not, and you still needed to be whitelisted to have the chance to contribute.
So, if I understand this correctly, you had a non-promoted invite-only platform which you think failed because not enough people contributed content?
I am confused. Surely it crossed someone's mind at some point...
but it's not obvious when/how
Any. Fucking. Time.
I notice that you tell people to come to Arbital, but it is still invite-only.
we couldn't settle on what we wanted the users to do once they were on the platform
"...the street finds its own uses for things" -- William Gibson
Have you considered letting users play freely and then learning from them instead of trying to construct an optimal-by-some-criteria maze that mice surely will joyfully choose to run through?
That sure was an expensive way to test demand for Arbital 1.0. Have you thought about cheaper ways to test demand for Arbital 2.0?
because it will help with x-risk
A new microblogging platform will help with x-risk?
I mean, I know Tumblr is bad, but it's hardly an existential threat.
Is there actually any sort of obvious path from a microblogging platform to "the original Arbital vision"? I confess I don't really see one, but maybe I haven't understood what the original Arbital vision was.
If you can't succeed without first getting mass adoption, then you can't succeed. See the 'success' of Medium, and how it required losing everything they set out to do.
If Arbital has failed, Arbital has failed. Building neoTumblr and hoping to turn it into Arbital later won't make it fail any less, it will just produce neoTumblr.
As a software engineer, it seems strange to me that Arbital is trying to be an encyclopedia, debate system, and blogging site at the same time. What made you decide to put those features together in one piece of software?
Eliezer said he wanted all of those features. (And he is using basically all of them.) But also we worked on it for 2 years, so a lot of features accumulated as we were trying different approaches.
Will this "Arbital 2.0" be an entirely unrelated microblogging platform, or are you simply re-branding Arbital 1.0 to focus on the microblogging features?
Is there a description anywhere of the history of Arbital development?
Specifically I once heard a 4th-hand description of Arbital as "the Stacks Project, but for everything." It seemed to me at the time that a good step would be to 'simply' extend the Stacks project until it encompasses all of math[*]. I'd like to know which of Arbital's difficulties would still apply there.
What were some specific ideas you had for "solving debates"? I was hoping Arbital would take the debate around a given topic and organize it into a tree. You start with an assertion that branches into supporting and opposing arguments, then those branch into rebuttals, then those branch into counter-rebuttals, etc.
This post seems like an overly brief and vague description compared to what I was hoping for and would guess the community would be interested in.
I see finding high-quality content producers was a problem; you reference math explanations specifically.
I notice that people are usually good at providing thorough and comprehensible explanations in only their chosen domains. That being said, people are interested in subjects beyond those they have mastered.
I wonder if it is possible to approach quality content producers with the question of what content they would like to passively consume, and then try and approach networks of content producers at once. For example: find a game theory explainer who want...
I expect that micro-blogging will be an excellent combination with the arbital-style of voting on things. I especially think that you could get very good results from voting on per-post 'related links' submitted by users. Tumblr has reblogs for responding to things, but those naturally become mediated by viralness instead of internal coherence.
I can imagine a neural-activation-like effect coming out of that, where frequently co-active posts naturally rise to the top of each other's links and become threads or topics.
Are you planning anything like that?
Thank you for the summary of the state of Arbital!
It seems that while you haven't achieved your full goals, you have created a system that Eliezer is happy with, which is of non-zero value in itself (or, depending on what you think of MIRI, the AI alignment problem etc., of very large value).
It'd be interesting to work out why projects like Wikipedia and StackOveflow succeeded, while Arbital didn't, to such an extent. Unfortunately, I don't really have much of an idea how to answer my own question, so I'll be among those who want all the answers, but don't...
Eliezer is still writing AI Alignment content on it, ... MIRI ... adopt Arbital ...
How does Eliezer's work on Arbital relate to MIRI? Little is publicly visible of what is is doing in MIRI. Is he focusing on Arbital? What is the strategic purpose?
This sounds great! There is no FAQ on the linked-to website, though. Is Arbital open-source? What are the key licensing terms? How's it implemented? How does voting work?
If we're all supposed to use the same website, there are advantages to that, but I would be less excited about that.
Also, the home page links to https://arbital.com/explore/math, but that page is blank. Er... https://arbital.com/explore/ai_alignment is also blank for me. Perhaps Arbital doesn't work for Chrome on Windows 7 without flash installed.
I had asked someone how I could contribute, and they said there was a waitlist or whatever. Like others have mentioned, I would recommend prioritizing maximal user involvement. Try to iterate quickly and get as many eyeballs on it as you can so you can see what works and what breaks. You can't control people.
This post is for all the people who have been following Arbital's progress since 2015 via whispers, rumors, and clairvoyant divination. That is to say: we didn't do a very good job of communicating on our part. I hope this posts corrects some of that.
The top question on your mind is probably: "Man, I was promised that Arbital will solve X! Why hasn't it solved X already?" Where X could be intuitive explanations, online debate, all LessWrong problems, AGI, or just cancer. Well, we did try to solve the first two and it didn't work. Math explanations didn't work because we couldn't find enough people who would spend the time to write good math explanations. (That said, we did end up with some decent posts on abstract algebra. Thank you to everyone who contributed!) Debates didn't work because... well, it's a very complicated problem. There was also some disagreement within the team about the best approach, and we ended up moving too slowly.
So what now?
You are welcome to use Arbital in its current version. It's mostly stable, though a little slow sometimes. It has a few features some might find very helpful for their type of content. Eliezer is still writing AI Alignment content on it, and he heavily relies on the specific Arbital features, so it's pretty certain that the platform is not going away. In fact, if the venture fails completely, it's likely MIRI will adopt Arbital for their personal use.
I'm starting work on Arbital 2.0. It's going to be a (micro-)blogging platform. (If you are a serious blogger / Tumblr user, let me know; I'd love to ask you some questions!) I'm not trying to solve online debates, build LW 2.0, or cure cancer. It's just going to be a damn good blogging platform. If it goes well, then at some point I'd love to revisit the Arbital dream.
I'm happy to answer any and all questions in the comments.