Flying through The Atmosphere with Dan Abermov

Justin Garrison (00:00)
Hello and welcome to Fork Around and Find Out, the show all about software maintenance, AI hacks, and being fed up with the Fediverse. I'm your host, Justin Garrison, and with me as always is Autumn Nash. How's it going, Autumn?

Autumn Nash (00:13)
Great, but that joke was so bad.

Justin Garrison (00:15)
I'm sorry, I'm

sorry. You know, like it's just it's it's a it's it's there.

Autumn Nash (00:18)
You're c now

we're gonna get Mastodon Hate Mail. Thanks, Justin.

Justin Garrison (00:22)
Mastodon's fine. but it is something I do want to talk about today because I I I spent some time with Mastodon and specifically their their protocol activity pub when I was first switching off of Twitter and figuring out like what the heck are we doing? And then Blue Sky came along and I was like, actually, I kinda like this design a little more and just the fluent like how it worked. I know people have opinions about those things. And so we brought Dan Abramov on the show. Welcome to the show, Dan.

dan (00:47)
Hello, thank you for inviting me.

Justin Garrison (00:50)
Yeah, thank you for coming because you've been doing a bunch of just like podcasts and blog posts and various things trying to share what app protocol is all about, what the atmosphere is about. And so we wanted to have you on the show and just talk about what it is and and how you got involved. So can you start there with what you were doing and why you started sharing all this information?

dan (01:09)
Yeah, sure. so I actually used to work at Blue Sky for a couple of years. And when I joined Blue Sky, I really was working on just the app itself. So there's like a React native app and I knew some React. and so I didn't really care for the protocol or really understood what it's about. And as I was working there I learned a little bit more about it. And like in the beginning, you know, I was thinking, well, what's

what do they even mean by, you know, developing a protocol? Like the protocols that I know are, you know, things like HTTP or FTP or like SMTP, you know, like the protocols that I've heard about. So it was kind of weird that, you know, why would you build an internet protocol? And is is that even like a real thing? Or are they just doing some kind of marketing around that? Or like what what is it? And then as I was kind of learning about it because

You know, when I was working on the app, I kind of started bumping into some pieces of Ed Prato or like the Ed Protocol at Prato is how they call it, like in a shorter way. I actually got pretty excited about it and I've so I've left Blue Sky a while ago, but I'm still, you know, blogging about Ed Prado and I I mean I think it's very exciting. like it's one of the most exciting kind of takes

on the web that I've seen, kind of possible evolutions of the web. And so I I think all attention has been you know sucked out into like the AI space right now. And I actually think like Ed Prado really deserves a closer look from the web developer community. And so I'm I'm just trying to get people Ed Prato built. And that's why I'm doing this, you know, this podcast tour. just trying to kind of

you know, get people excited about it.

Autumn Nash (02:54)
What was the change for you? Like what got you excited when you started looking into it? Like

dan (02:58)
Mm well I just thought it's like I got excited about like the idea itself made a lot of sense to me. and I I just thought like it really has the potential to kind of save the web. You know, that's that's like a big the the the that's like a big statement, but I think we've all been feeling that there's something going wrong with the web, like over the last ten years maybe. that like it's kind of

becoming less open, it's kinda getting more fragmented into these like closed spaces where, you know, rent seekers are just starting to like, you know, charge rent on APIs and all all of that stuff. And it just feels like it's very disconnected from this kind of open I think spirit of the web that initially, you know, when you learn about how the web works, like it's it's a very kind of open thing. And so I think we've been losing some of that. And I think at Prado kind of

bridges that gap between you know, the user experiences we actually want in our apps, while kind of porting them onto the open web topology so that it actually feels like, you know, like all these like social experiences, social apps we have, they can become a part of the open web like in a very meaningful way. And it changes the product dynamics as well. So like we can have more competitive marketplace where products don't really trap you the way like

social media companies today to you. So that is that is kind of why I'm excited about it.

Justin Garrison (04:18)
I I wanna back up with something that you said about this being a protocol and and not in the this is H E D P sort of protocol, right? Because it it does build on top of what this audience might understand. Like this a lot of our audience is like infrastructure network people that when they think protocols, they're thinking like TCP IP, right? Like these are like how do I send something on a wire? and in this case, this is more of like a a data transfer schema of how do I get data from one place to another. And in that

itself has gone over various evolutions of things like IMAP for email or I actually was just talking about this last week, SOAP, the XML sort of like pre-JSON APIs, right? Like we were s we were sending all this XML over the wire and it was called SOAP. And and now like app protocol kind of fits in that layer of like how does the data transfer? not necessarily at the lower level. This is a what layer seven, like app layer in the TCP stack. But

What is that built on? Right? Like the that fundamentally has other pieces that it relies on under the hood. And I know we've we've made new terms such as what is the the XRPCs and other commands that we do for PDSs, which stand for personal data service. I know not everyone else is is familiar with these these acronyms, so I'm trying to explain as we go. But what is what is app protocol build on?

dan (05:33)
sure. Yeah, I I mean F Proto kinda builds on top of a bunch of things that we already know. So it's HTTP, HTTPS, JSON, DNS, WebSock. I think those are the main ones. And then there's a few like more niche things like CBOR and there's stuff that they have to invent themselves. So that's that's mostly Lexicon and I think a couple of other things.

but I kind of like to explain the protocol from a different perspective. So I think if you just, you know, talk about these like different pieces, it's still like hard to understand, you know, what is this about. So the way I kind of converged on understanding like explaining it is that the idea is just to upstream two different things into the web standards. So one thing we want to upstream is a notion of identity, which is kind of similar to like how you might have like a Google account.

That you used to sign into everything, except that it's silly that this is tied to one company. So what if we had like a notion of like almost like web accounts, like upstreamed into the web standards? And then the second idea is just that what if we gave every web user kind of a folder? So if you had a folder, you know, kind of like a Dropbox folder, where apps can put stuff into your folder and they can kind of subscribe to updates from everybody's folders.

And this is really I think like largely all of these like other things they're really in service of these true ideas. So let's give everybody ability to create identities and then let's you know have some kind of place where apps can put pieces of JSON for those identities.

Justin Garrison (07:03)
And that's when you're talking about the lexicons, right? Like lex lexicons are are basically the types of files they would put in folders where someone's gonna send me a a blog post, it's gonna be in the standard sites. Lexicon, if they're sending me likes or something from my post on Blue Sky, it's gonna be different versus like images or videos, right? Like those are all the lexicons, kind of going I guess those would be like folders inside the folder, necessarily, right? Like that's the is that how you would explain that side of it?

dan (07:28)
Well, I think the the idea is that we want to you know, usually like when you for example, if you post a tweet on Twitter.com, that tweet goes into Twitter.com's database, right? So it's kind of like inside this closed thing that you really don't have any like real access to. The only way you can read it is through Twitter's API. you know, if Twitter closes the API then you would not be able to get the data from there.

You can't run queries on the database, right? Like it's because it's not yours. It's twitter.coms. It's like in their data centers. So in the sense like even if we create something publicly, it kind of still goes into these like silos that that are inaccessible to the you know, that are not on the open web. And then with that proto, the idea is that let's just lift all the public data out of the apps and into the open web itself.

So there is like a folder somewhere, kind of like a Git repository that is open, you know, that anyone can query and access and aggregate from. And that is where your JSON goes. And then like Twitter.com is just one you know, just one consumer of that folder that like aggregates stuff from everybody's folders. And so Twitter's database would become kind of like an index or like a cache of the entire network. And so if you know the

Then the problem becomes well how do you store if you have a folder with like all up all the different applications data that is, you know, public data from all the different applications that are yours, right? So maybe like if Reddit used that Protot, there would be like a subfolder in your folder that's like all your Reddit posts or like all your Reddit comments.

Or if you know, if Stack Overflow was on F Prado, then there would be like a Stack Overflow subfolder in your folder that has all your questions, all your answers, all your votes, right? And then the question becomes, okay, but like if we put everything, like all the applications data into your folder, how do we actually you know, like how do we decide their structure? And so the way F Prado solves this is it's actually like up to the applications to decide. So nobody needs to agree on like a single notion of

what a comment is or what a post is. It's kind of per app. And so in like my Ed Prado folder, there would be, you know, old Blue Sky follows. They're in the format that Blue Sky app decided like what a follow is. Or like old b old Blue Sky likes. Or all Reddit comments, right? Like all of these things are kind of like data types. And so the you could okay, then now the question is like how does an app specify data types?

And so one way to do that is with you know something like TypeScript interfaces, right? Like that's or like JSON schema or something like this. and so with Prado uses its own format b that is kind of like a JSON schema that is called Lexicon that can be compiled to different you know TypeScript definitions or Rust definitions or like to any language together with some runtime parsing helpers that give you strongly typed objects.

So Lexicon is kind of like a JSON schema format for, you know, defining schemas that can be converted to TypeScript interfaces or whatever that let you parse things out of users' folders to, you know, whatever format that your application actually needs, and you choose that format when you design the applications.

Autumn Nash (10:43)
I get really excited about app protocol because I think we've all spent so much time investing in building communities and then that community gets sold or it changes and you lose so much time. So it's really cool to think of the idea. I mean, like you know, Instagram was great and then it got sold to Facebook. Facebook changed. Twitter got sold, so

What do you think is holding app protocol back, I guess? Cause it seems like to me it makes a lot of sense and it seems like it got really popular when blue skies first started expanding. What do you think is stopping it from being like fully like the next Twitter or the investment in the amount of people that maybe Twitter or Instagram or Facebook has? What do you think is stopping it from getting to that level right now?

dan (11:25)
Yeah, I I I think it's multiple things. I think one is just that well a lot of people don't know about it, right? And it's it's kinda cutting in both directions. It's for like a lot of web developers have no idea that it exists. So there's a part of, you know, why would I use it for my you know, social app like if you're making a social app, or even some app with like identity here, right? Like where you offer users to log in and then they create something there.

just not being aware that this is a thing, that there's this atmosphere, which is you know kind of like blogosphere, but for like all the applications that I use in At Prado. so just not being aware of of this movement. I think right now it's kind of like a idealistic movement, maybe, because just like open source was in the nineties, right? Like if you're a developer and you

Like i you need to kind of buy into it to use it, right? Because it requires some amount of effort to actually like architect your application this way. So I think part of it is just like people not being aware of it. a large part is consumers not being aware of it because it is kind of like the explanations are kind of technical. and I think it's I think there are surface level explanations that are actually quite understandable, such as, you know, what if your all your apps

road to something like a job box folder and then they wouldn't be able to trap you and your communities, right? Like this this is something that maybe can resonate. But ultimately I think it has to win by having better products. So it's all like we need developers to start adopting it for you know for it to grow and for successful products. There there are some products that are getting I guess like somewhat popular. There's this tangled

which is like Etprato GitHub, and there's Leaflet, which is like Etprato blogin. there's actually like a couple of those platforms now. So I think it's the thing that's stopping it is primarily just more products becoming successful using Etprato under the hood. I think part of it's core part of the curse of Etprato is that it's actually designed to be invisible.

So it's you know, it's not designed to be this weird decentralized thing. It's kind of more like under the hood, your product just becomes, you know, much more interoperable with other products that use a proto. But like as a user, you don't necessarily even notice some of that until you see that like different apps can kind of show content from each other or like riff on each other's data in some way. so yeah, that

Justin Garrison (13:51)
Which is something

dan (13:53)
yeah.

Justin Garrison (13:53)
Like I feel like we've been

doing that for a long time. Right. Like that that idea of I want to pull content from somewhere else into like my blog. I remember back in the day I had a WordPress blog and I put like a plug-in that would show my tweets on the sidebar. Right. Like it was just a way to like mix

dan (14:06)
Mm-hmm.

Justin Garrison (14:06)
and say, This is what I'm doing, this is where I blog, but also here's me on social media and here's me on YouTube. And you try to do that already. We've been doing that for a very long time with mixing those things, but they were all different systems. And you always were like, Well, I have access to that API

s for now, if if someone removes it, I can't do that anymore. And I don't think that's new, but I think that combining that into who you are as like here's your again, a a personal data server is is like just your stuff. Like here's all your stuff, here's all your interactions on the internets, at least in this like social vein. one of the things that I think to make people care or at least people make people understand is like most of the things you described, I think we could also have built on top of

Activity Pub, right? As far as just like ways it interacts. I think that the fundamental instances in activity pub and in Macedon is a very different thing than a PDS. How do you describe those differences?

Autumn Nash (14:59)
I almost

feel like it's also like if the things we wish RSS feed could have like grown into or like if RSS feed and activity pub had like a baby.

Justin Garrison (15:09)
Well yeah, and I I

I mean, ActivityPub is basically just RSS feeds, right? It's just like a one way with like some web hooks and and app protocol is is a s a lot of the same stuff. I usually describe both of them as like an interactive RSS.

Autumn Nash (15:18)
what I'm saying, like

Justin Garrison (15:21)
It's just like it's web two for RSS.

dan (15:23)
What was the question again?

Justin Garrison (15:24)
How do you describe the differences between App Broto

dan (15:26)
Yeah.

Justin Garrison (15:27)
and an Activity Pub?

dan (15:29)
Yeah, I actually like the RSS comparison a lot more. And maybe because Activity Pub is, you know, it it doesn't really specify what you do with it. It's kind of like, you know, there's a mailbox and like what what is the actual architecture that that you want to make it? Like it doesn't really say like it doesn't prescribe or even like suggest any specific architecture. And so when we talk about activity pub, like I think in practice we often end up talking about Mastodon, which is, you know, the most popular

kind of activity

Justin Garrison (15:57)
Mm-hmm.

dan (15:57)
pop thing. but then if you criticize it, then people say, it's it's just because Mastodon is doing it wrong. But it's like, well, I I I'm I'm just gonna kind of collapse these two concepts into, you know, maybe like Fediverse or Mastodon, whatever we want to call it. But the I think the big difference is like to me like activ Mastodon world is kind of like

You know, you want to decentralize you know, take something like Twitter, right? Like you kinda want to decentralize or like give more p power to the people. And so what they do is, okay, let's make it hosted, right? Like let's make it a piece of software you can run yourself, like like WordPress, right? You just you just push it to your hosting and it's it's a server and you can log into it or register there and it's like it has its own database, you know, it's like a small thing that's like completely isolated.

And so like Mastodon's solution to giving power to the people is let's just have a thousand Twitter copies. Like let's host a thousand or like ten thousand or I don't know a hundred thousand of mini Twitter apps. Everybody runs their own app with its own database. And then let's somehow try to bridge them. And so in like in Mastodon world, you can, you know

Like normally you log into like yours you have to pick a server, right? Like you have to pick where which of these cause it's completely unrelated apps, right? Like they have nothing to do with each other. It's literally copies of the same thing that like independently hosted. And so you have to pick like which one you want to register on. And then they have these like mechanisms to talk to each other because they have to somehow give an illusion of a shared network. So if I want to follow somebody, but they're not on this instance.

these like servers have to talk to each other and tell each other, okay, like I want you to kind of email, you know, like I think of it almost like email, like activity pop. Like just like email me all posts that this person is posting so that I can ingest them and like also display them. And so it's kind of like many to many where you know all instances potentially have to talk to each other, but each of them is like a completely separate world.

And they're trying to s somehow like bridge those gaps. And that you know, that is kind of how like it's it's almost like a very like feudal system. Like I kind of imagine it as maybe you know, fiefdoms in ancient China is you know, my mental models with like horses running between those nodes with like messages. And then if you know if like if two admins decide to have a beef, then you're not gonna be able to like follow.

Or like you you maybe follow a friend from another instance, but this instance stops talking to your instance, you can't do anything with it like about it. Or if your instance goes down with your account, like your literal ID was you at that instance, right? Like you kind of belong to that clan or you know, to the to that specific thing. So it's like it's a very kind of I think of it as like very kind of disconnected sh shape.

And Etprato, I think, is much more like RSS and like the analogy I like to draw is like RSS and Google Reader, where the source of truth is everyone's h like everybody has a hosting, whether they know it or not. Right? Like when you register, just like you register with Google, like they're gonna host your data. Like it's kind of similar. Like you register on your first Ethprato app, it's gonna host your data.

But actually like hosting is disconnected from apps. Like hosting is app agnostic. It's just a folder. And like you can move you can move that folder who hosts it. And so it's kind of like the these two layers are completely disconnected. I can change where my you know, where my folder with JSON leaves. I can change the provider if I want to. And I can access all my stuff through any app. Because all apps are aggregating from every hosting.

So it's it's it's kind of just like a different shape where your like your identity is not tied, you know, it doesn't matter kind of where it's hosted initially because you can always change it and that like you don't lose any of your followers or any of your data or like any of those things because there's j they're just in your folder. Like you can always back it up.

Justin Garrison (20:10)
Yeah.

And I think that universal ID is the anchor point there, right? 'Cause like you had to you have to solve identity first. And a

dan (20:17)
Mm-hmm.

Justin Garrison (20:18)
lot of people in the infrastructure space are always like worried about lock in at certain layers of the stack. And and I've always I said that like identity is the biggest lock in you can have, right? Like as soon as you tie your identity to Active Directory or AWS IAM or something like that, like you're locked. Like that's it. Like you're not moving that thing. And

And your point there of like if I if I lose access to Gmail, right? Like I've lost my Gmail account forever because that's tied to Gmail server and I don't have any portability of that portion of it. but

if for like RSS feeds, like I've kept my Google Reader RSS feeds, like I've moved multiple times to different providers by like exporting my OPML you know, list of all my scroll all these RSS feeds and I just go to another app and I don't even care because all the data is coming from somewhere else and I can just aggregate it one layer above that and I can continue using that no matter where I want because it's just a standard open protocol.

dan (21:12)
Mm-hmm.

Justin Garrison (21:12)
My one of my biggest questions about Blue Sky and at Proto in general is is this original promise that I was kind of sold on when I was first like, okay, forget, I'm not doing Mastodon. For Mastodon, like I was I was really trying to like it. I was I was trying to simplify it and I think that disconnect of like an instance

is an active piece of software that owns everything was really hard for me because I was like, I have to run an instance of one for for one and it had to do more complex things than a p PDS had to do. And so that really was hard for me. But the the also Macedon in general, Activity Pub's the standard, it's it can it can l exist outside of Macedon, like Macedon in general had no business model originally,

dan (21:50)
Mm-hmm.

Justin Garrison (21:50)
you know, two, three years ago especially. And everyone's like, How are you gonna fund this? Like like

volunteers and donations doesn't fund these yeah they they don't

Autumn Nash (21:57)
It's also a very expensive process. You know what I mean? Like

Justin Garrison (22:02)
yeah the the actual people that were running this at larger scales like we had people on the lat on Shipit right that were running the Hackyderm and Hackyderm was costing a lot of money but actually the main thing that was a problem was like the legal record like the legal demands of if if someone if one of your users has CSAM or even if you peer with a

other Mastodon instance that is doing that, you're gonna be caching that information and now that is on your server and now that is a problem for you legally in a lot of places. And and I was like, wow that's that's a big problem. And that's something that in the app proto case doesn't exist because like my PDS doesn't talk to your PDS directly. We all go we all get funneled up to the application layer.

Autumn Nash (22:42)
I also think there's another layer of being approachable, right? Like for one, like you have to make a lot of decisions for Mastodon before really getting to use it and understand it. Like how do you know what server you want to pick? Like I feel like a lot of it requires a lot of knowledge that you may not have had previously. Also I think like a lot of people that are engineers like Mastodon, but it can be intimidating for other people. I think Blue Sky gives you the option, like it's kinda like Linux or like other places.

that like some developers want to be really into the weeds and they wanna pick every single like configuration and some people just wanna use it, right? And I think

Blue Sky gives you both of those. You can get as like into it as you want into the whole protocol, or you can use it like other social media apps where it's not super complicated and your like level of complication could or you know involvement can either expand or stay the same, but it's very approachable in my mind compared to the opposite.

Justin Garrison (23:37)
And and

and as a developer, like I I vibe coded a couple sites just for myself that only interacted with Blue Sky for the identity piece. Like I don't do any of the lexicon PDS storage. Like none of that's there. It's literally just like let me log into this app. It's just at that point it's just a single sign on that like Google was providing me before, right? It's like, hey, I just need an identity that I can trust and I don't have to pro you know, build all that into the app directly.

Autumn Nash (24:02)
The protocols made me more excited about RSS feeds and like the internet because I think like like there's so many people who invested years into like Twitter and I don't mean just like to have a fancy following, but I mean like the community. Like I met so many people and friends from Twitter and you invest a lot of time into like, you know, making videos for either YouTube or TikTok or like building a community on Twitter and then it's all just gone, you know? So I think that it's almost like revived a lot of the excitement

about social media and the web and interacting with people.

Justin Garrison (24:33)
And and that's my biggest question here is the promise I was I was provided when we were moving moving from Macedon into into Blue Sky was like this is what they call it, billionaire proof or whatever. Like no one can buy this thing 'cause it's not one thing. anyone can fork this and like just consume the fire hose like build another one. Dan, do you think that's real? Do you think that's actually the case?

dan (24:53)
that's a load of question. I think

Justin Garrison (24:55)
I know it is. That's why I'm asking you.

dan (24:57)
I think it's like it gives us the tools to, you know, reduce the like obviously like Blue Sky is the biggest app, right? but what they're trying to do is like they're upstreaming as much as they can into like the standards, right? Like so there's IETF working group for It Prado is currently focused on

I think like two kind of core pieces like the repository structure and something else. I I don't remember exactly. But basically the goal is to you know upstream as much as possible so that you know even if Blue Sky eventually turns like a instant proto, you know other companies can keep it running or like other people can keep it running. relays have gotten like run in the relay has gotten much more like

accessible because it used to cost a lot of money. Now it costs like thirty dollars a month, with the current volume, I believe. So some people are like renting stuff on Raspberry Pis and so on. So I think like broadly speaking, I think it is true. I think there is still of course, you know, like like Blue Sky as a company that is kind of behind the protocol has a lot of weight, right? So it's you know, if it's going to try to snuff it out, I think that

that would be bad for the community, right? Or if they try to like do something bad with it. but I think like all the tools that we want are there. Like, you know, all the important core code is open source, of things are getting upstreamed into standards. So it's really a question of like what do we want to do with it, you know, and are are we building something with it or are we just criticizing?

Justin Garrison (26:26)
And and how does that like any anytime there's free information on the internet, it gets abused. Right. And I think that that is something that all these like one of the selling points of having a PDS or or at protocol in general is just the fact that you can all of your data is there, anyone can crawl it.

Anytime they want. They can go discover all the PDSs and then just start crawling all the data. And that's right now fine because it's still a relatively small network with relatively small amounts of data. But I guarantee you all of the like AI crawlers are are just sucking up that data, right? If it's free

dan (26:57)
Mm-hmm.

Justin Garrison (26:57)
data on on the internet, they're gonna they're gonna consume it. And whether that's from the fire hose or directly from the PDSs, I know there was a a thing last year where it's like AI training was happening, right? And it's like hey

like this wasn't authorized by Blue Sky. Like this was just something that like they started doing and they started sucking up the data. So now what's how do we how do we protect that long term? Because that's the maintenance side of it that is the hardest problem. Once you put anything for free on the internet, it's gonna be widely abused.

dan (27:24)
Well, I think it's and maybe, you know, maybe that's just uncomfortable for people, like I don't know. But I think El Prado kinda takes like a very maybe maybe like a very direct stance there, which is this is a thing for efficiently disseminating public data. And it tries to level the playing field by making it equally accessible to everyone.

And so it's kind of like, you know, like for me, like my personal, you know, like take on it is you know, when I'm posting on Stack Overflow, for example, that that used to be a thing, right? when I was posting on Stack Overflow, like my goal is to contribute to kind of humanities knowledge, right, on this topic. My goal is not to contribute to Stack Overflow's ability to monetize on this or whatever. Or like when I post on Twitter, like it's it's really more like I'm

Like I think of it like I'm I'm putting a public thing out there. Twitter is just the medium through which it is accessed. I'm not like contributing IP to Twitter's, you know, worth it as a company or something like this. So

Justin Garrison (28:23)
But you are. I mean in

i every social network that makes money off of the content of its people, you're working for free. And that was the thing for me.

dan (28:30)
Right, but I mean like in terms of

intent. Like I it it's not what I'm intending to do. It's just like it just happens as a by product of me, you know, using this app. But my point is just like I think with F Prado, the idea is just that no, like if if if we're putting things in public anyway, let's just make them fully public. Just the websites are public. Right? Like when you put stuff on your website, there is no question of like who's accessing your website. Or like not you know, unless they're like

costing you a lot of money with with like crawling and so on.

Justin Garrison (29:00)
Well then that's what that's the problem right

now, right? Is like I I literally do see like the AI crawlbots come to my website and thankfully like my hosting has a a generous free tier and so it's not costing

dan (29:11)
Yeah.

Justin Garrison (29:11)
me money right now, but I can absolutely see that it would cost me money at some point, and then I'm like

dan (29:15)
Yeah, but that's a that's a completely

different problem though. It's it's it's not what we're discussing.

Autumn Nash (29:18)
I see what Dan is saying. He's saying that that's not

his intent, right? Like that's not why he did it. And like to a certain extent you only have so much control over what people are maliciously gonna do. And I think that that's gonna continue to e evolve. You know what I mean? If you told us five years ago we were gonna have to fight bots to the point of it costing people as much money as it does, like I think people would be like

You know, like I think that it's going it's gonna continue to evolve.

Justin Garrison (29:45)
And and

everyone has the agency to like spend their time wherever they want. And that's the thing that like for me was like when when Elon bought Twitter, I was like, I read his biography w like a decade ago and I was like, I will never work for that man. He is an idiot and a narcissist and I I do not want anything to do with him. And once he bought Twitter, I like, if I stay here, I'm working for him for free.

And I'm never doing that. And so I that's when I close my account. I'm like, forget it. Like I I don't even care where this is going. Day one. I'm like, I'm out. And

Autumn Nash (30:13)
Do you think at p

some possibility at some point we'll be able to blo block like certain websites that maybe take out protocol but we say we don't want to contribute to them? Do you think at some point we'll get to there where it'll be

dan (30:26)
Yeah. So so that that is kind of the I think you know like the the idea is that it's similar to personal websites, I think. except it's it's like you're Jason. So in that sense, like on a fundamental network layer, the idea is like everybody like everybody is a scraper, you know, j even like the first party app, like you're posting on Blue Sky, but actually the Blue Sky app puts JSON in your folder and consumes it.

So a blue sky app is just another scraper. And so that puts everybody on equal footing. But then of course, you know, there is questions, yeah, like do I want that to be indexed, you know, how do I express intent? Like maybe I don't want to like I I I I just want to separate this thing about AI is hammering my website and like I'm paying a lot of costs. Like that is a completely different problem from I don't want my content to be like scraped by AI because I don't want it to

you know, be represented in the in the training. That is that is like two different concerns. And I'm saying like that for the second concern for like I don't want to be in the training data, the solution there is I think similar to Robert's TXT. So we develop you know, we agree on some kind of standard for declaring do you want to be you know, do you want their content to be digested by AI? And then it's a you know then it moves into

like do crawlers respect that? is there legislation that forces them to respect that? And so on. So it kinda moves into, you know, like one layer below. but yes, you can't really, you know, you can't you can't kind of block it on a tech technical layer unless again it's done similar to how like Cloudflare tries to detect like AI traffic or you know, exclusions based on user agent and so on. So it's I think it's very similar to how it works in on on the web in general.

Justin Garrison (32:10)
And one of the things that we found that I think has changed in the last probably five years is the like robots.text is a

It's just like a convention, right? It's like, hey, we we please don't do this. Right. Like it's it's not an actual hard block of like you are not allowed to do this. Like, yeah, like all the stuff in the CGI bin, like, don't go there, please. and some of that was to like make your discovery a little better, but a lot of it was just to like, I don't want you in this spot. And it was generally honored by scrapers and and crawlers that were doing it for search indexing. And I don't think that that holds true anymore with like we need legal precedent.

that is like gonna get people in trouble if they do that. or we need to have harder blocks on the you are denied from this in that like my PDS is hosted on a you know like a five dollar a month

you know, VPC somewhere and I'm just like, it's like it just runs. But if like it started getting hammered, like I it would go down. I would get DDoSed or or something. Like I would consume a bunch of network if people like found it and and were like hammering it. Which is which is a valid concern of self hosting in twenty twenty six. But like I need to have some other way to like I don't want to ha set up like i IP tables to like start blocking this

dan (33:16)
Mm-hmm.

Justin Garrison (33:16)
stuff, right? Like that's that's the opposite direction of where we want people to have to go with this. And I know like

If you're on Blue Sky's network, great, like you're getting a free PDS and and you're, you know, being hosted by them. But say Larry Ellison decides that he wants to spend a billion dollars and come buy Blue Sky. Like a billion dollars is a lot of money. I bet someone's gonna say yes to that. And now Larry owns all the PDSs. So now what do we have to do? And everyone's like, well, you have to go self host your PDS. And there's some great migration tools out there, but it's not completely

you know, bulletproof from this happening.

dan (33:48)
Well, I think it part

of the question is like why why is it bad if Larry Ellison you know, like if the PJS contract is respected, a PJS is just a neural, you know, neutral app agnostic storage. It's like owning, I don't know, GoJaddy. Like, okay, maybe Go Jaddy is not a great company, but it's not really changing what's on your website, right? So I think part of the point of at Prado is it actually doesn't matter because if if there's enough incentives

to not break the network because otherwise, you know, all kind of all kinds of shit breaks. And to preserve its shape, its shape actually kind of you know, like it adds a bunch of new constraints that you don't really benefit from you know, like you can control the app layer, but then somebody can make another app. You can control the hosting layer, but if you do shady things, people can just swap to another hosting. So

Justin Garrison (34:38)
But but

GoDaddy can take down my website, right? Like as as the hosting provider, if I'm hosting stuff that's not that's against our terms of service, then

dan (34:46)
Yeah.

Justin Garrison (34:46)
then they take it down, right? And that's I think one of the problems we've had when we look at these buyouts, right? Like the problem with Twitter was that the ch the terms of service changed. Like all of a sudden things that we were doing a while ago, like are like, you're you're not you're gonna get blocked or banned or whatever. The the terms have changed because of the ownership and people

that don't own their PDS will just as easily get taken down as as you would at GoDaddy.

dan (35:10)
Sort of with a couple of corrections. I think one correction is just yes, GoDaddy can take it down, but then you can you know, if you have a backup of your site, right? Like if you just regularly back it up, you can put it somewhere else. Right. You can put it somewhere else, you can point the domain at that other server and it's not gonna break any links. So your site is kind of back on the web, you know, if somebody else is happy to host it. So GoDaddy doesn't really have the same kind of power that like if Twitter takes it down like

You're gone. You're you know, it's you can't like reappear somewhere else, right? That you're like completely gone. So I think that is one difference. Another difference is just like I don't think it makes a lot of sense to take like why you know, if you buy the app, you just ban people on the app level. Like I I don't know that it's that useful to ban people at the host and layer.

Justin Garrison (35:58)
There's plenty reasons I think to ban at the hosting layer, right? Like if if they don't want it there, if they want to really include people. But I I want to go back to that, like you could just move your stuff and have a backup, because at the very bottom level of identity in at protocol is essentially a certificate, right? Like a cert that is a a private key that signs all of my posts. I can't get that out of Blue Sky, but I can

right but I can create a second one and say, trust

dan (36:22)
Mm-hmm.

Justin Garrison (36:22)
this one too.

Is that right?

Autumn Nash (36:25)
But you get to take all your stuff with it, right?

Justin Garrison (36:28)
I mean but like i i it can be too late, right? Like to like if Larry comes in, buys it, my you know, i if you have your PDS there, it's gone.

like you're too late now, you should have been backing up before, right? That that has happened in like I see plenty of people with like Google accounts and GCP and stuff like that. Like, look, Google shut down my account. Like all of my websites down. Like it's gone because they hosted my stuff. Technically, it's all Docker containers or all VM images, or I have all the code, but like I can't just move that somewhere because I need to change DNS and I need to do a bunch of work to get there. And I think that at Proto and Blue Sky have the same problem.

of of yes, maybe the the data can be portable, but you can't just go create it somewhere else necessarily. And if I go create a new PDS with my alternate key and import all my data, like Blue Sky doesn't have to crawl me. It's like, no, you don't exist anymore. Like you can like no it's even worse than Shadow Band, right? Like you just you cannot be in this protocol layer.

dan (37:21)
I'm I'm not sure I quite understand what you mean. I mean it's also I think we're kinda starting with this premise that this is a protocol to be unbannable. I don't think that is the premise. because every app needs to be able to block its users. That's just you know, that's just like how apps

Justin Garrison (37:35)
Mm-hmm.

dan (37:35)
work. Gener like as a developer you want to be able to ban any person. So or like as a as a hosting provider you also want to be able to like arbitrarily suspend people. Right. This this is just part of how these things are supposed to work.

So I don't think like this this premise that the protocol is supposed to be resistant to that is quite true. but I still like I think with with the case of if your hosting is actually hostile, what you can do is you can register a rotational key with the with the identity registry. And then if you like if you don't trust your hosting, and then if the if like if your host and tries to do something, you can override it.

with a rotational key and you can point to another host in. So you don't need your hosting's permission to actually swap it out. I think that that

Justin Garrison (38:22)
Mm-hmm.

dan (38:23)
like one difference.

Justin Garrison (38:25)
And I guess that's something that I would love to see Blue Sky do by default and say, Hey,

dan (38:28)
Yeah, I but yeah, totally yeah.

Justin Garrison (38:30)
you signed up for an account, we are gonna own your identity, but here, go generate a pass key and you own that.

And if you ever need to

dan (38:37)
Yeah, yeah.

Justin Garrison (38:37)
create a new identity or whatever, like we're gonna trust the pass key first or something, just just to get like if you have a password manager, trust it on your device, whatever it is, like I would love to see that as default in Blue Sky to really say, like, hey, we we want you to trust us and we're gonna trust you. And and I I maybe we can't get the pass key back, but like we have your identity for the first certificates and and this is your PDS that you can own here. But yeah, if you won't need to go run another PDS somewhere, you can use that pass key to like bootstrap it and say also trust.

Yeah.

dan (39:05)
Yeah, yeah. Yeah, no, I wanted I think like I want it it's a tricky problem because you're also giving people like a way to, you know, get scammed. Like if you're if you're giving out like private key material to users, like they're

Autumn Nash (39:18)
Yeah, that's true.

dan (39:20)
they're probably gonna get scammed. So it's it's it's a balance, but I think like, you know, there's these things like s you know, like secure chips and whatever where like it's you can't really like pull it out easily. but the

I think like if if Blue Sky h hosting did backups by default, like I think that would have been great. Like if it would even like to actually to to just do my phone, right? Like if my if my phone had my like a rotational key that overrides blue sky's one, you know, the one that's on my hosting, and it made like regular backups, I think that that would have been you know, that would have made the network a lot more resilient.

Justin Garrison (39:58)
Yeah. I and I do think that like even my

PDS, like I I went through and I uploaded a bunch of my short videos when Blue Sky launched minute videos. And I was like, cool, I have a bunch of these. Let me just upload them and see what happens to my PDS. And like it did the you know, blobs aren't even stored there. So like blobs are still an external thing. So I'm like, these are just the te all the JSON, right? Like this is nothing. It's like you know, a few megs of information. and and it's super easy to actually back that up around and keep a copy of it as long as you can, you know, keep the key

dan (40:25)
Mm-hmm.

Justin Garrison (40:26)
secure and and you still have access to the blobs, which is something that I

think again it'd be nice to have like some way to back it up on blue sky as well. I know it's I think it's like S3 under the hood, but that's the data that people really care about. And and the thing

dan (40:40)
Mm-hmm.

Justin Garrison (40:40)
that Macedon fails at is this universal identity. And I know they're they're working on it, but it I don't as far as I know, no one's implemented, it's not there yet. But like fundamentally Blue Sky was saying, okay, we're gonna start with identity and who you are, which is one of my first like

big blue sky videos, I guess, was like how to set up the whole like DNS identity of like, look, your website is you.

dan (40:58)
Mm-hmm, mm-hmm.

Justin Garrison (41:00)
and in that PLC the D I D can point to a a you know a DNS entry. And and that was fascinating, but also I I realized like how many people that that locked out.

of you know, like

dan (41:10)
Mm-hmm.

Justin Garrison (41:11)
name squatting and stuff like that. Like it's really cool that I can just reuse JustinGarrison.com, but it and but it sucks that like anyone that never bought a domain, like that they're they just get, you know, whatever username they can they can grab.

Autumn Nash (41:23)
I almost think this is like everybody keeps talking about like the bubble and how everybody's doing the same thing and like at protocol to me is exciting because I think that we have like with just everything going on in the world and people talking about data sovereignty and like the distrust of all the billionaires with s like social media, it's almost like

People are gonna be looking more and more for new options and I think it's going to open up a realm of like businesses and tools and startups to start to address these issues and maybe or hopefully, like espe like new like open source projects and just different like I think they ha isn't there like a black sky too and

dan (42:04)
Mm-hmm.

Autumn Nash (42:04)
like a bunch of other social media like you know what I mean? Like I think it's

Justin Garrison (42:05)
There's a Euro sky, there's black sky, there's yeah, there's a bunch of those

Autumn Nash (42:09)
at least it's like

of found a foundation of something that we can start doing differently and not having to completely start again. Like do you remember when people were copying their Twitter tweets that were their like, you know, what they felt were their best Twitter tweets and then just like mass posting on Blue Sky. Like, I feel like even where we are right now is better than that.

Justin Garrison (42:29)
And I look back at like

From the early teens, right? This I forget what the lawsuit was, but like they provided big providers had to allow an export functionality of your data, right? And that's why like Facebook and Twitter and Google and all these they all have like take out or export export your data. I did that for LinkedIn, right? There was a new LinkedIn built on at proto,

dan (42:47)
Mm-hmm.

Justin Garrison (42:48)
and it's like go to LinkedIn, request your backup, and then we just imp import it and it just it doesn't do everything, but it'll get like your jobs and that sort of stuff. And I'm like, that's really cool because even if it's not in one place already, I can consume some.

of that LinkedIn information and export it and then put it in a place that I do control and I do own a little bit more. Yeah.

Autumn Nash (43:05)
Okay, I have a question for both of you though. Did you see the

new LinkedIn like report AI slop like button?

Justin Garrison (43:12)
Well, I mean, they should have just made the slo button the default button. They should've report this is not slop, right? It's not enough.

Autumn Nash (43:16)
Okay, but what do you think about it? Like of all over like

Justin Garrison (43:19)
Like this is ridiculous. Like everything on LinkedIn is slop. Like

Autumn Nash (43:23)
Do you really like

dan (43:23)
Is it is it

real? Like I I I haven't seen it.

Justin Garrison (43:25)
Yeah.

Yeah yeah.

Autumn Nash (43:26)
So you can

actually go to like the post, like you know how you have the three little dots at the top of a post

dan (43:30)
Yeah, yeah, yeah.

Autumn Nash (43:31)
and actually like do like this is AI slop and I just think it's interesting because it it just shows like how much all of this is evolving. You know, like there's been like think about if you look at like how we used AIM and like RSS feeds and chat rooms and then you think about how it grew into like MySpace and Facebook and like we're still kind of like I feel like we're even in a a

almost a more aggressive like cycle of like change and like dissatisfaction, but like looking for the next thing, you know, and the AI has made that more like complicated. Like want knowing what you see and read is real or being like just the fatigue of so much of it. Like I just think it's interesting that that's where we're at, that that is a button now, you know.

Justin Garrison (44:14)
I guess that that brings me to like an integration that App Proto has that I think is is not well it was advertised a lot initially and it's not well utilized today is labelers. And that's essentially what you're doing on this LinkedIn, right? It's like manual labeling to say like, hey, this this sucks, right? Like I'd get this out of here. And Dan, do you subscribe to any labelers?

dan (44:33)
I think I used to like not not anymore. I kinda think like from the product perspective they feel very bolted on. so I'm you know, maybe if they redesign it a little bit.

Justin Garrison (44:44)
The one the the few that I still like are there's one for like Twitter screenshots, right? Like, hey, we labeled this as a Twitter screenshot, so we just like hide the image for you. And that's like a default that like I subscribe to the labeler, so now all those images are hidden for me by default. And I was like, cool. Like a lot of times it sucks because it mislabels things, but I like say, Hey, this isn't labeled correctly. This isn't a a Twitter screenshots, let this one through. And so I I feel like

that was a a cool thing for like three months that people were trying to run these labelers

dan (45:15)
Mm.

Justin Garrison (45:15)
and figure them out and say like, how would this actually work? be able to label AI Slop immediately.

Autumn Nash (45:20)
I thought the

concept was interesting, but what really resonated with me with labeling was opinions. Right? Like it was very opinion heavy. And it I felt like people did nothing but argue whether something was racist or whatever or and it became more about like it being a way to like

I don't even know. Like it just people were fighting more of the opinions of the labels than the actual like using labeling and then I felt like nobody just talked about it anymore and it just went away.

Justin Garrison (45:50)
Yeah, I feel

like just faded away. And it was something that I I think would be a really there there may be a AI slop labeler that I don't know about. but that would be interesting to say, like, hey, this is

Autumn Nash (45:58)
I miss the fun

labels too, like when Barry put beans all over everything. Like we haven't had a good health re health red. Like they're like the world is so like

Justin Garrison (46:06)
They need to like un

unbreak health threads so that health thread can be a thing again.

Autumn Nash (46:10)
I

feel like it's like the internet's like so serious right now and like it is serious and a lot of it like we need to like pay attention to things but like the whimsy of the internet is like really lacking at the moment.

Justin Garrison (46:21)
And that's still my fundamental thesis of like social media online is like people gravitate to where people are having fun. And if you're not having fun, like no one wants to be there. And and just being able to like show people you're having fun on the internets is like a big thesis for just like getting people to join something. And and Blue Sky has been the most fun for me, I'd say, over the last couple of years.

Autumn Nash (46:42)
Yeah. Also the community though,

and I'm not investing a bunch of time into something else that is like getting acquired in five seconds. You know what I mean? Like

Justin Garrison (46:48)
Yeah. And and the thing that that

every social media sells is their moderation, right? Like the moderation is the key

Autumn Nash (46:54)
And that was a huge

difference between Mastodon and Blue Sky. Like it Mastodon got really like dark for a minute there. And Blue Sky was so much better at it. And even if maybe like they weren't pros at it, I think it just it was it was easier to sort through in my opinion.

Justin Garrison (47:13)
It at least had a moderation policy. Like every instance of Mastodon is like volunteers might

Autumn Nash (47:17)
And that's not a knock on Macedon. Like it's

not like I don't think they technically did anything wrong. I just think that like humans are not always well behaved and it's a lot to to try to like how do you police another person's server? You know what I mean? Like then you they had like a lot of times they were blocking whole servers and it just wasn't set up for that. It w

Justin Garrison (47:40)
I mean 'cause that's the only

lever they had, right? The lever was ban the server.

Autumn Nash (47:44)
That's what I'm saying. Like I feel like it was great if everybody behaved, but once you like it's like a lot of things, like kind of I just think of what Dan was saying, like these you have to think about the feature and then you have to think about how the feature is gonna be used wrong, right? So like a lot of ways that you don't want to be blocked on something. Sometimes there's people that you have to block, you know what I mean? And I think Twitter teaches us like what happens when like people need to be blocked. So you don't want to make it where it's

completely impossible to block people, but you also want people to be able to take their data and not start all over again. It's it's complex because there's humans involved.

Justin Garrison (48:17)
I try to think of that like that bring your data with you. like things I've had in the past that I've lost data from. Right? Like I've had I had I had my very first blog got it was WordPress and got hacked. And and I had a you know I could back up the files and I backed up the database, but I never trusted restoring it because it was hacked and I like, I don't know if it's gonna come back and and so I it eventually just like pulled out some of the posts that I liked and like rewrote them in the new platform. And over time that

taught me to like run run things that can't get hacked, kind of. And and the only things that exist on the internet that I know of that can't get hacked are read-only and they're like static file sites. Right? It's just like, hey, there's nothing that you can send here. Like everything is client side and and this is just a bunch of static files. And it's been one of the things that it was one of the first things that I tried to do with Mastodon was I said, can I run this as a static site or not? Can I run an instance that's static? And I got it down to like kind of I could I could make make three files that like appeared to be three

JSON files that appeared to be a Mastodon instance. And I can't do that with Blue Sky. And one of my worries is that there's going to be some vulnerability in a PDS somewhere that is going to let someone impersonate me, take over my account, add add data they shouldn't be able to, whatever. And not just a worry for me, but for anyone that's self-hosting PDSs, right? And then like all these instances are exposed, just like WordPress and like maybe.

eighty percent of them patch right away, but then the last twenty percent just become hacked and they always exist because they never update them and they just eventually go away because like people don't want to deal with it. How do we like that's the other side of maintenance to me that like is worrisome is I could still lose data. Like I still do have a backup. I I can restore my identity, right? It's in my one password, but like there is a series of events where I would lose everything here. Do

Autumn Nash (50:01)
I think I've

learned simplicity though.

dan (50:02)
A

Autumn Nash (50:04)
Like

Justin Garrison (50:04)
So

how so?

Autumn Nash (50:06)
I think like I've really fell back in love with like just writing and finding people's websites and RSS feeds, like to be honest. Like I feel like half of my old Facebook I wish we could lose. Like it's embarrassing.

Justin Garrison (50:17)
Hehehe.

Autumn Nash (50:18)
Like and I feel like there's a lot of times where like I don't know, take what you need with you, like take your writing or like I mean you can put a blog on GitHub and, you know, copy it on a website and just being simple and just I don't know, like

I think you're a lot of what you say and write and all of that like precedes you, you know?

Justin Garrison (50:39)
Mm-hmm.

Autumn Nash (50:40)
And maybe we don't have to do like so many extra things, like just keeping it simple with a regular website and hosting it and backing it up and just kind of like learning out loud in a way.

Justin Garrison (50:52)
And I that's still the number one thing that I wish that Blue Sky would do was allow me to hide likes and reposts on my own posts and other people's posts. Because like those are things that like I they it just drives the a different incentive of

I want to be on this platform for various things and it it like it drives that engagement bait of look at all the you know, in in in plenty of other social media sites have have faked those things or like Facebook had a huge lawsuit where they were like lying to people about how many views their videos got and they lost the lots of like, Yeah, we lied. It was it was us the whole time. We were just lying to get engagements and I think that that is something that those are the kinds of things that I wish like I

I want this to exist. I want it to exist longer term and I want it to be healthier for people to not care about driving someone else's business value going up, but just like worry about your own community and your own self and how

Autumn Nash (51:42)
But I kinda like reposts though. I think they're fun.

Justin Garrison (51:44)
No, I'm I'm I'm fine having them. I just don't want to see the number.

Right. Like just the just

the number alone of when I look at my post and I say, cool, that one got ten likes. Like that must have been a good one, right? Like I'm I'm on fire now. but just like, hey, you can leave the buttons there, but I don't want a number next to it. I want to just hide that completely. So it's just like I can still do it and I can still engage with things that I like, but I don't need any peer pressure of like, wow, this thing got like three thousand likes. Like, is that something that I like too? I don't know, probably. And I'll just click it, right?

Autumn Nash (52:13)
What do you think, Dan?

dan (52:14)
I I don't know. I mean I like my perspective on this is that it's really apt to application developers and the thing that a proto does that is novel is that anyone can be an application developer for an existing app, right? Like you can fork a product with all of its users in it.

Justin Garrison (52:34)
Mm-hmm.

dan (52:35)
like for example, like Eurosky, they're working on their own fork of the Blue Sky app.

I think currently it uses the Blue Sky server, so it's kind of like you know, just a risk in, like it's just a client. But like in theory, they could also fork the server, like Black Sky did. Like Black Sky actually has their own server. Like there it's not even a fork, like they actually like wrote it and you know, it's like a different code base. Yeah,

Justin Garrison (52:55)
Yeah. It's in Rust, right? Yeah.

dan (52:58)
yeah, yeah. And so you can kind of fork products, right? Like that's the new thing. And so if you you know, if you go to Eurosky and tell them, you know, maybe

Like i because they they they made their their client is called Mew that social, I think. And you know, you can come to them and say, it would be cool if you could hide likes or comments or whatever, and maybe they'll do it. Like up limitedly, like you can do it, right? You can make your own client for like Blue Sky server, or you can make your own server and you know, it's still gonna dis not like in Mastodon where like you make a server as like a separate thing. No, like you can make a thing that ingests everyone's data and offers a different

way to look at it. So there's kind of like a marketplace where people like you who want something different actually have an option to either build it or to, you know, to ask somebody to build it and to vote with their feet and to start using it. So that is like I don't think it's yeah,

Justin Garrison (53:48)
Yeah, my app view could hide it, huh?

dan (53:50)
like I don't think it's you know, it's it's like you need to pressure the Blue Sky app to be the app for everyone. It's more like now we have a way for everyone who, you know

Who differs enough in opinions about what the product should be to make a different product that still works with the same network?

Autumn Nash (54:08)
And you know what's cool about that? If you have a different way that you want to consume those things and you're on two different apps, because the data can pick up those, maybe those two people can still interact, but in ways that make them kind of like find that joy. Well the people that want to know their engagement, but the people that don't want to know their engagement, but you can still access the same information and create community.

dan (54:28)
Yeah, and there's actually an example. There is an app called Anisota. I think it's it's like a moth. It's the name of the moth. but there's an app called Anisota that is a blue sky client that is kind of a very weird client where I think it actually doesn't I d I don't know if it shows metrics, maybe it doesn't. Like it presents instead of a scroll, it's like it presents posts as cards.

You have like stamina that you spend by looking at them or something like this. Like I don't remember the whole system, but I think it's a good example of presenting the same information in a different way. And I think Anisota even has Anisota only posts. So posts that you know, they would not appear on Blue Sky. but Blue Sky posts do appear in Anisota. So it's you know, it's kind of like this multiverse of products where like

product you know, data can bleed between products in and it's up to the developers to kind of decide how that works.

Autumn Nash (55:20)
I really like that because like sometimes like all your friends will leave Facebook or Instagram or whatever, then you have to be on all of these different places to try to like see different friends even though maybe one might not be great for your mental health or maybe it's something you don't wanna consume. I love the idea of being able to pick what's good for you, but still like find the people where they're at, you know, and find people that have that same interest or you wanna catch up with, I don't know, your friend from like the second grade, you know? But

being able to know what you're exposing yourself to and kind of like what works for you.

Justin Garrison (55:52)
All right, my real last question now. Do you remember Dan? what was it called? Open ID. I think OpenID was the spec that we had back in the day

dan (55:58)
Mm-hmm.

Justin Garrison (55:59)
that would allow us to like get away from Google and Facebook as our as our identity provider and it was a portable identity. if anyone else is familiar with that, why do you think that failed versus what Blue Sky's doing with identity?

dan (56:10)
So this was a bit before I understood what the internet is. So I think I don't actually have like a clear idea of maybe i th does that build on OAuth or was that before OAuth?

Justin Garrison (56:22)
I don't even remember. I just remember

dan (56:24)
Okay.

Justin Garrison (56:24)
I implemented it on my WordPress. Was like cool, like I

dan (56:27)
But I think

Justin Garrison (56:27)
will never need an a new identity.

dan (56:29)
'Cause I think like a part of it is that you maybe had to run some software on your own site. Like d does it does it require you to do s yeah. So I think maybe that is

Justin Garrison (56:35)
Yeah. At at minimum you had to have

some files in place, but I think there might have needed to there had to be like some process there. I mean similar to like a PDS, right? Like 'cause I just signed into Anasoda, which hit my PDS, right? I have to have a server running somewhere that knows my identity to say, Yeah, you can go through this auth flow.

dan (56:51)
Yeah. But it's a it's a bit different

though, because in Atlant, identity is one layer above the data, right? It's it's like identity is like the thing that is the most decoupled from everything, right? Like it's it's persistent, but you can change you know it you can change where your hosting is. So identity is like one layer above all of that. Like it doesn't depend on your PTS. So you don't need to have a you know like a custom PDS.

Autumn Nash (57:13)
I also

Justin Garrison (57:14)
Wait a minute, but if my PD

my PDS I can't log into Blue Sky if my PDS is down. I guess blue Blue Sky specifically would require that.

dan (57:22)
Right, but your identity is essentially a JSON document that says where your hosting is. So if your host end is just completely down, you can redirect it to another host end. So identity is like one layer it doesn't live inside of your host and it's like it's like a thing that has a pointer to your hosting. And that's actually what makes you you know, what makes it possible to change your hosting over time. Because your host end doesn't really own your identity. It's like what the identity points at.

Justin Garrison (57:50)
I could have a layer of auth authentication or identity in front of my PDS or separate from my PDS.

dan (57:56)
N no. So the the PDS is the PDS is like the implementation, right? Like it's it's like the the thing that you log into. But which PDS your id like you have identity which is your ID, you know, and your JSON that says which PDS is currently serving you. And so if your PDS is like you're unhappy with it, you can point your identity at another PDS. And then now this PDS will be the thing that you log into.

Justin Garrison (58:22)
Mm-hmm.

dan (58:23)
Right, so it's just identity is just like a pointer that you can change over time.

Justin Garrison (58:26)
Right. The redirect

for where where to go to find it.

dan (58:29)
Yeah, yeah.

Autumn Nash (58:29)
But also I

don't think this is purely technical. I think it's also like the environment in which Blue Sky was like birthed from, right? Like I think Elon did a lot to make to make people like they had a whole Elon effect and like look at how many people like joined Blue Sky when he got rid of Brazil. Like I'm not saying that Blue Sky is only successful because of Twitter, but I think a lot of mistrust with Facebook, Instagram, Twitter and just

a lot of what people were experiencing sold them on app protocol or just something different, anything different, you know? It's great motivation to learn something new and to get out of your comfort zone when people are just ready to throw the whole thing away.

dan (59:10)
Yeah, it's it's very nuanced, I think, because it's you really can't like you know, we don't have like an alternative timeline to compare to. And I think it's it's such a you know, there there's just so many contradictions here that in some way I think in some way Elon did kind of create a lot of incentives for people to push people towards open social solutions, right? Like an open

These like open protocols, just by kind of showing by example what what can go wrong. Right. So in a sense, he kind of actually put a lot of you know enthusiasm into this movement by kind of opposing it. But on the other hand, he also like if if not like if that didn't happen, then like originally it's Prado was not supposed to be

You know, it it was not like there was no blue sky, right? Originally Prado was a proof of concept of moving Twitter itself into open web, right? Like that was that

Justin Garrison (1:00:07)
Mm-hmm.

dan (1:00:08)
was kind of the idea. And so it would be Twitter that would have been on Eprato had he not intervened, right? Which is like such a different future because then it's a mainstream like it's an already mainstream product built on it. But then the question is like would incentives really

line up, right? Because like Jay was insistent on doing it as a separate company so that Twitter doesn't really have control over it and like she ended up being right that that was the right choice. And so it's like who you know, like what is is the pressure that Elon created by being Elon like has it kind of propelled the protocol into the spotlight or has it hurt it by making it like a fringe thing that only people who are like unhappy with Elon go to?

And I think there's like a second order kind of effect there that if you look at the kind of people that went to I mean, sure, it's it's like bro strokes, but it's mostly people who are unhappy for some reason. So it actually created this like negative, you know, this like perception of like blue skies full of negative people who are like angry about everything and so on, because it's it's just the most critical segment of the mainstream platforms that has moved there. And so I think that in a sense that actually hurt the protocol because they

You know, that created this like cloud over a blue sky of like, blue sky is just like angry lefty people or something like this. And then it's it's actually and like a lot of them, in a sense it's the success of the protocol that a lot of them don't care about the protocol, right? Like a lot of people say, I don't care about this if protocol bullshit, like I'm just you know, I just want Twitter with better moderation and so on. And it is a it is a success of the protocol that

people who don't care about the protocol actually end up kind of using it without knowing it, but it's also it also kinda hurts it because it means that there is no you know, it doesn't create like momentum around the protocol itself. So it's such a bowl of contradictions, I think. It's it's very hard to say like how things could have turned out.

Autumn Nash (1:02:03)
I think app protocol has the same issue as S R E's do. Like when it's doing its job you don't know it's doing it great. You know? So it's like

Justin Garrison (1:02:10)
No yeah.

dan (1:02:09)
Exactly. Exactly.

Autumn Nash (1:02:11)
because it's been so good, like it's

it's meant to be invisible, you know? But also I think a

dan (1:02:17)
Yes.

Autumn Nash (1:02:17)
lot of like so I think for a long time it was that negativity of people that didn't like Twitter. But I wonder also if like, you know, like there's a lot of discussion in how people are you know, data is being harvested. It's a lot of discussion into like flock cameras and like just all

dan (1:02:32)
Mm-hmm.

Autumn Nash (1:02:33)
of these kind of like data and like living in a like security state and how people are using your data. And I think the more that mistrust grows, the more people are looking for other alternatives and what

whether that be at protocol or anything. But I do think we're in a point where and even just look at the laws that people are making and data sovereignty or just the way that like the government got involved with TikTok. Like I think we're in a stage where like our laws are not keeping up with it and people are really starting to be more and more concerned and realizing how like technology works and if they want to be a part of like technology. So I think that so much is changing that I wonder if the

the focus on data and what we use and how we consume will benefit blue s like the protocol in general, not just blue sky, but just, you know, finding different ways to interact with the world that you feel comfortable with.

Justin Garrison (1:03:20)
And with all

dan (1:03:20)
I they

Justin Garrison (1:03:21)
go ahead.

dan (1:03:23)
Yeah, I just want to say I think like the contradiction there is that F prato like is very kind of scraper first mentality where it's actually like if it's public, it's equally public to everyone. Right? Like that is that is the stance, which I think is kind of opposite from how a lot of people, you know, this like your data is private. It's more like it's it's not you know, one company is not special. You know, it's not special because they post into it.

So I think that that is maybe like a difference, but they're also working on extending the protocol for permissioned data.

Justin Garrison (1:03:54)
Yeah.

dan (1:03:54)
So for things that are, you know, private only that only the apps that you give access to see it and only like, you know, the scope that you kind of give access. So I think that is an exciting kind of development that they're trying to finish like this year. So we'll see if that is also like another, you know, it's like another blocker to adoption is that you can't do permission data on the protocol and

We'll see what happens when you can.

Autumn Nash (1:04:19)
I think even though like the protocol is like public things being public, I think it's just the fact that people are more interested in un in learning about what's public, what's not public and all of that. And then the fact

dan (1:04:29)
Mm-hmm.

Autumn Nash (1:04:30)
that the protocol is like so is taking so many like contributions and is w it's willing to bend and kind of listen to what people want makes it like a good possibility, you know.

dan (1:04:42)
Mm-hmm.

Justin Garrison (1:04:44)
Yeah. Well, Dan, thank you so much for coming on the show and talking about all that with us. if you want to follow Dan on Blue Sky, the FAFO.fm account, which we do have the domain as as our handle, has a starter pack of all of the guests we've had on the show. And and Dan, I just added you, so you're in there now too. So anyone that is coming to

blue sky for the first time, you can sign up for the starter pack that's in the show notes. and you will get everyone that has been on this show. Or everyone that has been on this show that has a Blue Sky account, which is not everyone, unfortunately. and then yeah, you can go from there, follow people and s and start owning your own PDS. I do have some videos on on my personal YouTube on like how to run a PDS. I was running on a Raspberry Pi at my house for a little while and I I learned just how unreliable my uplink

internet connection was and so I moved that to a hosting provider, but the the rules still apply and if you want to migrate to a PDS, I also have a video for I'll throw that in the show notes just because I think it's fascinating. I do think that more people should own their data no matter what they're doing. whether that's whether the network is is billionaire proof or not, just having a backup and being able to say this is my data and I can move it around wherever I want, I I think is important. And that's what something I like to promote.

Autumn Nash (1:05:54)
And I also think that app protocol is one of the few things that feel accessible that people can try something like that for the first time. Maybe it works, maybe it doesn't work, but it's such a cool project and something to like try, you know.

Justin Garrison (1:06:06)
Mm-hmm. So thank you, Dan, and thank you everyone for listening to the show, and we will talk to you again soon.

dan (1:06:13)
Thank you for inviting me.