Citadel Dispatch

Share on social media:

support dispatch: https://citadeldispatch.com/donate  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠
EPISODE: 122
BLOCK: 836527
PRICE: 1448 sats per dollar
TOPICS: relays, clients, censorship resistance

FIATJAF: https://primal.net/p/npub180cvv07tjdrrgpa0j7j7tmnyl2yr6yr7l8j4s3evf6u64th6gkwsyjh6w6
HODLBOD: https://primal.net/p/npub1jlrs53pkdfjnts29kveljul2sm0actt6n8dxrrzqcersttvcuv3qdjynqn
PABLO: https://primal.net/p/npub1l2vyh47mk2p0qlsku7hg0vn29faehy9hy34ygaclpn66ukqp3afqutajft
MIKE DILGER: https://primal.net/p/npub1acg6thl5psv62405rljzkj8spesceyfz2c32udakc2ak0dmvfeyse9p35c

new to nostr? try ⁠https://primal.net⁠

nostr live chat: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://citadeldispatch.com/stream⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠
nostr account: ⁠https://primal.net/odell⁠
youtube: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.youtube.com/@citadeldispatch⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠
podcast: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://citadeldispatch.com
stream sats to the show: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.fountain.fm/

(00:00:02) Intro Clip: The central planning of money and its impact on society

(00:39:14) How Blastr works and its impact on censorship

(00:39:33) The need for spicier content on Nostr

(00:42:47) The challenges of relay selection and monetization

(01:17:48) Discussion about the chat and its functionality

(01:18:15) Proposal to put something generically out there to gather more information

(01:20:07) Discussion about proxies and their potential benefits

(01:53:48) Introduction and gratitude for audience support

(01:54:11) Ways to support the show

(01:54:34) Appreciation for audience engagement

(01:55:08) Future plans and hopes for the show

[00:00:03] Jack Mallers:

With central planning money, which is such a core technology to society working. Think about it, Ed. You go to work every single day, and you pour your blood, sweat, and tears all of the time you spend on this planet in exchange for what?   Money.   So the central planning of money convulates and complicates   so much of the inner workings of society. Let's go back to basics. Our government is in debt. Right? Yes or no? The US government. That is what all governments. Right? Our government's in debt. Okay. Traditionally, Ed, if I owed you $20,   I'd have 2 options. I'd wanna have to default on that and say, you know what, Ed? I hope you still consider me a good friend, but I'm not gonna make whole on that $20.  

The other is I could pay it back. Those are classically the 2 options that anyone in debt has. Right? Now the government, because they centrally plan and control our currency,   unfortunately, has a third, and that's that they can print more money, devalue   the debt that they have and that they owe, and allocate more capital to themselves. So our government   can't default. The US, the United States of America cannot default on debt. It would collapse the entire planet.   We also cannot afford to pay it back. So if you just like, listen. I didn't even go to college, brother. This is just 101 basics how the world works. If we can't default and we can't pay it back, what's the only option that they have to do? No matter what they sit and tell you at the Fed chair meetings and all of the economists,   they have to issue more dollars. And so if there's going to be more pieces of green paper, you want them competing for the most fixed thing. Right? There's more dollars that are competing for a fixed amount of Bitcoin. And Jack. Yes. Real estate's gonna go up too because there's more dollars competing for real estate, but they can make more real estate. They can find more gold. They can't make any more Bitcoins. And that's just I mean, even a college dropout can understand that, my friend.  



[00:02:30] ODELL:

Happy Bitcoin Wednesday, freaks. It's your host, Odell, here for another Citadel Dispatch,   the interactive live show focused on actionable   Bitcoin and Freedom Tech discussion.   That intro was rider die freak and good friend,   Jack Mallers,   on Bloomberg.   Duncan on the host telling him he dropped out of college and still understands the system.   I have a great, great crew,   joining me for dispatch today.   The the focus of the conversation will be Noster.   I have a subset of the Noster Illuminati   here on dispatch to discuss Noster development goals, the path forward.   We have   Mike Dilger of Gossip,   the Gossip client. He was on   dispatch a couple weeks ago. How's it going, Mike?  

You're muted. We went over this earlier.  

[00:03:31] Fiatjaf:

His name is literally Mike.  

[00:03:33] Mike Dilger:

Is it I actually pressed the end button, but my mouse was in a different window. You're good. Is it is it other than that and I'm I've got my coffee, so I'll try to wake up and be a bit more alert for this. Mike, we did a full rip,  

[00:03:47] ODELL:

and I realized afterwards I didn't ask you. Is it Dilger or Dilger?  

[00:03:52] Mike Dilger:

Depends what country you're in.   You know, the Germans will say Dilger, but we never said that because I grew up in America. So it's Dilger.  

[00:03:59] ODELL:

Okay. Dilger it is. We have Fiat Jaf,   the satoshi of nostril. How's it going, Fiat Jaf?  

[00:04:07] Fiatjaf:

It's good.  

[00:04:08] ODELL:

Good to have you good to have you back on the show.   I think you've been you've been on the show many times. It's always good to have you.   We have Pablo who does many things in in Noster. How's it going, Pablo?   Very good. Nice to be here. Are you cool with that intro? You do a lot.  

[00:04:26] Pablo:

I do a lot. Yeah. That sounds good.  

[00:04:29] ODELL:

His his main well, one of his projects, definitely not his main project, is Wikifredia,   which is a Wikipedia competitor that's actually censorship resistant.   That's not just, like, controlled by Jimmy Wales.  

[00:04:42] Pablo:

Yeah. Which Gigi told me a couple of days ago that when he first heard it, he thought it was retarded and hated the idea   and that it took him, like, a week to realize how genius it actually is.  

[00:04:55] ODELL:

Would that he hated Wikifredia?  

[00:04:57] Pablo:

Yeah. The the concept of no canonical entry for Wikipedia,  

[00:05:00] ODELL:

that there is no one entry for anything. And it's also different. We live in a post truth world now.  

[00:05:07] Pablo:

Exactly.  

[00:05:08] ODELL:

Yeah. Pablo, you might be on your wrong mic. Tap your mic. Goddamn it. Yeah. You're on your wrong mic. Alright.   And we have Huddl bod, who I had the pleasure of meeting in person for the first time, in an Austin bathroom.   He is, the maintainer of Coracle,   the Nostril client. How's it going, Huddlbot?  

[00:05:28] Hodlbod:

Very well. Good morning to everyone except for Pablo.  

[00:05:31] ODELL:

Autobot, by the way, you have,   you have the title of being there's only 2 people I've I've ever met in a bathroom before is you and Craig Raw.  

[00:05:42] Hodlbod:

So Good. Well, I'm happy to be, be in good company, I guess. Yeah. There you go.  

[00:05:47] ODELL:

Okay, guys. Well, first of all, like, we plan to go pretty deep on this show.   If you've never used Noster or tried Noster,   you should still join us. You should   come along for the ride. It's good to hear technical discussions and and kinda learn through osmosis.   You'll figure it out eventually.   But if you've never used Nostra, consider checking it out.   It is   quite fascinating and quite powerful. And,   yeah, I mean, if I recommend a client here, like, I'll probably just get shit on. But if you go to the App Store and just search Primal   in your favorite App Store and download that, it's a good place to get started.  



[00:06:30] Hodlbod:

Company around bags, I see.  

[00:06:32] ODELL:

You know, I fund I fund 22 Noster clients. So,   yes, I am I am pumping my bags a little bit.   Would I mean, do you agree that, like, primal is the best place for a beginner to get started with Master, or am am I off base with that?  

[00:06:48] Hodlbod:

Yeah. That's probably fair.   The, the integrated lightning wallet doesn't.  

[00:06:55] ODELL:

Yeah. Because people need to feel the zaps in the beginning,   and Lightning kinda sucks   to use self custody for the first time. So it it does suck that that primal is using a custodial wallet,   that has some KYC attached to it. Specifically, you have to do email verification,   which I don't love,   But it's pretty smooth onboarding,   and it works for most people.  

[00:07:23] Pablo:

Given the state of things, it could be far worse than email.  

[00:07:27] ODELL:

Yeah. Exactly.   Yeah, Jeff, you you have an opinion on my client recommendation for beginners before we get more technical?  

[00:07:36] Fiatjaf:

No. Let's let's get more technical.  

[00:07:38] ODELL:

Okay. That's what the flame client flame wars. Yes. You don't wanna do client flame wars? Yeah. Okay.   There's many great clients. I do we have a repository for all the clients? Is it nostr.org?   Does that does that list actual clients? Or nostr.net   has a giant list. Nostr.net. Everybody's busy. If you wanna overwhelm yourself, you can go there,   or you can open the App Store and just search Primal, get started, and just trial by fire.   Okay.   So we have a booming Nostril live chat. You can get to that at sill dispatch dotcom/stream.   That's powered by zap.stream.  

Shout out Kieran for maintaining that project. Awesome fucking project. You do not need a Nostril account to join us   there. You can create one on the website.   Okay, guys. Where do you wanna get started? I woke up in the morning, like, 5 days ago,   and I was, like, being   kind of, like, peer pressured into doing something by Fiat Jaffe. It seemed like there was, like, a little bit of protocol wars brewing.   What is this about? Like, why should people care?   And maybe that's a good spot to start.  

[00:08:49] Fiatjaf:

Okay. Everybody's silent. So what happened was answer.  

[00:08:52] ODELL:

Oh, I'm I supposed to do that?   No. Go for it.  

[00:08:59] Fiatjaf:

Well, what happened was that I did a small test that people would say was statistically   invalid, but I just posted   to my usual relays   that I just choose more relays that I post through.   And and I posted the exact same content to the big relays. And   in 5 like, 5 minutes, there was a dramatic difference in the amount of responses   between the 2. And then I did the test again with a comment on the same experiment,   and the results was the same.   Like, it was 10 times more   comments and likes and replies, etcetera,   on one versus the other.  

And I concluded,   that this was the problem with the,   like, with the censorship resistance properties of Nasr. If I was being censored   by the big relays and was only able to post to my   2 small relays,   and then my my post was not reaching the people that have voluntarily decided to follow me, and that was the problem.   And that, I think, sparked a lot of discussion.  

[00:10:11] ODELL:

So, basically,   Noster uses a system of relays. Right? Anyone can run a relay.   And if you right now, as Noster   Noster is pretty early in its adoption cycle, but right now, in Noster, if you if you publish your note if you broadcast your note to one of the the large relays,   you get more engagement.   That was the conclusion of your test.  

[00:10:37] Fiatjaf:

Yeah. Yeah. The idea is that, there is a much of people that are supposedly following me, but they're not getting my notes or they're not getting them, like, they're maybe they're getting them the relator, but doesn't matter.   And then it's kind of, like,   it's a kind of shadow bound, you could say,   because their clients are not connecting to to the small release, I'm I'm publishing my notes too.  

[00:11:04] Pablo:

Yeah. To to add a bit more color to it, the the way more   the way the way, pretty much all clients, especially, like, the big big,   clients, they all do it this way. You define a set of relays, and they connect to that that list of relays.   And then whenever you follow someone, it asks those relays give me everything that all these a 1,000 a 1,000 pub keys that I follow give me everything that you know about them.   But if you don't publish to to those relays, then if someone is following you, they're simply not going to see it.   So   one of the things that we've been discussing is what we used to call the the gossip protocol.  

We gossip means a bunch of things in a bunch of different contexts,   in lightning networks. It it means one thing.   So   we we are trying to change the name to to outbox model,   which is more descriptive of what we're doing. And the idea is you your pop key says I'm going to be writing on these relays or I'm going to be reading on these relays.   So if one of the main relay sends for you, it doesn't really matter because the clients will go to those relays and find the notes   that that you wrote.   So the the the difference is very stark. I I've been using   Deimos   and and a few other clients   with, a small set of relays, and I'm   constantly not seeing responses. Then I go to primal in primal because,   because primal is using the caching server.  

Even though I am using the,   I am using the same, the same set of relays, I am reading for from the caching relay, so I see all the responses that I would have missed.   But that only works because the caching service is indexing all these other relays.   But if   the the network fragments so, for example, probably FiatChefs pyramid relay, which has, like, 10 users,   is not being indexed. Right? So if I don't have that relay on my Well, he's probably indexing that, Milan.   It's a very,   I mean, it's a it's a it's a known relay, but it's still a very small relay. But the point stands that if I run a small relay, that really is not going to be indexed. If if I span a new relay and I just put it in my NIP 65 on my on my list of relays that I that I write to,   probably probably, primal is not going to index that. It's not going to be listening for all the relays that are published, and there's one person I mean, the bigger issue the bigger issue is that,  

[00:13:33] ODELL:

you know, the the primal caching server can omit posts. It's it's a centralized service. I mean, it's it's false. It's open source. Anyone can run a caching server right now. The Unity Wallet guys run a caching server as well. Presumably, there'll be more you could run your own, technically.   But I think the bigger issue with the primal caching server is that   it can lie to you and say that a note doesn't exist, that does exist, and and and you wouldn't know unless you check a different client or switch your caching server.   Mhmm. I mean, I know Milan's goal is to index, like, all of Nostr   and every relay and everything.  

I will say using primal during Fjotchoff's test was really annoying because   I saw all the posts. I saw all the notes.  

[00:14:18] Fiatjaf:

Well, the product, like, after 10 minutes, like, after maybe 20 minutes,   the notes on my small relay, they got broadcasted through all the big relays. So People were rebroadcasting them? Yeah. Yeah. So if if if you arrived on any client, if you arrived   a little later, a little after I posted, then you'll see both. And then you say my test is flawed, but in the beginning, you would not   see both. I don't know about about what happened on primal, but I know primal index is my small relay too. So   primal is not a is not a good parameter. Yeah. It does.  

[00:14:52] ODELL:

By the way, was I intentionally excluded from pyramid because you don't like the caps? Was that why I wasn't part of the original group of users?   No. You can be you can be invited, I think. I'm no. I'm in. I'm in now. Island got me in.  

[00:15:05] Fiatjaf:

Well, the idea is that people would invite other people, but I I kept inviting everybody.  

[00:15:11] Hodlbod:

So okay. Yeah. Yeah. The rebroadcast thing isn't isn't a totally irrelevant point either because, like, even if your notes are eventually making it to the big relays,   there's that time delay. Even if it's, like, 1 minute, 2 minutes,   Nostra feeds are so,   so biased towards recency   that the people who wanna see your stuff, are a lot less likely to see it because most most,   notes are seen at the top of the feed.  

[00:15:35] Mike Dilger:

Yeah. Right. So the ultimate concern is People are saying people are saying that Zap ZapStream died. Is that relevant? Is that something you need to fix? Did ZapStream die?  

[00:15:49] ODELL:

It wasn't it just wasn't updating?  

[00:15:52] Pablo:

Obviously, no. Yeah. It's it's comments. The stream ended. There is no stream on Substream. Why did the stream end?  

[00:15:59] ODELL:

Kieran,  

[00:16:01] Pablo:

fix it. This is the censorship we're talking about, Matt. So this is what we need to fix.  

[00:16:07] ODELL:

It says it's streaming on my side. Kieran's probably freaking out right now. The chat still works.   Continue to use the chat freaks,   while Kieran fix it. We're on YouTube. You know? Fuck Google, but  

[00:16:20] Fiatjaf:

it doesn't just There are other clients.   There are other clients that can see the the live stream. Really?  

[00:16:26] Pablo:

Amethyst? Yeah. I don't see it on Nostril though.  

[00:16:30] ODELL:

I think it all just goes through ZapStream.  

[00:16:32] Pablo:

No. No. No. Yeah. It's just not through events. So, like, on Amethyst, you can see it.  

[00:16:38] Fiatjaf:

There's this app stream server, which is different from the client,   but it's tightly integrated, but it's different. I don't know if you It's probably your streaming service  

[00:16:46] Hodlbod:

that's that's gone down.  

[00:16:48] Pablo:

I don't think so because we're still on YouTube. I think you just marked the, the stream as ended in in some way   Because I it looks like the event changed to to ended.  

[00:17:00] Hodlbod:

I think it watches the stream. So whatever's reeling the stream to, to ZapStream went down. But if YouTube's still working, then that's bullshit.  

[00:17:09] ODELL:

YouTube probably a some kind of fan out thing. Okay. Well, we're not gonna, like, troubleshoot this right now. Technical person to help us  

[00:17:16] Hodlbod:

with  

[00:17:18] ODELL:

this. Okay, guys. Let's let's continue the conversation.   The the live chat is still working. Hopefully, it comes back.   So so, I mean, the so Gossip is is a client that that does this,   Would you say does this the right way? I mean, we have Mike here.   Like, the whole point is it's it's supposed to read your your relay list and then   automatically check the relays that you're broadcasting to depending on on how you're, you know, viewing your feed.  

[00:17:49] Mike Dilger:

Yeah. The Gossip client is the first one to do it this way as far as I'm aware.   And   by the time   that   was sort of formalized   and we had niche 65 to say,   here's here's an e here's a better way to do what Gossip was doing, but   being more straightforward about it,   there were already some 20 or 30   nonster clients.   So   most of them had already developed   you know, we're working on other things, and we're we're not thinking about that problem.   And so that's where we're at today. We have all these developed clients   that, well, they already built themselves up in a certain way,   and to go back and change this now is, I think, a little bit a little bit trickier. There's a little bit,   hesitancy.  

So that's where I think it's, you know, where all of us are trying to, you know, push push to get people to move more towards that because as Fayette Jobs points out, you know, if we don't if you don't check where someone posts, you're not gonna get their posts or you're gonna have to copy the posts.   And then there's a you're relying on a copying service, which could be a sensor,   or you're relying on a proxy service, which could be a sensor.   And so,   yeah, it's just   but, yeah, the gossip client was   was where this started. That's why people were calling it the gossip model.   Yeah. Yeah. And now now there's quite a few clients that that implement this model. What other clients do?  



[00:19:22] ODELL:

By the way, I fixed ZapStream. I just ended the stream and restarted it.  

[00:19:27] Hodlbod:

Well, go ahead. Go ahead. Cool. Yep. I can speak for myself. And I I know NDK, you're putting you're adding support to it. Right, Pablo?  

[00:19:35] Pablo:

No. It had support for the past,   whatever, like, 6 months or something like that. I don't know. Yeah. Like, I don't know. For for a while,   I'm improving support and improving and improving   and everything after 2.1,   which was a while ago,   has outbox model by default.   So I have a list Loom has it too. Yeah. Loom. Loom uses NDK, so so do it does.  

[00:20:10] Mike Dilger:

Nozzle. I don't know if Nozzle is still around. Canvas app, is that still around? I don't know. Snort   uses it. Iris,   OpenVive is in the process.  

[00:20:21] Pablo:

So Nastur Hasid.  

[00:20:24] ODELL:

Horrible name. Nastur Hasid. Horrible name. I didn't know. Yeah. With a u. Nastur with a u.   One u. It should at least be 2 u's, but whatever.  

[00:20:33] Hodlbod:

The the interesting   it's kinda interesting though because, like, having the outbox model isn't really like a a Boolean switch.   Yeah. It's it's incredibly complicated to get right,   even if you're doing just the outbox part of it.   Yeah. I was reviewing my my implementation this morning because,   right now, feeds,   don't exactly   implement it because I have a cap on how many,   relay connections the client can open at any given time. And so Coracle actually weights towards, the biggest relays currently. The fuck?   Yeah. Sorry. I rugged you guys. I'm, I'm trying to fix it right now, but,   and then there's the whole, like, bootstrapping problem. So what relay do you connect to first on a fresh on a fresh, like, install of your client?  

And how do you find everyone's Relay selections in the first place? That's a whole whole other   thing that,   that clients have to do right or wrong or or figure out some way.  

[00:21:31] Mike Dilger:

Yeah. Even the gossip client isn't doing everything right. Like, I as we were discussing DIP 65, I I realized I'm not looking for replies   in the inbox of of the author of the event that I'm looking for replies in. I I don't I never read from an inbox. I only write to an inbox, unless it's my inbox.   But that ends up being a really great place to find those replies because if all the other clients are using the outbox model, they will they will be putting replies into that inbox so I could find them all there.   So now I have to go change my code. So I'm not even, you know, doing it all right in the gossip plan.  



[00:22:08] Fiatjaf:

I I don't think there's a kinda tweaking it a little bit. I don't think there's a correct way. I think there's, there are many ways to do it. There's   each client can have a different approach. You can have, like, to think big manual. You have to click a button to add a new relay or accept connection to a new relay, or you can do it automatically.   You can only fetch reply from relays that you you deem to be safe of spam or,   into the release that are not safe of spam. You restrict that to some accounts only or something like that.   You can   pick relays only from the NIP 60 like the NIP 65   list of relays that people publish, or you can use relay hits, which I think you should   even as a secondary fallback mechanism.  

Can you use other other ways to track, like, the the   to keep track of who where people are posting and how to connect?   You can do one thing that I think some clients are doing. It's like like you optimize   to only connect to some relays. You you like, you   you pick relays that a lot of people are publishing to, and then you connect to those instead of connecting to, like, a small relay that's just one person who publishes   as long as that person is covered.   You can do that or you cannot do that. You can use a different rate to select the relays when you're asking for post.   I don't think that's correct, man.  



[00:23:32] Pablo:

I, yeah, I agree. A few a few months ago, like, 6 months or so ago,   we had a bit of a discussion on the roster whether we should formalize   how to implement the outbox model.   And we realized that there it's more of an art than a science. It's it's something that we it's not settled on what's the right way. Like, for each of I don't think there is a right way. There is just a bunch of different approaches, and we are using the term outbox model or gossip   as a as a very loose term. It doesn't necessarily need to be the outbox itself.   It's like what the the   the discussion we're having is about   what's the smart way   of reading and writing from relays.  

And when we want to like client relay selection.  

[00:24:16] ODELL:

Like, how clients handle relay selection.  

[00:24:18] Pablo:

Yeah. Yeah. Yeah. Exactly. I mean Like, one one of the things he said right now, we are at the stage where, say, for example, you have I I used this example today on on Noster,   talking with, with Shavie 55.   And   imagine that you have 5 main relays and you are following Trump, and then Trump is banned from all those 5 relays.   And he moves to relay.troopsocial.com,   but nobody's read reading from relay.troopsocial.com.  

[00:24:49] ODELL:

But they would if you switched to it.  

[00:24:52] Fiatjaf:

Well, actually, Trump had what? Like, a 100,000,000 followers? In truth social, he has 6,000,000 followers. Yeah. But truth social is a fucking shit coin. It fucking sucks. Like, if you just have to go into settings and just, like, change which relays you're looking at, like, it's a way lower lift. But it's it's a very bad idea to do that because, like, you're going to connect to truth social, and then you ask for all your posts from all the Bitcoiners you follow from the truth social relay that has not no one   just has Trump. Like, this  

[00:25:20] ODELL:

is so bad. The client should be architected. Wait. So True. I mean, there's 2 things here. Right? Like, Mike There are many things. Mike Mike there's many things here.   And I saw pseudo Carlos,   commented like, oh, this feels like one of my work calls. I get this is like a nostril work call is what's going on here, and I hope you guys enjoy it.   We try and mix it up here on dispatch. I know I'm enjoying it. Okay.   I because I I need Noster, and I love Noster, and I I use Noster. And I I wanted to be as censorship resistant as possible because I know the first censorship battle is gonna be against all caps,   and we need this to be as as censorship resistant as possible before they try and stop my caps.  

The there's there's   like, Mike is building a desktop only client, which I'm I'm not gonna say is easy, but it's easier   than if people are coming on mobile. Right? Because, like, the big trade off here is, like, if if I'm hit if my mobile client is hitting a ton of relays,   like, it's gonna hurt it's gonna suck up data,   and it's going to   hurt my battery life. Right? Am I correct in this No. You're you're  

[00:26:25] Pablo:

you're not.   So the way the way NOS works is when you fetch events, if when you fetch notes from other people you follow is you literally send your whole follow list to each relay saying give me everything you know from this 1,000 puppies. Right? Okay. Follow-up a 1,000 puppies. So if you're connected to the my to the main 5 relays,   the main 5 relays are all going to have, most likely,   the the events for all 1,000 of those puppies. Right?   So you are essentially downloading if you're using 5 relays, you're downloading the same events   valid and 5,   5 times. Right? If you're connected to 10 relays, you're downloading the same event 10 times. Each event for all the people that you follow to, 10 times. In the outbox with the outbox outbox model, you will query the the relays where you know that these people are publishing. I just want Pablo from this relay.  

Exactly. But then you don't send Pablo to all the other 20 relays you're connected to. You said Pablo from this 1 or 2 relays. That's it.   So it's actually less data consumption.  

[00:27:32] Hodlbod:

The, the overhead that I think Vitor has pointed out is from establishing a bunch of connections to different relays, and that's like,   I guess, SSL,   encryption   negotiation.   And I don't know how true that is, but I trust Vidor. He he works on a mobile client and I don't. So I think that is substantial, but  

[00:27:53] Mike Dilger:

I'm I'm like what you said, Odell, I'm I'm lucky I'm on a desktop. I don't have to deal with this complex problem. But   it seems to me,   like like Pablo says, it's less data,   which is oh, you're getting it from potentially a lot more connections.   And so the only way that that could be more overhead is if the per connection overhead was high,   and which makes me think of SSL because obviously this is all over SSL.   But the the   the cost to your battery life of establishing an SSL session is really in the public key,   key exchange   stuff,   And the cost,   to your battery life of verifying an event is is, like, 3 times worse   with,   store signatures,   on on this curve that we're using. And so,   to me, it seems like if your client   on a phone is verifying events,   that's far more of your battery drain than the,   the connection overhead of SSL. And so even in this case, it seems that the app box model would be less battery drain   because you're not getting, you know, 101 the same event 10 times, and you're not verifying the same event 10 times.  



[00:29:06] Hodlbod:

Yeah. There there are ways around that. Like, so in Coreacle, I keep a cache of,   event hashes to signatures. I think I stole this from NDK, actually. So I don't I don't validate the same event multiple times.   But that's like that's an in the weeds thing. But, also, I wanna point out, like, this whole network topology thing is one conversation, but the optimizations,   there's that's a whole different conversation. So I'm I'm very much in favor of,   proxies,   and these can take all kinds of different forms. I I wrote one about a year ago. A number a number of them have been written. And the proxies can do a lot of this deduplication for you. And even if you trust it, it can,   check signatures   for you. If you really trust it, you can give it your private key, and it can sign things for you.  

We don't have to debate the trade offs with that. But,   you know, if you're talking to a single proxy that you trust and that proxy is swappable, you get all the censorship resistance, you get the network topology that you want, and you only have to establish a single network connection and download events a single time. So, like, you know, just keep that in the back of your mind when you're thinking about all this stuff.   Like, battery life and connections and validating signatures, these are not real reasons to not use the gossip or the the outbox model. I can't stop using the saying gossip. I'm sorry. Gossip's a better gossip's a better name than outbox.  

[00:30:28] Pablo:

It's a weird it's a horrible name, man, because it means something different things.  

[00:30:33] Hodlbod:

But why do I call it gossip? It knows who it is. It's Twitter model. Why did out Outbox get chosen as the name?  

[00:30:40] Pablo:

It it's more representative of of what's going on, really. But it's also not a great name.   It's really We need, like, a naming podcast on this space.  

[00:30:50] Mike Dilger:

I have I have 2 naming podcasts. It wasn't meant to be a marketing name.   It was meant to help describe describe what's going on to other developers.   Got it. Yep. People have is the only main People have an inbox and an outbox. If you wanna send something to them, you put it in their inbox. If they put it in their outbox, you can look there to see all the things they posted. That's basically it.  

[00:31:10] ODELL:

So the outbox is is my small relay that I'm running?  

[00:31:14] Mike Dilger:

Potentially.   Or there's someone else's running? You can have you can post to 10 relays. I wouldn't, but you can post   to, you know, multiple relays and that your outbox is actually multiple relays. Okay. So let's let's  

[00:31:27] ODELL:

let's I mean, let's start naming names because I feel like it you're, like, pussy footing around the names of of of clients and stuff, and and it makes it more difficult conversation.   So is the biggest concern here of, like, how Domus is handling things? Like, is Domus just   like it breaks if if you're not publishing   to big relays?  

[00:31:48] Hodlbod:

No.  

[00:31:49] Pablo:

Nothing is broken right now. From big relays.  

[00:31:52] Fiatjaf:

Go, Fiescher. Nothing is broken right now. We're just preparing for the future. Like,   if Nasr is not censorship resistance,   then I don't I don't see the point. Like, I don't I don't know what the the people that are not   supporting the outbox model,   why do they like Noster if   if Noster is not censorship resistant? Or why why did they came?   It's like the a bunch of developers writing clients.   I know, like,   JB   55 points is is that you   can manually go there and add a new relay whenever someone is censored. But I I don't think that I think the client must help you somehow to find out that someone is  

[00:32:34] ODELL:

automation because the users aren't gonna fucking do it.  

[00:32:37] Pablo:

Yeah. Yeah. Like, the default right now, the default is that when if someone is censored, if someone is not using,   the main relays or the relays that you are using,   they simply don't show up. Right? You follow them, and you think, oh, this person is never publishing and is never posting anything. It's like a shadow bunny. You they simply don't show up. You don't know that if something is happening.  

[00:32:57] Hodlbod:

But In theory.   But I think it's time   to to mention,   because of blaster,  

[00:33:03] Pablo:

it it does show up. Main or main? Ends up. Yeah. No. No. Not really. Not really. I I've been using I've been using small relays for the past 5 days. And   on on Deimos,   most of the responses don't have like, most of the responses don't have what they're responding to, I cannot see. And then I moved to primal, and then I see what the response was to. But it's not in theory. It's it's already happening. Like Yeah. And the primal is not a solution. Right? Once again. Primal is not a solution that scales. Yeah. It it's part of a solution. When it will scale, centralization  

[00:33:35] ODELL:

always scales.   Yeah. Yeah. It's, it is centralization, but making that indexer swappable would be would be It is swappable. It already is swappable. There you go. It's in settings. Yeah. But what I would like is a good step. What I would like is a good step. No. No. Well, the proprietary format is bad, but the the the architecture is is okay. Well, what I would like is for your your client to like, the primal client to automatically,   like, dub because the the big thing the caching server is trusted with is it can omit   notes that exist. Right? So I think there's something there to be said about, like,   doing a sample of your relays.  

And if you notice your caching serve like, if the client notices the caching server is lying to you about notes that exist, it automatically switches to someone else's caching server. Like, you have, like, extra caching servers, like, in settings,   and it'll just automatically flip over.   Does that help at all?  

[00:34:33] Hodlbod:

Well, there are different kinds of caching. Right? There's you've got caching, but then you also have indexing, and those are kinda 2 different things. Indexing is a particular cache, but it's meant to help you find the content in other places. Caching is is like a more more comprehensive scenario. And then there's stuff like search, which is a different interface for the same data, and that is a whole different set of trade offs. But, you know, like, the indexing part of it is really important. And I think the best solution we have for indexing so far is purple pages,   which Pablo made.  

And that just kinda looks around and grabs 10,000 twos and zeros and threes. Is that right, Pablo? That's right. Yep.   So you can pretty reliably,   look at Purple Pages to grab,   the list of relays that any given user posts to and then go there for their actual posts. And And that means that Purple Pages doesn't have to be super huge. It doesn't have to keep these giant indexes of Zap counts and all this stuff.   And, you know, the the primal caching server is just too thick,   and it doesn't need to be thick.  

[00:35:37] Mike Dilger:

Hey. Where does how does purple pages get its data?  

[00:35:41] Pablo:

It's just it's just indexing. I wrote an indexer,   but it's only getting those kinds and storing them. And for the longest time So it's it's copying events?   Yeah. Yeah. It's just copying events. Now a lot of people publish to them,   so it all it only accepts 03,110,002.   But, yeah, for the longest time, I was running it on a $5 VPS just to prove the point.   Then I realized that I was the only one that knew that it wasn't a $5 VPS, but it's a very, very But now we all know.   Now we now we all yeah. Trust me. But, I actually moved into my main relay because I was thinking I'm spending $5 for nothing here.  

[00:36:18] Mike Dilger:

No. I'm gonna change I'm gonna change my code to refresh relay lists way more often than so we can bring your VPS down a little. Wait. Wait. Let's just go.  

[00:36:28] ODELL:

So let's just bring it back to real world. And first of all, you know, is Pablo's relay a small relay? Like, I saw, like, Dorsey   switched his relay list, and then he added Pablo's relay. Like, is this all just a conspiracy to, like, get him over to your   relay? Like, why is your relay considered a small relay? I feel like that's a big relay.   I'll do profit. Yeah. I I just I just want to censor all caps of ale. It's not about profit. It's about censoring a billionaire. Like, it's there's there's it's enjoyable. That's the end game. Yeah. Yeah. That's definitely.  

[00:36:59] Hodlbod:

So No. I do wanna say that that big relays are not bad.   This is not about using only small relays. Big relays have,   plenty of utility.   They they do support some of those optimizations. So if you post to a big relay and you're not censored, then everyone can just connect to the one relay and get everything they need. That's fine, but it should be treated as an optimization, not as a default.   It it just needs to be possible   for clients to find people who have been banned from the big relays. Yeah. Yeah. Let's And that's not a heavy lift, really. Well, let I mean, let's yeah, go on. Let me put that slightly different. Mike.  

[00:37:34] Mike Dilger:

I'll put that a little bit different and say, you know, I don't actually think we should all be getting on the small relays. I don't think we should all be spreading out, and everybody's setting up their own small relays. Good.   I think we should all generally, you know I don't know how many 20, 30 relays is plenty. We've got a lot more than that now,   you know, because they're not all gonna get censored.   They might. When when something starts happening,   people are gonna have to jump the ship of those relays and go set up new ones.   And you need to have some process where that automatically, you just keep following them. And that's what this outbox does. So I mean, yeah, there's some It's it's it needs to be possible   to set up smaller guys, but what doesn't mean we should all be doing that right now?  



[00:38:18] Hodlbod:

There's some helpful back of the envelope math I did yesterday that kinda puts this this situation   in a little bit clearer terms. So if if if you have a 10% chance of being censored from any given relay, if you connect to 1 relay, then you have a 10% chance of being censored. If you connect to 2 relays and,   you have a 10% chance being censored from either one, you now have a 1% chance.   If it's 3, you have a point 1 percent chance. 4 is point 0 1% chance. So   you you,   small amounts of redundancy go a really long way towards censorship resistance. That math is very wrong because there's censorship.  

[00:38:53] Fiatjaf:

If you're a per a person that's likely to be banned from one relay, you're more likely be been from all the others too. So Yeah. Yeah. But, like, only one of those has to be one that you run or that you trust.  

[00:39:04] Hodlbod:

So, like and that's that's totally, like, ignoring any kind of relay selection heuristics. If you're naive and you just, like, choose whatever relays, that's that's, like, your worst case scenario.   So,   but, like, you know, the way blaster works is it's it's sending your notes to 100 of relays. And now you have a 10 to the negative 300th power chance of being,   censored, and no one needs that. Like, it's not worth,  

[00:39:31] ODELL:

you know, storing those notes in in so many duplicates. Well, I mean, I think the biggest issue here is, like, we don't have spicy enough content on Nostra yet. Like, people should be trying to push the envelope. Like, there's there's not much sensor worthy material out there yet.   And I feel like I don't wanna see anything worse than what I've seen. I I feel like you start we'll start to see like, it's all about incentives. Right? And so it's not only about incentives. Like, it's about,   like, open systems only really evolve and improve with pressure.   And it's it's we haven't really seen any kind of censorship pressure, and it's hard for me to imagine,   you know, making substantive changes or clients improving or user behavior improving until, like, people actually are getting   censored from major relays.  



[00:40:17] Fiatjaf:

Well, we don't have enough people we don't have enough people using also, but we should be prepared for when that happens. So,   no one's being censored. I'm not being censored by the big relays even though I choose to only post through the small relay.  

[00:40:32] ODELL:

Yeah. And then You need spicier content.  

[00:40:35] Fiatjaf:

What kind of spicier content? That's   I can't think of anything.  

[00:40:42] ODELL:

Wait. So let's talk about a real world example. Right? Real world example,   Will woke up hungover one day. It was like, holy shit. Like, this video of me on Noster is super embarrassing. Just deleted the whole Domus relay.   Biggest relay in the world. Just decided   decided you wanted to delete the relay. Were any notes lost?  

[00:41:02] Fiatjaf:

Probably not. I think there were some note notes lost. Yes.  

[00:41:06] Hodlbod:

I think so. I've had a harder time finding things at least.  

[00:41:09] Fiatjaf:

But that doesn't matter. It was only old notes that were Is this theoretical? Is this theoretical or did that that didn't happen? Right? No. He did. He did it in November, I think. October, something like that. Maybe he did do it recently as well. But  

[00:41:21] ODELL:

he literally just woke up in the morning and was, like, fuck this shit. Delete.   Just deleted the relay.   And I I didn't notice any notes lost. I mean,   if a note was lost and you didn't notice, did it actually get lost?  

[00:41:37] Hodlbod:

I can't find my first note anymore.  

[00:41:39] ODELL:

So that's what you should've used blaster. Maybe you should've posted it to as many relays as possible.  

[00:41:46] Mike Dilger:

I do have an archival relay. I don't know what else. People, even if there's no outbox model, most people, they're they're configuring more than one relay. And so the note if one relay goes down, doesn't get lost.   And then on top of that, you've got things like blaster   or even clients that copy   other people's notes from relay to relay,   that sort of replicate that stuff. And so it's kind of like we're all,   you know, it's kinda like we're riding bicycles with training wheels on.   And,   you know,   we we don't know who's actually able to ride a bicycle   because we can't I can't see if my software was working or not because,   you know, it just   the problems get patched   by this this event copying. I mean, it it it solves real problems. Right? But then we don't know the problems are there, so we can't fix them. So it'd be really nice if there wasn't any event copying because then we could easily see that your my client's not working, and and then we'd fix it.  

But I don't know how to make that world happen.  

[00:42:48] Hodlbod:

Yeah. Well, it's it's interesting because you can't stop blaster. Like,   unless relays choose to protect themselves,   by authenticating who's writing to them, then Blaster is just part of the   the situation.   And, like, I agree with Mike. It would be a lot easier to solve,   the these outbox model problems,   from the client side if we didn't have blaster.   So it would be nice if blaster would would cut it out for a while so we can feel the pain. Because you were saying, like, you know, you experience pressure and then you solve the problem. We're not experiencing the pain that exists as, like, a core part of the protocol. And I'm I'm pretty sure, like, Tony doesn't believe in Nostra as, like,   a censorship resistant long term solution. I think he said that in his blog post from last year when he joined. So, you know, does he does what they have?  

Does he? Oh, man. Well, that says everything, I guess.   But, like so he's not incentivized. He's not interested in solving this problem, which is fine. Like, you know, he he's, we need cats to knock things over.   So now we have to figure out how do we react to blaster? How do we how do we I don't think even blaster is the is the worst culprit. Like, I think people will naturally  

[00:43:59] Fiatjaf:

select the big rebate, like nos.lawanddemos,   nos. Wine, and then they will just publish to these and read from these, and this will happen naturally. And the network will centralize naturally without.   And if you're if you're not in these, then   then then you're not heard. And and people can also   rebroadcast manually.   Like, I think there were some bots that were broadcasting stuff. Like, the the the primal the primal   service has some kind of listener that listens to all the relays in existence.   And there are there were there were many other bots in the past that were not blaster. They were just broadcasting stuff.   So   it's not it's not Tony's fault.  



[00:44:41] ODELL:

Blame Tony. But I  

[00:44:44] Hodlbod:

Yeah. And and and, broadcasting has a lot of good use cases too. So, like, even in the NIP 65 spec, it says,   kind 10,002 should be broadcasted as aggressively as possible.   And most relays,   can afford to keep an index of of those, Even if there's 9,000,000,000   of them, like, that is not a huge relay. It's it's a fairly big relay to store 9,000,000,000 records, but it's like, that actually does scale,   reasonably well.   So broadcasting that kind of stuff so that you can find the actual content you're looking for, that's a great trade off. Well, but, like, I mean, first of all, let's unpack this really quick. Like, blaster is  

[00:45:23] ODELL:

is is this relay that's hosted at noster.munitywallet.com.   And if you broadcast to that, it broadcasts to every known relay.   And if if that relay will accept your notes, it accepts your notes, and now you're on that relay.   From a user perspective, like, once again, coming back to incentives. Right?   As a user   who is actually in like, wants to use Nostra as a censorship resistant communication medium, a way for me to connect directly to the people I care about,   without   without   being censored and being blocked and and and being stopped.   Isn't my incentive to broadcast to as many relays as possible? Like, I don't understand   like, I understand you guys trying to make a point about, you know, you should use small like, I I broadcast to a small relay, relay dot bitcoinpark.com. Like, only members of Bitcoin Park   can write to that relay. But then I also broadcast to any fucking relay that will accept my notes. And if they don't wanna accept my notes, they shouldn't accept my notes. But if they're gonna accept my notes, like, I want that seed to spread as far and wide as possible.  



[00:46:31] Hodlbod:

Yeah. Exactly. It's it it doesn't seem   You want that seed time person to get here. It's sorry. Go ahead. Oh, really? You're gonna just do that to me on my own show? I'll explain it. I'll explain it. But go ahead, Pablo.   No. No. No. You go ahead. Explain it. Okay. Yeah. So it's it's it's high time preference behavior. You're you're taking advantage of free resources that other people are providing,   to serve your own, your own interests. And, like, yeah, those resources are there. They're not protecting themselves. So I think that's how how, like, blaster eventually dies is relays realize, you know, I need a way to not have to scale my VPS   past $20 a month. And so I guess I'll upgrade to a relay implementation that allows me to protect myself. So and that introduces its own problems. It's kinda it it is interesting. If you if you say, I'm only accepting   like, that's a Relay's prerogative,   and that is gonna result in Relay's having different datasets,   which is going to make it impossible for blaster to function, and it's going to incentivize   the, the outbox model. So, like, over the long term, that's just what's gonna happen. It's that we need to, like, build this now and,   blaster makes it kinda hard and relays that accept public notes.  



[00:47:52] ODELL:

So you want less relays to accept public notes?  

[00:47:55] Hodlbod:

Yeah. I've always been about that. Relay decommodification   is what I call it. And it's like, you know, Fia Jaff has talked about this too, but it's, making relays interesting.   You know, right now, it's like, if I wanna join a relay, it's like, well, I'll choose the one with the lowest ping   and the fastest response times and the biggest database. That it's all like computational characteristics that that inform relay selections right now. But if it's like, well, I wanna talk about, you know, like, I don't know, Dustin Hoffman   podcast.  

Bitcoin Park. That's a better example. I don't know. I don't know why Dustin Hoffman hopped hopped in my head, but like very specific.   Very telling. R r slash Dustin Hoffman.   Very specific,   you know, areas of interest.   You know, relays can support that,   if if the owner is,   is,   aligned with with that interest. Right?  

[00:48:50] Mike Dilger:

Yeah. I don't see I don't see how relays are going to just allow anybody   to post,   some   some will will abuse that and just fill it up with a bunch of junk events   or use it as a disk drive,   you know,   You know, put take a movie and slice it into a 1,000,000,000 pieces of different events and stick them all up there to share movies.   There's gonna have to be you know, paid relays is where it's moving towards, I think.   And so there needs you know?   I don't know where I was going with that.  

[00:49:27] ODELL:

I mean I mean, I guess your point is paid relays are gonna be more of a thing.   But I assume not, like, pay once because that also doesn't seem to make sense.  

[00:49:36] Fiatjaf:

Yeah. We need some pay once the pay once thing was supposed to be about spam, like, you're not you're not accepting notes for anyone. It's not it is never supposed to be   a business model. I don't know why people thought that.   It's just an interest payment mechanism.  

[00:49:52] Hodlbod:

But there are all kinds of ways to make special purpose relays. Like, paying is one one way. There's, you know, relays might wanna pull in particular kinds of content,   to preserve it. There might be archival relays that only accept notes after a certain period, or there might be up to date relays that only accept new notes and then purge their cash.   There might be relays that represent,   actual communities. So based on, a social graph, you can just have, like, a seed pub key and everyone they follow. I mean, pyramid is a good example of this. There's relays that only accept certain kinds. So purple pages is an example of that. They all have functions,   and we need a way to differentiate one from another,   both on a, like, client level and on a user level so that people can make intelligent selections.  

And, obviously, like, this shouldn't be offloaded onto users because that's a terrible user experience.   But, you know, the the network should   sort of partition,   naturally,   as people use different things for different purposes.   So  

[00:50:53] Mike Dilger:

Yeah. There's some problems with with this, especially the user experience   and how do they know which use which relays to choose for their outbox or for their inbox.   New users   are probably gonna wanna sign up somewhere that doesn't charge them,   but then you have an open relay again, which can be abused.   And then for an inbox, well, if if I wanna accept, how do you make it so that your relay accepts any note that tags you?  

[00:51:23] Fiatjaf:

I mean,  

[00:51:25] Mike Dilger:

what I did in the Chorus relay was this. It you relay accepts any note that tags you, but doesn't send it back out again, so it's can't be abused.   And then the moderator has to actually say that, okay, anything from this user or or any of these notes is allowed to go back out again,   so that it could still be an inbox without without opening up that that abuse possibility.   But this is difficult, you know,   figuring out how to get the relays right is a whole another   whole another challenge. It's that's quite difficult. But  

[00:51:57] Hodlbod:

Yeah. And then there's monetization, of course. You know, we've talked about the cost side a little bit where small relays get bloated with a bunch of useless content, and then that increases their cost. But relay monetization   is an unsolved problem for the most part, especially for these, like, public bootstrapping relays, where if you got a new user and they wanna join the network, they do need, like, a public right thing like domus or no.law.   And,   you know, how those relays actually make money. It has to be attached to some other business and, like, may maybe to a client,  

[00:52:29] ODELL:

but I'm not sure how that what that would look like. I mean, like, Bitcoin Park is simple. It's like, okay. They're members of Bitcoin. And, like, community realize I mean, I know how about you a lot of your perspective is like, I want to move my church group off of Facebook Messenger.   It's like a church group is like a very easy like, of course, like a community relay, like, that makes sense to me. There's other funding mechanisms, and this is just one value add that the community organizer,   whatnot is offering.  

[00:52:57] Hodlbod:

Yeah. That can be an on ramp for sorry, Pablo. One   that can be an on ramp for certain people,   but not for everyone.  

[00:53:04] Pablo:

Go ahead, Pablo. Okay. But play play that part out. Play that case where you onboard your church,   and they are in tune after their party   participating in the community with with the the the churchgoers   within that church community,   but they start using Nostar beyond that community. And they're writing to their church relay stuff that is more public, so kind ones, global post, the Twitter, the micro blogging experience. But now because they're not writing to Deimos Relay, because they weren't onboarded via demos, they weren't onboarded via primal,   they maybe maybe their their events are blasted outside.  

But now if they're using demos, for example, they're not gonna see responses to their content.   So this is the part if you play out the part where a bunch of communities onboard people into Nostr, that's when the outbox model becomes even more and more and more clear.  

[00:53:58] Hodlbod:

Yeah. I totally agree with that.   Yeah. And, like, even if your community relay has public read,   and your 10,002s are replicated across the network as they should be,   if if you have a client that doesn't support,   support the the outbox model, it's not gonna see stuff,   unless it gets blasted, and then, you know, we're back kind of in the same situation we're in now.  

[00:54:20] Pablo:

Yeah. Unless unless it get no. No. But unless it gets blasted. But if the community   relay that you're running for your church is right protected as it should be, then replies   might not go into the relay.   Right?  

[00:54:34] Hodlbod:

Right. Yeah.  

[00:54:35] Pablo:

And you follow say say you follow Matt, but Matt is not part of that church. Now you're reading only from the community relay,   but you think you're following Matt. You just think that Matt Matt  

[00:54:46] ODELL:

never posts anything. Well, I mean, I would Yeah. Exactly. As it stands right now, like, if they would accept my notes, I am broadcasting to their relay.  

[00:54:56] Fiatjaf:

Yeah. But that's we we established that this high time preference.  

[00:55:00] Hodlbod:

No. No. No. No. No. No. It it it it it has   even but even if But even  

[00:55:05] Pablo:

right. It projects to right. And so even within that model It's so tight. Blaster   blaster Yeah. Publishes to the top 300 relays. So it gets from Nosto dot watch, it gets the 300 most popular relays, and it blasts everything to those 300 relays.  

[00:55:19] Mike Dilger:

But as relays   as relays protect themselves   and say, hey. I'm a special purpose relay. I only deal with certain things. You know? I've got my personal relay. It only accepts events from me   or events to tag me.   And so if blaster wants to put events there, and I keep seeing something keeps trying to put events there that have nothing to do with it, starting with a a really old event from FitchApp about Brandy.   And it just keeps it just keeps blocking it. It says, no. That's not related. You know.   Then blaster's not going to be able to put the events there.   And then,   you know, blaster doesn't fix the problem. You have to use the outbox model.  



[00:56:00] Fiatjaf:

One question I have is,   imagine Notar   succeeds and gets, like, a 1,000,000,000 users.   Will there be free public relays like we have today?   Yes. And who who will be running them? Well, Prime will  

[00:56:15] ODELL:

be will be running one of them for sure.  

[00:56:18] Hodlbod:

Yeah. People who wanna harvest data. Noster.band  

[00:56:20] ODELL:

will probably be running 1.  

[00:56:25] Fiatjaf:

Okay. I didn't have a follow-up for that. I think you you really don't need to be able to support,  

[00:56:31] Hodlbod:

for businesses   that that need that data. Right? So they'll host the relay. They'll suck in the data because everyone's posting to it, and then they'll power things like paid search or paid recommendations. And, like, Hootsuite and shit. Like, okay. I'm Coca Cola,  

[00:56:44] ODELL:

and I wanna know, like, how my new product is is being   discussed on,   on Noster, and I I need to see a full   a full view of the network as a result.  

[00:56:57] Pablo:

Like I do Yeah. Completely agree. I mean, it really starts as cost centers. For most cases, there's just gonna be cost centers attached to whatever Yeah. Product you're selling. The problem is is if there's only a few of them.  

[00:57:10] ODELL:

Right? Like, I've Fiat, John, this is what you, like, originally sold me on a Nasr. It's like, it doesn't have to be decentralized.   It just   like, it it is it works because it's in a lot of ways, it's very centralized. But if if if there's not friction if there's not that much friction to move and there's not that much friction to run, you know, competing relays,   then when censorship   does rear its head, we can kind of route around it.  

[00:57:36] Fiatjaf:

Yes. Yes. That was the idea.  

[00:57:38] Pablo:

But only only if you can find the relays where you're publishing   or you move to. Right? If you cannot find the relay, then you're fucked.  

[00:57:46] ODELL:

Okay. Besides calling me high time preference, because there's nothing high time preference about quitting Twitter and doing noster only,   during the greatest bull market of all time.   Like, there is no incentive for me not to   broadcast my notes to every single relay that will accept me. Like, that is that is what any user that's concerned about censorship should be doing.  

[00:58:10] Mike Dilger:

That's fine. I'm predicting that the relays will over time not do that, not accept that anymore. And it's not it's not just the monetization that you guys are talking about, that relays are just a cost center, and only if someone wants to, you know, collect a bunch of data analysis that they would bother to run one, it's going to be the abuse.   Because if you have an open relay that just allows anybody in there, that's are you gonna monitor everything that's going on? Or are you going how are you going to stop   someone that writes a program that just generates random pub keys and start shoving things in there to try to take you down?  

So   relays are going to have to defend themselves against abuse, and I think that's what's going to eventually close it all down and require people have to sign up for some service on SunRelay.   Yeah. Even beginners.   I that's just my prediction. I mean, I don't know how it's actually going to go, but I don't see how to stop that problem. At a fundamental level, it seems unstoppable.  

[00:59:07] Hodlbod:

Yeah. I mean, you're right. You're right, Matt. Like, there is no incentive against   blasting notes for individuals. And if you if you're care caring about getting something done today, growing your platform or reach or whatever,   like, that's how you do it, and that's fine.   But, like, maybe as a developer, you know, I just wanna be able to see my notes because   yeah. But I I actually want people to not see my notes if gossip isn't working.   Right? Because that will mean that   and, you know, it it'll reduce my reach, and I my reach is, is probably reduced because I don't use any big relays.   But that's okay because I feel the pain, and now I can be a squeaky wheel and and yell at Tony more often.  



[00:59:48] Pablo:

Matt, there there is no ins there there is,   how did you phrase it? There is incentive for you to run to use blaster, or there is no Well, I mean, let's take blaster out of this. When I had Mike on the show  

[01:00:01] ODELL:

4 weeks ago, right, by default, I was connecting to 32 relays.   Like, forget about blaster for a second. Right? It's like I'm using it on desktop.   I wanna see everything, and I wanna be I I I don't wanna get censored.   Right? It's like my incentive is to connect to as many relays as possible.   But but that's only  

[01:00:22] Pablo:

like, you are doing all these and it's fine because it's not a big problem, but you're doing this inefficient way of using Nostra   just because most clients don't do it. Right? Right. Just because you're not using Outbox, you have to do it in this way. Right. And you find out about this new pleb who is on this other relay, and you add this other 33rd relay, and now you're requesting for everybody you follow from that relay as well. From all 33 relays you're requesting from everybody the same thing.   And it and it's fine. It's not a big deal. It's a like, a little bit not as efficient as it could be, but who who cares?  

But the the the issue is that there there is,   currently,   there is an incentive to use,   there is no incentive not to use plaster and to publish everything you want everywhere,   but there is   there is no incentive not to do it. There is no incentive because   why wouldn't you? You want You might as well blast it. You might as well blast. Right? And you are going to be something out of it. This is why Jack banned himself,  

[01:01:26] Hodlbod:

because he has an incentive   to to make no store work in the long term. So he's he's much more sensitive to this,   you know, whereas Matt has no incentive to, to make no story. How do I have no incentive?  

[01:01:40] ODELL:

I have a huge incentive.   I'm just kidding. But, like, you know, every person like shits and, like, the whole world knows he shits. Like, I I have a bigger incentive than than Dorothy does in terms of Every person who,  

[01:01:52] Pablo:

maybe they're not the worst. Like, term.  

[01:01:54] Hodlbod:

Every person who starts posting to small relays or and bans themselves from the big relays is is one more person that people What is a small relay? Like like Pablo's relay, like, connecting to Pablo's re does that really solve anything if we just all connect to Pablo's relay instead? No. No. We we shouldn't all do it. But right now, like, really interesting accounts who do post to, you know, not the top, like like, 15, let's say,   those people are   breaking domus   by how they're using the protocol,   and and Amethyst and anyone else, not to pick on Domus. But they're breaking things, they're breaking clients by putting their information somewhere that those clients are not gonna see it. And then people following them are gonna say, like, well, what's the deal? Like, why can't I see Jack's post? I follow him.   It's well, it's the fault of the client that you're using. They're not using Nostra properly. So that's what I mean by, like, low time preference is saying, I'm not gonna build my platform right now. I'm not gonna let people see my notes so that long term, I have access to the censorship resistant protocol   that,   will ensure that my notes can be seen. So and and we have, like, Twitter. So just just cross post to Twitter. Be a blue check-in the meantime if you Are you really cross posting to Twitter during this whole process? No. No. I've I've deleted my Twitter account. Because that's really hypocritical. That's the biggest relay. I walked in. Oh, you're, like, you're cross posting to Twitter.  

No. But if you have, like, a business you need to run, like, post to Twitter because no sir is not ready for you yet. No. Nostra only.  

[01:03:24] ODELL:

Lower your time preference.  

[01:03:26] Fiatjaf:

Okay.   Blast all the relays. Well, yeah, good question you're asking right now, like, what's a small relay? I've seen a lot of people asking these questions since we proposed that Jack and Odell started publishing through small relays.   And some people are interested in doing that, but then they ask, what's a small relay? I don't think that that question   is easy to answer, but,   it's it's basically the one relay that people don't know and they're not configured by default. But I don't think people should be searching. I don't think that's the correct way to use Nasr. Like, you open a list of relays and then you choose 1. I I don't think that's a good idea. Like, you you should   in a normal   way, in in a world in which Notar is successful, you could you should, like, already know some relays.  

Like, when you join Allstar, you'll join through a relay or   through someone that invited you and pointed you to a relay   of a community or something like that. Like, the client should handle it, basically, for the most part. No. No. Someone should invite you. Yep. And then you get to the client and the client points you to the relay of the people, like the person who invited you. Something like that. I don't know. Maybe a local relay. I don't know. What happens when someone joins Twitter for the first time?   Like,   people was was always people were always asking, like, how do I get how do I know who to follow when I first joined Oscar?  

I don't know because, like, this is the same question. Like, what relate do I pick?   When you first join Twitter, and if you don't know anyone,   like, you just follow a bunch of NPC accounts   from media   people. Like, that's not the experience people like on Twitter.   So   you you should join the thing already knowing something. I think that's the natural way that will happen   in the future.   So whatever.  

[01:05:19] Hodlbod:

We have some nippo 5 maxis in the chat.   Yeah. Speaking clearly,   why can't we use nipo 5 relay list to advertise where to find our notes? All caps.   The answer is, we should. Like, that's that's part of clients,   bootstrapping, basically, you know. If you're given an NPUB I mean, sorry, if you're given an a a NIPIFYB address, then you you should have relays that you can start with, and then you don't have to consult an indexer or something like that. NIPO 5 is just an indexer solution that works over HTTP and DNS   rather than over relays.  



[01:05:50] ODELL:

But then everyone's just gonna use, like, the primal nippo 5. Like, how does that solve shit?   It doesn't solve shit.  

[01:05:57] Hodlbod:

Well, hopefully, primal will allow you to set your relay, like, it like, nifa fives should mirror your 10,002's.  

[01:06:04] Fiatjaf:

Most people are not 50% sensitive. Most  

[01:06:09] ODELL:

we need the content the content needs to be spicier. I think the I think the big issue is, like, there's not enough sensible content. Like, if there's more sensible content, then we'll fucking figure it out. Go go browse global go browse global on a Japanese relay and then come back to me. No. Don't global's a straight point. You don't check global. You never check global. Primal, you can't even view global on primal.  

[01:06:31] Fiatjaf:

Doesn't because you you are screen I've never seen Your computer would crash. I have I've still never seen global.  

[01:06:37] Hodlbod:

Don't do it. Don't do  

[01:06:40] ODELL:

it. Yeah. Connect to 32 relays and then click the global tab. That's a horrible idea.  

[01:06:47] Hodlbod:

So you're saying that, people who actually have some sort of,   following should spoo should post spicier content. No. No. No. I'm not sure what spicier than, like, Jack posting,  

[01:06:58] ODELL:

everything is a rich man's game. Okay. Jack hasn't posted anything spicy. It's not spicy enough yet.   I I look. I think I think that Noster  

[01:07:08] Mike Dilger:

You wanna start the censorship early so that we can fix it early. Noster's gonna have, like, it's Alex Jones moment.  

[01:07:15] ODELL:

And we I like, we can, you know, have these super technical conversations,   and   Pablo can ship, like, a 1000000 different fucking projects, and, like, all this different theoretical conversation can happen. But, like, it's not going nothing of of substance is gonna change   until we have, you know, a fire rod   that gets censored off of all centralized social media that moves to Noster that could starts getting centralized on the major relays,   and then there's, like, actually pressure to, like, change things.  

[01:07:46] Hodlbod:

I think the problem with that is the major relays right now are run by people who care about censorship resistance, and so they're not gonna censor very much. The precondition for that not a problem. That's good.   Yeah. But but the precondition for people get getting banned   is bringing people in who wanna make money, and so they run a huge relay and outcompete everyone because they wanna make money, and then they wanna make money, and then they ban people. So we need to, like, get giant,   companies that, that wanna censor us on the Dosto first. Maybe, maybe threads is the beginning. I don't know.  

[01:08:23] Pablo:

The thing is is if we go through that test, if there was, like, a court order saying, Will, you have to censor this pop key. Like,   we would fail the test   so hard   unnecessarily, but we would fucking fail it so hard. It will be   laughing stuff.   Like, Bitcoin was used by Wiki p by, WikiLeaks.   Right? Like, they were able to use it. Right. We would not be able to sustain an attack, like, not even a little bit.  

[01:08:53] Hodlbod:

Unless everyone uses that account. Use that?  

[01:08:57] Pablo:

Alright. I mean, look at Pierre Chefs' test.   Just I saw all the tests.   Afterwards?  

[01:09:04] ODELL:

No. While it was at the hospital  

[01:09:06] Fiatjaf:

What if primal what if primal had received a court order?  

[01:09:09] ODELL:

Then I would be I would be like, primal received a court order to censor fiatjaf. Like, stop using fucking primal.  

[01:09:16] Fiatjaf:

You know you know, Humble and Locust, these 2 big platforms there, they don't serve any content to Brazil anymore because they received court orders from Brazil   asking them to censor some specific people. So they shut down the entire site for Brazil.   I think they did the the most honorable thing possible,   And, well, people are still using them   on the United States, IMS, I guess. People who are very   mindful about this censorship stuff.  

[01:09:48] Mike Dilger:

Which platform was that you said did that?  

[01:09:51] Fiatjaf:

Humble and Locos.   Oh, right. Yeah. It's it's called Humble. So they're not they're not Humble. Humble. Humble. Humble. Right? I don't know if you Zoho. Okay. Don't make fun of me.  

[01:10:02] ODELL:

I wasn't making fun of you. I was like, shit. There's a platform named Humble, and I didn't know about that.  

[01:10:08] Fiatjaf:

It's it's basically the same thing. Right?  

[01:10:11] Mike Dilger:

That's that's why I asked. I thought that's what you said, but I couldn't if I tell. You should rebrand those here to Humble. Yeah.  

[01:10:17] Hodlbod:

That's a great  

[01:10:19] ODELL:

What did we call it? Or making fun of Wait. Instead of the outbox protocol, Humble protocol.  

[01:10:24] Pablo:

I like it. There you go. I like it. Just fucking go.  

[01:10:27] Hodlbod:

It's a thing now.   Stay on humble and snack stats.  

[01:10:32] ODELL:

But, I mean, look, Theo Jeff, that's, like, a perfect example. Right? It's like there is a bunch of people that actually believe that Rumble is censorship resistant, and then their whole country gets banned from it. And they're like, oh, no. It's not. We need to improve the situation.   But until that happened, no one was gonna improve the situation. I have a reply guy on Noster who fucking loves Rumble. He's, like, constantly telling me, like, caps are not not okay. Like, I shouldn't do like,   he calls me a retard. It's like, oh, well, you're the reply guy to the retard, so who's the fucking retard? You know?   People believe rumble is censorship resistant until it's not.  



[01:11:09] Hodlbod:

Like, that's I have the same thing with with Gabb.   When I first started looking into this stuff, I was like, oh, Gabb, they're, like, redoing their architecture so that they can run on multi cloud   kind of stuff. So if cloud providers censor them, they can still be up. But then, like, they're just   relocating the locus of censorship in themselves. It's it's so dumb that no one has done this. I just can't even believe.   I I can't believe it. It's   yeah.  

[01:11:36] ODELL:

So what's the what's the yeah. Go on. Mike.  

[01:11:40] Fiatjaf:

Mike? Nope.  

[01:11:42] ODELL:

I'm just page topic Sorry. Already. Okay. Let's talk what do you wanna talk about?  

[01:11:48] Fiatjaf:

Wikifreeda   is censored in my articles.  

[01:11:55] Pablo:

Get a better protocol.  

[01:11:57] Fiatjaf:

I've created I've created 3 different key pairs, and I've copied   some content from, like, a movies database and another movie database and created an articles automatically for each movie that exists.   And, also,   some other feedback about the names, name like, the history of names. Like, there's one article for each name, and it's published by this   different key pair   to my own relay   that has all this crap.   And I can't read anything of that in Wiki 3rd year, and I don't know how to fix it, how to properly   select relays   for that use case. It's not the outbox model.  

I don't know what it is. The humble protocol. Yeah. Yeah. The so on on the humble protocol, we have,  

[01:12:41] Pablo:

HIP we have HIP,   34   That,   did you put the,   that new relay that you created on on one of the whatever relay kind we created for for Wiki?  

[01:12:56] Fiatjaf:

No. Forgot the name. I did not. What's the name? There you go.   Well, but  

[01:13:03] Pablo:

go to nostril.com   and read how it works.  

[01:13:08] Fiatjaf:

No. But what's what's the flow? Like, someone wants to read about some movie,  

[01:13:13] Pablo:

at the it's that Yes. So so movies published. So if I if I if I were to follow that those pub keys that you wrote from,   or someone within your web of trust follows them, it will check for the relay   list of Wikipedia, of Wikifredia entries,   of Wiki entries, and it will connect to those. So it will do outbox,   to to those really, so it will find them. But if it's just a random pop key in the middle of nowhere,   and there is no way to get to that What's what do you mean by that? Web of trust. Do you have a degree in web of trust or a master's? You can configure it on the no. I a degree. No. There is a degree of the configuration of the web of trust, but,   I got fired from synonym, so I don't know.  



[01:14:01] Fiatjaf:

So it is it's like, what what's the default? Like, 2 2 hots away?  

[01:14:05] Pablo:

3. Okay. A friend of But if you go to settings, you can change it.  

[01:14:09] ODELL:

Okay.   By the way, in the in the live chat,   this is why I love Nostril. Like, Kieran's, like, debugging his features in my live chat,   Live.   Post a clip says don't use a clip. Don't use clips.   Big pink box.  

[01:14:31] Fiatjaf:

What What about the inbox model?  

[01:14:33] ODELL:

What is the difference? So inbox is where I want to receive my notes. Is that the  

[01:14:39] Fiatjaf:

difference? Yeah. Yeah. You could say that, but we we are horrible at names. Like, in the current 65,   there's a concept of inbox, which is where we want to receive   reactions, replies, and   mentions, stuff like that.   But   Alex Glayson suggested some days ago, and   I proposed the NIP for the inbox model. I don't know if it's a good idea, but I just wrote it. I hate it.  

[01:15:04] ODELL:

I hate  

[01:15:06] Fiatjaf:

it. So you 2 are wrong, and I think my I think the  

[01:15:09] Hodlbod:

the basic idea is interesting where you just, like, put something out there that's, like, I want this kind of stuff, and anyone who wants to give it to me can put it in my inbox. No. That's not that idea. That's that's yeah. That's much looser.   Well, it's similar. I mean, it uses the uses an advertisement.   Yours is way too way too strict, way too structured.  

[01:15:27] Fiatjaf:

My idea is, like, I wanna I wanna subscribe to I wanna follow Pablo, and Pablo   signals that he implements this   this model. And then I send a follow request or something like that to Pablo to Pablo's release release that Pablo advertised.   And   then the next time Pablo publishes any note,   he will send like, in this in this follow request, I I include the relay.   And then Pablo will sub submit his note directly to my relay that I note like, I that I wrote there.  

[01:16:00] Pablo:

So so so, Matt, to to make to translate that to English, this has the feature.   This has this this model that FireChef proposes has the feature where you can actually block people, like, for real. Like, they cannot see your posts.  

[01:16:15] Fiatjaf:

So But you you you you use this together with the Outbox model. You would still publish stuff to the relays that you advertise. So people can still follow you there.   It's just a  

[01:16:26] Mike Dilger:

Mhmm. Instead of instead of you going and and getting Pablo's notes from his outbox,   you tell Pablo to put his notes in your inbox.   Right? So you kinda flip it around. But what you're basically you're taking the locus of control,   and you're handing it to Pablo. And so then Pablo could say, I'm gonna cut you off. And then not send you the notes anymore.  

[01:16:48] Fiatjaf:

Yeah. But then the client, after   realizing that for a week, Pablo hasn't posted anything, your client goes there and checks if Pablo is publishing through his outbox. And then you switch  

[01:16:58] Hodlbod:

Pablo publishes.  

[01:16:59] Mike Dilger:

Well, I I mean, I think it's If it comes this potentially   I think there's use cases for it, but   I I I don't wanna use any of those use cases.  

[01:17:12] ODELL:

Doesn't the doesn't the Outbox model just solve this?  

[01:17:15] Fiatjaf:

It does. Outbox model solves everything. But But I'm I'm I'm saying that there's a different there's a different way of doing the thing.   Like, you you you don't have to equate.   But you don't have to equate to us or with the other one.   Yeah. You do things multiple ways. No? You don't?  

[01:17:34] Mike Dilger:

It's okay. Experimentation is good. Criteria for acceptance of NIPs, There should be no more than one way of doing the same thing.  

[01:17:47] Hodlbod:

That's right.  

[01:17:48] Pablo:

It's it's stuck, Matt, the the chat.  

[01:17:52] Mike Dilger:

I don't I don't think it's the same thing. So I I think it's Oh, Kieran Grove is there. If   if someone finds a good use for it, I'm fine with it.  

[01:18:01] Pablo:

It's so much more complicated. Yeah. I'm like I read it. I was like, okay. What's the goal of this? And he was, okay. I I don't see   a point. It's really complicated.   And, yeah, I don't know.  

[01:18:15] Hodlbod:

So something that's similar that,   that I was talking to Nanya Bidness about,   he,   he he proposed this idea, is to just put put something more generically out there. So rather than delivering it to a particular person's,   inbox and saying, I want your stuff, you just put something out there and it's like, I want more information about   Bitcoin or whatever. And then people can and the thing about that is it not only means that people would deliver stuff to your inbox about Bitcoin, but it would mean that people know that there's a market for that or like a, you know, market of attention. And so people might be incentivized to actually create more content. I think that's that's the best application of this sorta architecture that I can think of.  

[01:18:59] Fiatjaf:

That's a completely different idea and has nothing to do with what I said, but I think it's a good idea too. Although I like it. It doesn't work   in the naive way you said, but I think it can't be can't be used can have uses.  

[01:19:13] Hodlbod:

You're calling me naive.  

[01:19:17] Fiatjaf:

Well, if someone creates, like, a 1,000,000 puppies saying that they want more content about,   I don't know, basketball. Are you going to start getting basketball content? Like, it's just just going to scan you of your time   and and then of your bandwidth having to publish stuff to a relay that doesn't even exist,   where supposedly someone is looking for basketball content.   But Yeah. Yeah. I think it could could, like, could have its users, but then it would be, like, 10 years ago. In in 10 years, we can think about that.  

[01:19:49] Pablo:

Yeah. I'm not gonna do it. Or you could just publish an event saying, oh, I would like to read more about   spinners or flowers or whatever. Right? And it's it's just people can't fulfill that. Just go to Corigall. You'll know everything you need to do.   All your spinners are fulfilled.  

[01:20:07] Hodlbod:

Can we talk about proxies?  

[01:20:10] Pablo:

Okay. By the way, MBK wants us to talk about OMIMO.   Jesus Christ.  

[01:20:18] Hodlbod:

I'm not smart enough. You'll have to do it for me.   I think proxies are cool and, like, you know, they're like a a partial solution to a lot of these, optimization problems, from a client's perspective   because they can implement the the outbox model.   You know, if you have a proxy, you can just ask it like, give me stuff from these pub keys. And then it can go figure out where those pub keys are and keep an index of all that stuff internally. The and that makes clients a lot simpler potentially,   so reduces the barrier to entry to to the outbox model.   Some interesting things about proxies are you can't proxy auth right now. It's impossible.  

But,   I love this idea that Pablo mentioned to me.   No one has been talking about it, but I really want it to exist. I'll build it if I ever get around to it. But making NIP 46 signers that are also proxies.   Right? Because you already trust the NIP 46 signer with your private key. And And it's a common do that   yeah. Yeah. Well,   all all this all Net 46 signers. Right? I mean, you gotta have the Yeah. No. I'm just I'm just giving reference to to Matt.   Oh, okay. Yeah.   So, so what you can do is you can reduce a lot of round trips with that signer by just sending an unsigned event to the signer and saying, publish this where it should go. Then you don't have to do relay selection for publishing either,   and you, you reduce the number of network requests, the signer can double as a proxy.  

I think it's great,   with the caveat that you have to put your private key somewhere. But,   proxies would be easily self hostable.   The database wouldn't, wouldn't grow any bigger than, any clients, database.   So I think it's totally doable.  

[01:21:55] Fiatjaf:

And, someone should do it. Maybe I But you should should you do it, like, through the normal   Nasr,   transport?   Like, you're you're going to subscribe  

[01:22:04] Hodlbod:

with the rack thing to your own proxy, or should we have a different protocol there? Oh, yeah. I would open a direct WebSocket to the proxy. It doesn't need to go over DVMs because then you'd have to, like, you know, proxy. I'm not talking about DVMs.  

[01:22:17] Fiatjaf:

I'm not talking about DVMs. I'm I'm saying like Well, yeah. But, like, the NIP 46 is currently,  

[01:22:23] Hodlbod:

works with the same architecture as DVMs because you publish,   like, the transport is relays. So I would just assume intermediate relays and do a direct connection to the signer. Sorry, Pablo. But  

[01:22:34] Mike Dilger:

I I think you could go one step further.   I think you could take your your proxy idea. Well, that's a that's also a a a bunker,   and then actually combine it with the client into 1 big program?   Yeah. You could.   If you want a back end. A client should run back end. He's just going backwards to where he came from. But,   I think it's a good architecture if you have a a a weak   system, like a phone to run your main visual part of it on.   But then you have to have some sort of trust relationship   between that phone and that proxy, which is, I think, the biggest sticking point in most situations  

[01:23:14] ODELL:

unless you're self hosting. Yeah. But most people are not gonna be self hosting a proxy.  

[01:23:20] Hodlbod:

Yeah. And you can partially solve that by paying for the proxy because that aligns interest. But, yeah, your your password is still stored on a server. So if you don't want that, then run a desktop client or a console. Trying to say is that you shouldn't treat the proxy as a relay. Like, if if the proxy is just an agent, like, it's it's just a a back end to your to  

[01:23:39] Fiatjaf:

your thing client,   and it's doing all the work on behalf of your client.   It's not it's not acting as a relay for your   client. Your price is just a front end, and this product is just a back end. Isn't this basically what Prime is doing?   Yeah. It's basically what Prime is doing. It's also what CBD Social does.  

[01:23:58] ODELL:

Wait. Does ZBD Social still exist? I haven't seen any of their notes. Are they getting censored?   Well, we don't follow anyone from that, I think. I feel like they've all been they've all disappeared.  

[01:24:10] Fiatjaf:

There's a bunch of people there. They're posting bunch of really hot chicks. That they they want. Too much.  

[01:24:16] ODELL:

I don't know. I use Monster every day, and I I feel like I they've disappeared.  

[01:24:21] Hodlbod:

Time zones.  

[01:24:23] ODELL:

Probably not, though. It's probably not time zones.  

[01:24:26] Mike Dilger:

Maybe they're Oh, they're using the outbox model.  

[01:24:29] Fiatjaf:

They are using the outbox model. It's a custodial outbox model like what Autobot is is proposing. Like, you have a trusted server that does the outbox model for you and just served you   the content already stripped of anything, like, just gives you   the the  

[01:24:47] Pablo:

the thing meshed. I don't know. It sounds like it's a If there is any hate speech, you can take it out. I like it.  

[01:24:56] Hodlbod:

What I'm talking about is personal web nodes that actually work. We need more hate speech on Aster.  

[01:25:01] Pablo:

Agreed. Spicy.   This is going to system. This is a we need to get this over here. Calvin Air go with get Calvin Air to come to Noster and publish some of  

[01:25:11] ODELL:

his. That's what we need to be cash. We need the the   Satoshi's visionaries to  

[01:25:17] Hodlbod:

Yeah. There you go.  

[01:25:21] ODELL:

ZBB has the hottest girls. Do you read in this chat? Fucking crazy.   I haven't noticed.  

[01:25:29] Fiatjaf:

If you can open the   the pure zbd feed, if you go on coreco and type,   I think it's coreco.social/relay/  

[01:25:38] ODELL:

Are they all just posting to the zbd relay only? G g.   Is that what's going on?  

[01:25:46] Fiatjaf:

The ZBD users know they they post, like, the the the stuff is blasted   to a bunch of relays. Yeah.   But they all on the z b like, the the z b d relay,   they only had only has the the post from ZBD users. So, Jeff a bunch  

[01:26:03] ODELL:

of women. Yeah. Jeff, we have you on the pod.   We have you on dispatch. It's a live show unedited.   ZBD pays your rent.   Andre literally comes into your house and pays your rent every every month.   How do you feel about custodial   clients? Like like, where the user can't even pull their NSAC out. How how does that make you feel?  

[01:26:26] Fiatjaf:

I think   that was always my position. I think it's   I think it's okay for most people, but I wouldn't like it.   Everyone's No. You should you should ask you should ask you should ask Pablo because he's the biggest proponent of custodial,   NOSFET clients NOSFET teams.   And I think with the with the key migration thing,   that that probably gets   very much do you mean, like, it's it's much smaller?  

[01:26:54] Pablo:

What do you mean the key migration to the   the the the thing that we discussed about,   migrating the identity where you can rotate your end second to your end pub into a new end pub.   And there are signaling, and there's all kinds of shit to to You're basically just, like, signing a note that says, like, I'm gonna rotate to these and put some pictures to that with some Basically, yeah, with with certain guarantees that Yeah. But it's a hack. Different ways.   Yeah. It is a hack because you cannot change math. Right? So it is what it is.   But the the cool thing is that we could go into   something that PHF loves, which is have some kind of,   master public key so you can derive,   you can derive children,   child keys,   and,   basically, like, having, like, an expat.  

But, of course, we can't with the current keys, So we need a scheme. We could have a scheme that migrates from the current scheme to, to a new scheme where we have these these x pups, and we can do rotation without this kinda I like to hack. The hack works.   The the the hack is required, basically.  

[01:28:01] ODELL:

We can take it further or not? You just post a note that you're like, I'm gonna rotate to this if I get hacked. And then if you see a note that says I got hacked, then you rotate.  

[01:28:11] Pablo:

Yeah. But you have a whole kind of problem with, you know, saying But it's basically that.   It it is it is basically that. You're just anchoring to to to the Bitcoin blockchain to so that if someone has your key,   they cannot publish, an event that is older than you're saying, oh, I'm rotating to this new key, and they rotate to a to a key that they control, but you don't. So it it's basically that, but with open time stamps. Yeah. What if Peter Todd senses you from his open timestamps server?   Well, then we're fucked. What what is she doing with Peter? Tell us to do this.  

[01:28:45] Hodlbod:

Put it on Ethereum.  

[01:28:50] ODELL:

Is it Nostr or Nostr?  

[01:28:53] Fiatjaf:

Jesus Christ. Again, this question.   Someone asked it in the chat. It wasn't my question. In Brazil in Brazilian people, they're talking about they they say no STR   in Portuguese. No STR? No STR.  

[01:29:05] Hodlbod:

Like the It's both.   That's what I think. Wars.  

[01:29:09] ODELL:

I like Nostr.  

[01:29:10] Hodlbod:

The protocol the protocol can, can handle partitioning and so should the pronunciation of the name. Fair enough. It's Nostril. You mean the Humble protocol?   Yeah.   I pronounce Nostril humble.  

[01:29:24] ODELL:

All caps.   You couldn't tell from the pronunciation.  

[01:29:30] Fiatjaf:

Can we talk about Omnimo now? Or Yeah. You wanna talk about it?   Omnimo is like a protocol for encrypting   chat   that has been abandoned,   like,   7 years ago, I think. But NDK wants us to implement this.  

[01:29:48] ODELL:

I mean, he's just trolling. Right?  

[01:29:52] Fiatjaf:

Not when Matrix uses? Somehow. I think that's my that's my understanding.  

[01:29:57] ODELL:

How is this relevant at all?   It was just what NVK was was talking about. Just ignore him. I'm not gonna ban him from the live chat, but just ignore his his his notes. Okay.   I'm not gonna give him the satisfaction of of getting banned.   Is there anything of relevance you guys wanna discuss while I have you?   While we have you?  

[01:30:24] Hodlbod:

Pub key hints.   That's another wrinkle of this whole thing.  

[01:30:28] Pablo:

Oh, lord. You're you're gonna you're gonna make this even more niche.  

[01:30:32] Hodlbod:

Oh, yeah. Oh, yeah. Let's go.  

[01:30:34] Mike Dilger:

This is definitely what works all about.  

[01:30:37] ODELL:

What is what is pub key hints?  

[01:30:39] Hodlbod:

So right now we have Relay Hints. Right? If you e tag or a tag something, you have a Relay Hints so you know how to find it. But that is,   I don't know. There's probably   some use cases for it, but I can't think of any because the Relay is like, it's high time preference again. Like, it's all stuff in the short term, but only as long as those relays exist and actually have the data and clients are choosing good relay hints. And most clients don't even have relay hints. They just put an empty string there. It's totally pointless.   And the the big thing is that relays are gonna go offline,   and also DNS is in the mix. So they're just extremely fragile. There's all kinds of ways that these can fail. But the thing that is They they can even be wrong in the first place.  



[01:31:20] Mike Dilger:

Yeah. Yeah. Exactly. There's, like That's not a signed piece of data. I mean, I mean, it's signed by the person to put the hint there, but the fact that the that the event is actually on that relay, who the hell knows? It might have been a lie.  

[01:31:32] Hodlbod:

Yeah. Relays it's it's the whole, like, state versus values thing. State is named values. So relay is a name for a bunch of values, but you don't it doesn't have it's not a hash. It doesn't have anything to do with the values that it's actually holding. So because of link rot, we need better hints, and a perfect hint is a pub key. So and there's, like, there's a proposal to add add this, but,   you know, a tags already have pub keys, pub key tags already have pub keys, obviously. Etags don't have pub keys, but if we can add a pub key to etags,   then all you need to know is the 10,002 for that pub key, and then you can find stuff. So I think it's a great idea, and we should do it.  



[01:32:09] Mike Dilger:

Yep. I support it. I've I've already started working on the code.  

[01:32:13] Hodlbod:

Woo hoo. Ship it. Done. Oh, that was easy. Alright. I thought VHF was gonna argue with me, but  

[01:32:19] Fiatjaf:

I like the relay hints. Don't don't remove the relay hints.  

[01:32:24] Mike Dilger:

What do they do? What do they do? The thing the thing that's that's not the greatest is that if you're using that same slot that's been used for URLs and now you're gonna put pub queues in there, Then your code has to sort of older code is gonna, you know, choke on that. Hopefully, it just Yeah. But they have to handle it. Ignores ignores the error. Right? Because yeah. You have to be, you know, hardened.   And the newer code is gonna say, well, if it wasn't a URL or maybe it checks public key first, then then check it the other way.  

[01:32:56] Pablo:

Kieran is asking on the on the chat, how do you get your the 2,002 if you're in an obscure relay?   That's all the way. No. No. Not really. Like, you can ask basically any relay.   That's I I think Bitcoin, like, Bitcoin when you when you run Bitcoin d, it uses DNS to put strap to find the peers.   And I think that's that that's the idea that I had when I when I started working on the Peruvial Phases Relay is a Bootstrap really doesn't have any data other than this. It's easy to run. Now we have,   no social also runs a a relay like this one, and it's just having a few of these relays to start finding the rest the network. You have you need to have some kind of, bootstrapping process.   This is one of them. It doesn't need to be the only one, but that's   that's the way. Yeah. But the 10,002 should totally be blasted.  



[01:33:47] Hodlbod:

You can also check,  

[01:33:48] Pablo:

nibbo fives. You can check like, you could nibbo 5 stop.  

[01:33:53] Hodlbod:

Yeah. They do. But, but it is another way to find the relays.  

[01:33:56] Pablo:

I I still have BRB on my NEBO 5,   and vacation.  

[01:34:02] ODELL:

You can also build What a waste of time that was.   BRB.  

[01:34:09] Mike Dilger:

Sorry. Continue. How about Oh, yeah. My my nip o 5 keeps getting out of date   because I keep forgetting to update it. I update my my my 10,002 in my client, and then I'm like, oh, shit. I've also got this other file. I have to go update it. And then I leave a fucking trailing comma in there, and it doesn't work for a month.   Yeah. That's the worst part. Like, I haven't been I haven't been pretty good about that, but, you know, it it'd be better if it was somehow automated that the client could automatically keep the NIP 0 5 in sync.  

[01:34:37] Hodlbod:

It's totally possible, but every NIP 0 5 provider would have to implement that or or use the software. It's not that hard a problem, just no one is bothered really with it. But all all this stuff is DNS. Wait. What does this mean for someone who self hosted NIPO5?  

[01:34:51] Fiatjaf:

You'll just have to run your own domain, run the right software, or update it yourself. No. Forget about that. Just add it manually. Just a file. You you already have a website with a domain name that is just you, and so you you have to update your Nipple 5 manually. I think that Nipple 5 thing was   I I don't know. I I never liked these   providers that showed up, like, a bunch of   and NOSR whatever and my NIPL five. I don't I don't think those serve any purpose. Like, Nipple 5 should be for   platforms that have already   names, like users register their names and   or people that self host their domain. I don't think these products these providers   have any point.  



[01:35:34] Hodlbod:

I don't know. I think it works fine as an address,   but that's all it is.  

[01:35:40] Fiatjaf:

The relay stuff's not bad.  

[01:35:44] Hodlbod:

I was playing with this, you know, disintermediating   DNS for bootstrapping,   you know, pub key relay selections. Obviously, that doesn't completely take DNS out of the picture because WebSockets rely on DNS. But,   I I played around with GunDB   and, putting,   putting these puggy putting Kine 10,002,   events on GunDB.   So you could just open up gun and grab a key.   Obviously, it's a DHT, so, you know, Pablo and Fiat Jack are laughing at me. But,   you know, it it proves in theory that you can use different indexes   to, to bootstrap your connections. So there's that.  

[01:36:28] Pablo:

Yeah. I think the right approach is just using ordinals.   We'll we'll all make more money. Let's just do that. Yeah. This is really devolved.  

[01:36:45] ODELL:

I mean, I I think Nastr is good. Do you guys think Nastr is good?  

[01:36:50] Pablo:

Nastr is good enough for casting. Small. I'm warming up too. It's too small.  

[01:36:56] ODELL:

You're warming up too, Mike? Is that what you said? Yeah.   I mean, it works. Like, it works. Like, you got I like, I appreciate you guys going deep and,   you know, trying to make it as robust as possible, but the fucking thing works.  

[01:37:11] Fiatjaf:

Yeah. That's true.  

[01:37:14] Pablo:

That's amazing. It it works under these conditions.  

[01:37:17] ODELL:

There's, like, 10 people that have made it an hour and 40 minutes into the show that literally just wake up in the morning, go on to Nasr,   use whatever app they want to, they say good morning to me, and then they just receive Bitcoin, and, like, that's Nasr to them. And they're here. This works.   Yeah.  

[01:37:35] Pablo:

At this scale, it works.  

[01:37:37] ODELL:

At this scale? At a 1000000000 people, it's gonna break.  

[01:37:40] Fiatjaf:

How do we get more users?  

[01:37:42] ODELL:

Spicier shit.  

[01:37:45] Hodlbod:

Communities.  

[01:37:46] ODELL:

We need the spicy shit. Communities.   Not communities.   I mean, how about I like I I want your church group to move off of Facebook a 100%.   I just did like, that's not how we get a 1000000000 users. Maybe.  

[01:38:01] Hodlbod:

Communities have network effect.   You because your network is smaller,   you you can get people into a protocol that doesn't have everyone they wanna talk to because they have a subset of the people they really wanna talk to on it.  

[01:38:16] Pablo:

People go through hoops   to get to the people that they so   the two moments where we had the most   user base increase was when Jack joined and when Snowden joined. And when I dealt with   And well, okay. I mean, that that's a given.   But   but,   when   when people that,   like, the all those people figure out the keys, they figure out all the relays, all the hard stuff because they wanted to see what Jack was up to. They wanted to see what's what's Odell was up to, and they want to see what snow Snowdon was up to.   So in those sense, those are communities. It's social capital, basically. And when I say community, I don't know how about when I say communities, that's what I have in mind.  

[01:39:04] Hodlbod:

Yeah. I agree with that. Things that exist outside of Nostriland.  

[01:39:10] Mike Dilger:

Yeah. But if we had some really spicy content that was censored everywhere else, the only place to get it would be Nostril.  

[01:39:17] ODELL:

And, yeah, that's the same. People whole bunch of people will show up. I mean What is that? I don't know. But are you gonna tell Alex Jones to use small relays only, or does he use blaster?  

[01:39:28] Pablo:

No. No. No. No. No. That's the whole point. Alex Jones can if with outbox, Alex Jones can run the relay.alexjones.com.   He knows he's not going to be censored because he runs it,   and people can find it without having to do anything at all. He cannot be shadow banned. At with the current model, he can.  

[01:39:46] ODELL:

Without box, he can't. That's that's the whole thing. Okay. And Prime was Prime was gonna implement this. Right?  

[01:39:53] Mike Dilger:

Yeah. Yeah. Yeah. Yeah. So we're good. E even people that don't We we don't always moving in the right direction. 100%.   You know, I could be in a conversation and and on some thread, and then Alex Jones could reply to me like it's gonna happen. But,   and then I could go, oh, Alex Jones. Monday, Mike. Monday. Click the follow button, and I have no idea that he's on his own small relay. I just now I'm starting to follow Alex Jones. But it never happened that way. We go the other way. I don't know. No. Let's be honest. Alex Jones is gonna download Domus.  

[01:40:23] ODELL:

He's gonna post directly to the Domus relay, and then there's gonna be, like, 3 weeks where he tries to figure out his lightning address and then set up his lightning address.   He's never gonna run his own relay, and it's gonna be   great.  

[01:40:36] Hodlbod:

Everything will work out. And then jb 55 will deplatform  

[01:40:40] ODELL:

Alex Jones. And then we'll finally get some robust   substantive changes,   and we can have a follow-up conversation at that point.   There we go. I look forward to it. See, I think we solved it. I think we got some homework.  

[01:40:56] Pablo:

Weird. I think I think the we we moved the the overturn. Like, we we've   started talking about this a little bit, like, a a lot,   and at least people are aware that this is a thing.   And Primal, to to me, the most important part is that Primal is going to start   doing,   is one of the biggest clients and is gonna start doing outbox, which I think is fantastic.   They are going to start leaving relay hints, which I think is fantastic. Random relays for new users instead of the same relay list. Yeah. Yeah. I   yeah. Yeah. So the idea is to use NIP 66, so nostril dot watch,   to get relays that are fast, publicly accessible, and fast and reliable   close to the person where the person that is is using Primal is.  

And they write to write relay primal.net   and also to these, whatever, 5, 4, 6 different relays.   I think it's fantastic. I I think that's gonna be a big, big, big step.  

[01:41:56] Fiatjaf:

Well, Demos has already   said they will implement it,   many times, although I don't know what exactly would be the would the implementation look like.   So   maybe this discussion was for nothing. Everybody's already agreeing.  

[01:42:13] Mike Dilger:

And no one understood any Everybody here already kind of agreed, so we didn't we didn't get any we didn't bring in Tony or  

[01:42:20] ODELL:

or Well, I mean, to be clear to be clear, my relay list is Not that they disagreed, but My my relay list is only one small relay. It's relay.bitcoinpark.com.   Then I have big relays. I have relay dot pyramid fiatjaf or whatever that relay is. I have Nostra Band. I have Primal.   I have Snort Social, and then I have blaster, and I just blast everything.   So I'm the disagreeing one.   You're here to learn from the I know. I just want I I want my notes to be everywhere. Anywhere that'll accept my notes can have my notes.   What did Jack say? Jack said I don't wanna be a part of a relay that'll have me. I wanna be a part of every relay that'll have me.  

Except for the blue check ones. No blue checks. I will not broadcast to x.com. I haven't since August.   There you go. The biggest relay.   I got nothing else to talk about. Dave, did you guys have any final thoughts?  

[01:43:17] Pablo:

I I have one that I think the conversation around this has been very centric around   these, like, outbox against cluster.   It's 100%   not against blaster,   and I think there is a I was running blaster until last week. Like, I had my own blaster instance. It was doing more numbers than what I see totally posted. I would it was doing, like, a half self hosted your own blaster instance? You can't self host because it runs on Cloudflare. So all blaster runs on Cloudflare. But I was running a a blaster on blaster dot, f 7 zed.xyzed.   So it it's very very much not,   Outbox versus   versus, blaster. I think part of the misunderstanding   is that there is this tension between the 2,   but there there isn't. It it doesn't matter if you're using plaster.  

Things are still,   broken in some way because plaster is doing the right part, but not the rip part. Right? So, for example,  

[01:44:19] ODELL:

if Jack now is using these, 4 small relays Pablo, it's not your relay is not a small relay, but continue.  

[01:44:26] Pablo:

My relay is small, man. It's running on a Celeron. It's running on a Celeron computer   along with my fulcrum.   So Continue. I interrupted you. It's a very small relay. I will publish a picture later. You'll see how small it is. Oh my god.   Of the relay. She's   the,   the,   the the the,   the thing is people can can blast   Jack notes from this so someone can go in. They can read the notes from Jack, from these 4 small relays, and publish them via blaster, and they will end up in   300 different relays, and that's fine.   If Jack is using a a client that is not using Outbox,   he will not see responses   unless people send them to these very small relays, which could change,   they he will not see responses. So this is where, like, it's it's not   an an or it's not there is no tension. We need outbox regardless of blaster, and I think there is a case for blaster.  



[01:45:32] ODELL:

Yeah. I support the outbox  

[01:45:33] Mike Dilger:

model. I I think I just got that. I didn't realize that you you blasted the 300 relays.   On one of those relays, somebody replies onto that relay,   and that's not something your clients ever pulling from, so it never sees a reply.   Yeah. Exactly. On on demos, it looks I see, I'm not even familiar with the landscape and what it looks like outside of the outbox model because I've been using outbox model from day 1.   No. I'm But, yeah, that's that's not good.  

[01:46:00] Fiatjaf:

That's not good.  

[01:46:01] Hodlbod:

Yeah. I do think blaster has its use cases. Like, if if it limited itself to 10,000 twos, I'd be just so happy. That would be great. I mean Do the change you wanna see in the world. You can do that.  

[01:46:13] ODELL:

I'm not gonna run any relays. No public relays for me. Okay. Well, I mean, it's a public blast. It's Cloudflare is running it.  

[01:46:20] Hodlbod:

Yeah. Forget Cloudflare.   I like to do the things the hard way. I'm kinda curious. Like, so  

[01:46:28] ODELL:

I got pretty excited when I saw a Fiat Jeff's note. Right? And I was like, we need to run a dispatch.   I like, I think this is, like, actually just a noncontroversial   thing among most Nas like, either Nas users don't give a shit or they support the outbox model. Like, I feel like it's what that's what Mike said. Right? Like, I think,   this seems very noncontroversial.  

[01:46:50] Mike Dilger:

But I think Will says he's gonna implement it. Tony's not against the idea. He just that's just this what this it's it's a fake   it's a fake debate. It's not there's no real debate here. There's no debate.  

[01:47:03] Hodlbod:

Noster's lot of confusion.  

[01:47:05] ODELL:

But   there will be   there's gonna be wars in the future. Right? Like, people were we're gonna start disagreeing. Like, we're gonna get to a 100000 users, a 1000000 users, a 1000000000 users,   you know, a 100 clients, like, different developers around the world. Like, there's gonna be a bifurcation. Like, there will be   there will be NASDAQ protocol wars. Will there not?  

[01:47:28] Hodlbod:

Yeah. And that's, I mean, that's, like, that's the real threat is   some big company with that's well funded coming in and taking over the protocol. I think when that happens,   we'll probably just wanna say, okay. We're done advancing the protocol. It is what it is. No one can add anything. Oh, you wanna you're for Nostril protocol ossification.   Not yet. But,   but when, when threads is done murdering,   Activity Pub,   then I will be for ossification.  

[01:47:57] ODELL:

I think it's kinda beautiful that it does it's like there's no consensus   there's no global consensus with Nastr. Like, I think,   like, if if a client falls out of sync with the global network, like, whatever happened to ZBD's client for, like, 6 months, who fuck gives a shit? Like, it'll it can eventually come back   or not or or different clients can evolve. Like,   I think kind of the beauty of Noster is the chaos element, which isn't it doesn't exist in Bitcoin because you have to have this global consensus.  

[01:48:26] Hodlbod:

Mhmm. Yeah. Pablo, say the thing.  

[01:48:29] Pablo:

Which one? There are many things. Embrace the chaos. Embrace the chaos. Yeah. Embrace all caps.  

[01:48:35] ODELL:

Always. Always.   I appreciate it. My, Nystratia keynote with that. I'm pretty excited for the Nostra Protocol Wars. Like, I would agree with Mike that this was a little bit of disappointing conversation because it's just there wasn't enough debate,   not enough on the line. No one's getting censored. Has anyone gotten censored yet, Anastar?  

[01:48:55] Mike Dilger:

I don't know if we can tell.   Only accidental is getting censored. It's it's hard to notice the dog not barking.  

[01:49:03] ODELL:

Yeah. I feel like I would notice if someone got censored. I don't think I've noticed yet. I mean, if if you don't notice, like, did they get censored?   Maybe. Yes. Probably.  

[01:49:12] Pablo:

That's probably the way It will look like they got shadow banned. It wouldn't look like they got censored.  

[01:49:17] ODELL:

So it's hard to tell. It's it's the lack of something. Right? I mean, this is where it does does matter, like, who the ride or die original nostril users are. Right? Because it's not just, like, I love our developers. Like, I I all of you guys here are doing great work, contributing to open source software on the nostril protocol, but, like, a lot of it does also come down to, like, who are the original users, who are the router dot users. And I think a lot of Nas users just wouldn't accept it. Like, if there was, like, one person that no one really gave a shit about   and they were like, I'm getting blocked,   I think even that right now would get significant   traction.  

I would blast if they did it, I would blast it out to all the relays.  

[01:49:59] Hodlbod:

Yep. Another way to look at it is everyone's getting shadow banned right now.  

[01:50:03] ODELL:

Why?   Because there's only because of all the stuff we've been talking about. There's only 10,000 users, so it's like, who the fuck is shit? Twitter is reading this stuff for me. I promise you people on Twitter are reading the shit.   They pretend they're not post something spicy.   You post something spicy, the blue checks realize right away.   I've done this test. My test was more successful than your test, Vijay.   Pablo knows exactly what test I did.   I literally I I sent a single note out that was, like, who rugs first, Swan or Binance, and Corey Klipschian was in my DMs in fucking 15 minutes.   Straight to the big relay.  



[01:50:52] Hodlbod:

Okay, gents. Alright. Here's my final thoughts.   Nostril will I yeah. Nostril will un gay the frogs.  

[01:50:58] ODELL:

We're gonna un gay the frogs. Tell Alex Jones. Pablo, final thoughts.  

[01:51:03] Pablo:

Can I do a shield?   Of course. Or the final thoughts?   Alright.   Very soft.   Apply for sovereign engineering.   We're running the 2nd cohort.   If you if you made it into the rest is to into the end of this,   podcast or livestream,   You 100%   belong. You are a 100% in.  

[01:51:27] ODELL:

Now Congratulations.   You've been accepted.  

[01:51:29] Pablo:

You you are fully accepted. Book your ticket.  

[01:51:34] ODELL:

Sovereign Engineering dot io. What is Sovereign Engineering? Can you actually give a proper Yeah. So yeah.  

[01:51:40] Pablo:

Good idea. Yeah. So Sovereign Engineering is a program that we're running with Gigi here in Madeira.   The idea is to bring 21 people,   that want to work on Freedom Tech, that they want to make a difference shipping something that matters. So you should be a developer. You should be able to ship stuff.   We also have a few designers,   but it it's like you should be someone that can ship their own their own thing. And   it it's great. We did the first the first cohort.   It was for 2 months from, January to March.   We got a bunch of shit done.  

Npad.cash came out of it. Blossom came out of it.   A bunch of really exciting things,   confidential DBMs.   A a bunch of really interesting things came out and will be coming out. It's a really good opportunity to to   just bond with people that have the same values as you do. So, yeah, consider it.  

[01:52:36] ODELL:

Yeah. I mean, I would just wholeheartedly recommend Sovereign Engineering.   You guys are crushing it over there. And   if you're a young aspiring developer and you love Nostra and you love Bitcoin, you should definitely   definitely check it out. And you made it this long, so you already accepted.   Fia, Jeff, final thoughts.  

[01:52:56] Fiatjaf:

I don't have any.   I hope people enjoy the games.  

[01:53:01] ODELL:

I do. I really did. If you have no thoughts, you can't get censored.  

[01:53:05] Fiatjaf:

Exactly.  

[01:53:09] ODELL:

Mike. Mike,   thanks for joining us. Final thoughts. I don't have any final thoughts either. I don't I don't know. Mike, like, completely regret he's like, I shouldn't have joined the podcast. I shouldn't develop on Nostril. Like, what the fuck am I doing with my life? So just post the truth social.  

[01:53:25] Mike Dilger:

Oh, no. No. No. No. No. I love Nostril. I I love Nostril. I'm I'm I love where I'm at. I love that I can do this for a job. That's fucking amazing.   So,  

[01:53:35] ODELL:

let's fucking go. Well, I appreciate you all. I appreciate all the freaks who joined us in the live chat,   even though there were some technical difficulties.   We're working through them. We're improving, proving over time.   As always,   sale dispatch is a 100% audience funded. We have no ads. We have no sponsors.   We're here purely to serve you,   and   I'm incredibly grateful for everyone who sends Bitcoin to the show to to support the show and and and keep   you don't really keep the lights on. Like, I I pay out of pocket for the lights, but but thank you for your support. It it does mean a lot.   Easiest way to support the show is to subscribe on your favorite podcast app. Share it with friends and family. Leave a review.  

Ideally, a good review, but, you know, you can leave whatever kind of review you want. And, podcasting 2 point o apps, Fountain Breeze,   I do read all the comments.   Thank you if you left a comment in this app.   I appreciate you all.   Guys, you guys have been fucking crushing it. Thank you for building out Nostr and allowing me to shitpost without limits.   It truly is beautiful. I couldn't I I   all my shitposts were would not be possible without you guys.   So thank you. Thank you for making dreams a reality.  

[01:54:55] Fiatjaf:

Yeah.   Thank you for shaming the Twitter people   all day.  

[01:55:00] ODELL:

I'm gonna continue I'm gonna continue to do that. That's that's my role,   and it's beautiful. So thank you, guys.   Maybe in, like,   4 months, 5 months, let's bring the whole crew back on. You know? Hopefully, someone gets censored at that point.   Hopefully, we, like, we have our Alex Jones model moment. Maybe we have, like,   by the way, I'm using Alex Jones as a example,   not him specifically. I mean, maybe he specifically joins.   But, yeah, it'd be cool to   reconvene and and do another work call and and and see what the flex going on.   Cheers.  



[01:55:42] Pablo:

Cool. Thanks for coming out, man.  

[01:59:56] ODELL:

Love you, freaks.   Keep it spicy.   I didn't think Pablo realizes that he's still in the stream.   I wasn't.   I I didn't. Pablo, thank you.   Freaks, I got the CEO of Phoenix Wallet coming in on Tuesday, 1700 UTC,   And then I have Tony,   the creator of blaster and mutiny wallet coming in the week after that. Also, Tuesday, 1700 UTC.   Might add something in between there. Not sure yet. Love you all. Stay humble, stack stats. Fuck the suits.