Transcript (free, so it's mostly right)

00:01

Hello. how you doing? I'm doing pretty good. How you doing? Well, you're making me a little nervous there. Sorry. I spaced out in the shower, I think. I was like, what? How long did I take? OK. Let's see. Your audio sounds good through Zoom, so the rest of this should be fine. Let me switch over to the main view.

00:31

We already have a couple people like hanging out in the chat at the stream place stream. Oh gosh, that waveform needs a color.

00:52

All right. So the only, well, since we've got Sim and Bailey already in chat, I need you to talk about something for 30 seconds so they can come and take you. OK. I am in Seattle. It's kind of overcast today. I'm hoping that it gets warm again. I'm not really ready to give up on summer.

01:20

Summer starts pretty late in Seattle, um and it's like perfect. But I don't know. I feel like some of the climate change stuff starts a little bit earlier. Sometimes August, we get summer. Sometimes we don't. And it seems kind of touch and go. Are we going to get an early fall? And then I was gone for two weeks, and so I feel like I missed it if this was the end. That would really suck.

01:53

I New York. They say you sound good. All right, cool. Because I had to, like I said, I was more concerned about routing your audio in than anything else. I was going the route of Zoom and I'm using my dad's account and it's supposed to have RTMP out, which I had tested last night to a certain extent and I thought it was fine. But then when I started to call up this morning, it was like, no, I just want to do Twitch. And I was like, I haven't logged in to Twitch on this laptop. What are you talking about?

02:24

Anywho, I'm piping all of this through OBS and then back out to Streamplace. And since Bluecast has a 60-minute limit, I am going to smash start streaming now. ah I'm worried, though, it's only going to capture the audio from my AirPods. But we'll see what happens. Like I said, I've got it both recorded locally and in Zoom, and it'll all be for edits later.

02:54

And ah we are good for all the things that need to go. We've never talked before. Hi. Hi. It's so nice to talk to you. I feel like we talk all the time on Blue Sky. Well, I'm always like blasting off a billion posts and you're like, look, man, I'm just trying to do some work and just occasionally pop in.

03:23

Yes, I go through uh stretches of poster madness, and then I get back to work. I think that's probably a healthy balance, Because you shouldn't be online all hours, all day long. You should be like, oh, I went outside, and then I remembered to check in on everybody two days later. That's fine. Exactly. That's how I justify it to myself when I do it. I'm like, oh, I could be working right now, but I'm like, yeah, this is part of the balance.

03:51

Ray, you gotta take a breather every now and All right, are you ready for our warmup exercises? Yeah, let's do it. Okay, they are yes or no questions. Okay, yes. Don't overthink them. I've asked questions like this to other people before and then they tend to go, well, and I'm like, no, no, it's just a yes or no. You don't have to think too much. It's just yes or no. All right, do you have a pet? No.

04:20

Days off are for sleeping in. Yes. Have you traveled to more than one country? Yes. Are birds real? Yes. Do you have posters madness? No. Are you moots with someone that has posters madness? Yes. Did you go to the inaugural atmosphere comp? Yes. Have you worn butterfly wings for a video call? No. Do you keep a journal or notebook?

04:50

Not anymore. you ever intentionally gone off grid for more than 24 hours? No. No. OK. All right. Do you feel good about all your answers? As good as I can about not having worn butterfly wings to a Zoom call before. Hey, I'm just as guilty. I debate it all the time. I'm like, man, if Paul can pull that off, I bet you I can wear butterfly wings.

05:17

It just takes a degree of planning that I usually do for calls. Maybe for Halloween, though, the butterfly wants something. no, no, no. Halloween, then it's just basic. You're just being Halloween fun. You need to pull it off on February. And when nobody expects it, that's how you do it. That is true. That's more fun. There was one guy at my high school who would just occasionally wear Superman outfit.

05:48

And everybody loved it. It was completely random, but just occasional enough that we kind of forgot that he did it. And then, boom, show up dressed as Superman. Right. Smart, smart. Because you don't want to do it so often. People are like, is that guy weird? Whereas if you do it occasionally, then it's like, oh, that's fun. And then you forget about it, like you said. And then, oh, that's fun again. Yeah. It's like he's keeping us on our toes.

06:16

Well, we've got a number of people that have popped in the chat and they're all saying hi. So we got Tyon, we got Nick, we got Sri. Bailey's just talking about having a mood. I think he's talking about the just popping up a Superman. So we got a surprisingly large crowd for the first call. I expected like two people to show up at the top, but that's okay. All right. Pressure's on. Pressure's on. So introductions.

06:46

I'm Patrick. I'm just some dude on the internet. And I just like to post a bunch of stuff. And I try to be positive and supportive of devs because in my real job, they drive me crazy. And they make me go slightly insane. And they don't think about security the way that I want them to because I work in the security field. ah And I reached out and general and said, hey, wants to do this? And Reed was the first one to say, yeah, I'll do it. Not knowing

07:16

much about me other than I'm a poster. All right, read, introduce yourself, give yourself five words to describe yourself. I'm the CTO of Skylight. think I will need four. of Skylight. Fine. All right, that's five. That's good. All right, why did you agree to join the call? I think that we want to talk to the rest of that pro community about what we're doing, and especially that the developers

07:44

To some degree, we were pretty heads down, especially when we started, and kind of maybe allowed there to be this gap between us and everybody else. um So we definitely want to do any kind of uh talking with other proto-builders or people interested in that proto-building. uh That was what was so great about going to the conference, because we actually got to meet everybody. And they're like, you should join this Discord. Why are you not in this Discord?

08:14

when you're working on this app. Everybody's wondering where you are. And so, yeah, I think that that's a major reason. And because I do think that there are some things that we're doing and the way that we think about building apps and stuff on that proto that's a little bit different from some other people or from a lot of other people. And that it's a nice balance.

08:44

Well, I would tend to agree because I've fooled with a majority of the iOS apps since I only have iOS. And the two things, I am a weirdo when it comes to use case, the things that I enjoyed about Skylight was the video editor was really good. And it didn't ask to save anything to my local storage, nor did it record to my local storage because I don't like things recording to my local storage. So was very happy about all that, which is why I posted a whole bunch of junk from last weekend.

09:15

from when I went to Los Angeles and that was a lot of fun and people were like, hey, what did you do all that? And I was like, and this brand new app that's brand new. And they're like, how brand new? I'm like, like Munsell, not even a year. And they're like, wow. And they're like, this is the new TikTok. And I go, sure, it's the new TikTok. Cause I didn't want to get into a technical conversation. Yeah. It is the easiest thing to say. You just go, yeah, yeah. Just like TikTok, but you know, not TikTok.

09:44

Okay. Probably the best one of those is just like, it's Mark Cuban's unbanable TikTok. Wow. So you're just going to jump right into it. Mark is a catalyzing force on the Blue Sky app. Because many people are like, no billionaires, and then here you have a billionaire. So how's that working out with working with Mark? he just ceded money and wants

10:13

occasional updates or is it more involved in that? It's definitely more the first one. We have asked him his thoughts about things and he'll give advice, but that's pretty much on our terms. He's a lot going on too. um I think his idea was like, like the idea of there's obviously some disruption around the TikTok stuff. I'm totally willing to put in some money to get somebody started building an alternative.

10:43

Like, that's good for us. I think that I'll talk about it a little bit later, but yeah, one piece of advice that kind of stuck out and came up when we were thinking about Skylight V2.

10:56

Belly B2, are you already talking about B2? Yeah, why not? All right. All right. All right. Let's go with another rando question.

11:10

This goes into how long have you been on Blue Sky? And everybody has been, whether they wanted to or not, subjected to some of the lore of Blue Sky. So what lore have you been part of, cultural or technical? I probably later than a lot of other people or the people in this chat, came in, like I think of it as like there are the three waves, there are like the two that didn't really grab, like hold on to people in the third one that did around the election. And I think the second wave I joined

11:40

But it didn't quite do it for me, but I was pretty frustrated with X. And then the third one, I just switched over. I was so done with X after the election. Wow, you hard switched over. Yeah. And I'm happy I did. I was like, I have a better time on here than I do on X. And that was before we ever thought of doing Skylight or anything. I think the first time I checked it out was more because I saw Dan's

12:10

talk over the summer, Dan Abramov's talk over the summer about the web without walls. And I was like, this seems really interesting. So I kind of joined the first time for that. But it was like I couldn't find the content I was looking for. then it kind of faded out. But then afterwards, there were enough people that it was kind of like, OK, cool. I can hang here. And I don't have to see some of the toxicity and self-promotion that's on X. OK, so it's interesting that.

12:38

one, your experience started strictly as a, I'm getting off of X and doing something else. And then you said that you couldn't find the content you wanted. So pivoting off of that and with the fact that you're working on an app, helping surface content. ah How do you feel about finding stuff now? Like how much is involved? um With Blue Sky or with Skylight? Well, let's just say AT Protocol.

13:07

Right. Because they both had similar challenges, right? Which is people are making stuff and they're putting it out there. And then how do you surface that? And I think both app views have a similar challenge, How do you find the things that you want to find? Yeah. think that's ultimately, I guess I'll even take it one step further back is that when we think of a social media app and if we kind of do the hierarchy of needs thing,

13:37

Like the very foundational part of that, like the very bottom of the pyramid or whatever, is the IC content that I like or that's relevant to me. And it's like that's the thing you have to get right. And then you can worry about other features. uh that's at least my philosophy on it. And there's more than one version of what right looks like. It just kind of has to be right for some group of people.

14:07

One of the things that we've been thinking a lot about is what did write look like for the early adopters of these other social media apps? Because obviously Facebook, the right thing to show people was who was in there, like people from their college in the posts about those people, not people from other colleges. And then the right thing for a Reddit user was very topical. It was because they're in the right subreddit. Like whatever the conversation is, has to be appropriate to that subreddit.

14:36

And then Twitter was based off of people and the following people, and that they show up in your feed and stuff. And all those are just valid versions of right, and they appeal to different people. uh But the app did facilitate on kind of an app-view way, like how the app is structured and designed, and stuff did facilitate those kind of sweet spots.

15:01

Okay. Okay. I see where you're coming from. That might've been a little much, like... No, no. was great. That was great context because you talked about how really, if we think about things in terms of AT Proto, a lot of social media, however the content is, it really does come through how you view it and interact with it and what you're looking for in that context. And you covered some really good context points.

15:29

And if we think of AT Proto with each application being its own app view, then feeds become another tangible part of that in the sense of how do you formulate the right feed? And then if everybody can consume a feed, it still matters on how it's presented because of how the app view works, right? So for example, I imagine you're working on the algorithm for Skylight and you're going, this is how our feed is generated and piped through to people, but then you would

15:59

said earlier, I think it was a post on Blue Sky that you said, no, really one of the things that we've just discovered slash rediscovered was that there are still content curators out there and they're doing that. And that's where the feed paradigm of AT Protocol kind of works in with that as well. So how are things going with you on the feed front with Skylight and feeds for other people in general? Yeah. So that's where I think that

16:27

Like Skylight V2 is in some ways a pretty big departure from V1. And it's part of this overall strategy. V1 was like, we didn't really put a lot of thought into what we wanted to do differently than TikTok. It was like, hey, TikTok might get banned and people might need a place to go. So we're going to try to be as similar to TikTok as possible. Basically be like a life raft for people. was kind of like our original intent is that just because this app gets banned doesn't mean the social network that had like

16:57

grown on it had to get banned, that there was kind of safe harbor on app protocol. And so from that perspective, was, OK, well, TikTok works like this. has like, you just swipe up and it shows you different videos. And the only way to really make that work is to use machine learning to try to recommend people videos. And so that was the product approach, essentially, that we took for building Skylight V1.

17:24

But that didn't work very well. And what we kind of figured out was that the machine learning algorithm is more of an accelerator of some social network that's already happening without it. And because we started with the machine learning algorithm, it actually had nothing to really accelerate. So the videos that it would show you were just

17:50

the same videos as if we showed you the most popular videos or the most well-liked, which is not at all what it was supposed to do. And so that was part of what led to my Blue Sky post, was basically like, we're not going to focus so much on the machine learning algorithm. We're kind of just going to turn it off. And that's what we're doing. We've already actually done that. um Because it was doing more harm than good.

18:18

Instead, we want to build a social network that works without it. And then we could add some kind of machine learning in to help that be a little smoother. But that's for this idea of curators comes along. And custom feeds is a really high touch way of doing curation. Obviously, you could programmatically determine everything about the videos that show up. But

18:42

There's also just reposts and stuff like that, these ways that people can say, hey, I like this video, and think the people who follow me will like this video too. And trying to then surface the curators of good videos so that people could follow them instead of only following the people who post the video is kind of what Skylight V2 is all about. So you find the people who have similar tastes to you.

19:10

You follow them, and then you'll see the videos that you like. So that's how we kind of fulfill the bottom of the pyramid of the hierarchy of needs or whatever. Like, is that we don't need an algorithm because we're kind of, we're leveraging the fact that people are already watching these videos and they have taste and they have like, uh probably an interest in watching more videos and being somebody who is a tastemaker.

19:38

and just allowing them to actually do that on our platform in a way that TikTok and Instagram and stuff don't. And so that's the strategy. so feeds are one way of doing that. Reposts are another way of doing that. We almost think of you probably get started by reposting. And then at a certain point, you're like, oh, I actually want like, now that I know what I want out of a feed, I know that I should actually go make a feed on Graze or something. And then that will be the feed that I kind of

20:07

push everybody to for some given topic. let me see if I can restate this and make sure I understand. You feel that with the first iteration, ML was really dependent upon just being a lot of volume. It had to start with somewhere, and then people would give it inputs. And now you're kind of going, all right, we

20:34

we really kind of need to focus on what early internet was for social media and go, no, we really got to depend on people's ability to find people that they like, that like things that they like. And then that becomes the way that they can find more stuff that they like, because they said, I'm going to let these people share what they are into. And I can say, oh yeah, I liked 70 % of that as well. And that's kind of what's going to drive things going forward. Is that right? Exactly.

21:02

I think that that's like the, what we learned is that like, it's not a coincidence that all these apps started this way and then added like algorithms on top of, on top of things. And I think of TikTok. TikTok's the one where in hindsight, I think everybody believes it's always had an algorithm. And I think we believe that too. But what we kind of figured out was that before they had an algorithm,

21:31

The way that they were matchmaking people with similar videos to their interests was just using the music that played. Because every song had to have music, you could almost think of that song as like its own subreddit and the posts being a part of that subreddit. And when you would listen to uh or like a video about a certain song, they would just show you more videos with that song on it. And people would self-select certain kinds of content into certain songs.

21:59

And that was the organic social network that developed. And then they built an algorithm on top of that. And that's kind of like, so there's all these different ways to do it, but you do have to give people the ability to do it manually, really, at first. And it has to be satisfying and work well as a manual experience before it could ever be kind of streamlined.

22:27

Okay, it's funny that you touched on the music topic, because that's honestly been the only reason I've even tried to maintain some interest in a lot of social media is that I love music discovery. love finding new types of music. And I am more importantly enjoy it when somebody goes, I love this type of music. And this is why and it doesn't have to be something that I would be into right? Like a good example is a mutuals with Daniel.

22:57

and let config and I would say a lot of his music isn't in the genre that I would search out. But when he recommends something, I value his input and I will listen to stuff. And he shared a playlist recently and I was like, all this is good. I've discovered something and I don't know that I ever would have found this on my own. So we rely on other people to help us with our tastes. And we are becoming a taste maker while letting other people.

23:26

be tastemakers for us. And I think, you know, I never thought about it, but the only reason I've ever cared about using TikTok in any way whatsoever is it's what helps me discover new artists and new bands because one, to be relevant, they got to be there. But two, I've trained TikTok enough where I hate everything other than finding new music from relatively new artists, right? Like people who are less than 10,000 followers.

23:53

Those are the ones that I'm interested in and we'll go like, yeah, I'm going to check this out. And of course with three minutes, you can only give me your hook or your, part of your track. That's just a real banger. Of course that's going to get me drawn in and I might give it more of a listen out because you know, if you're flipping through songs on whatever service you're on, people tend to focus on the beginning of a track rather than the middle of a track to determine whether or not they're going to give it a chance. And I always try to skip to the middle because that's when you've gotten to the real.

24:20

bulk of what the music is. Anyways, I digress. I digress. So, um, getting back on track here, which is the main reason I wrote this doc was so that I wouldn't wander off on a crazy ass tangent. Um, so let's see, we've already, we've already kind of talked about your product to a certain extent. Um, but let's, let's go, let's double back to the win. When was the point where like you and Tory were both like,

24:48

We need to make this happen. it explicitly the impending TikTok ban that never really happened? Or was it more of you said, oh, I want to do this and Torey's like, I want to do it too. And then the two of you kind of jelled. Like, how did that kind of play out on? When did you get the idea that you were going to do this? I think Torey had the idea first because she kind of mentioned it to me. And I was the one who kind of said no. And this was in like September or not September, December of last year.

25:17

And she was kind of like, you know, we could build this TikTok thing. what if it gets banned? And I'm like, it's not getting banned. And then, like, I think I was just kind of burnt out from previous stuff we were doing. And so I kind of wasn't in the mindset of doing a startup at all. And so then we get to like,

25:44

the week of the TikTok ban and the place turns somber. And everybody's like, this is actually going to get banned. And I felt so incredibly guilty because it was like, I knew that we could do it. I knew that we would do it on the app protocol and that exactly how we would have done it. um just the idea that it was like, this is this

26:13

could actually happen. And then all these people are basically just going to get assimilated by these existing apps. And all of the stuff on TikTok is basically gone. Thanos snapped out of existence. And then that was the same time when we saw the Mark Cuban video where he's like, I'm going to fund somebody who does a TikTok alternative. so that also then made it safer.

26:43

Really? Before it was social is expensive. We have to pay ourselves to do it. It's not like it makes any money in the short term. And so for Bart, he would basically just say, hey, I'm just going to pull the trigger on something. Whoever's actually executing on this stuff. And it's like, I knew we could execute. And I was like, I guess that's better than a job. If I was just going to go get a job to make some money, I could just have this be my job.

27:13

This is a job that's like somebody should do and for whatever reason, nobody did in like the months leading up to the TikTok ban, which I will never understand, but I guess I'd be a hypocrite because I didn't do it either. somebody just had to do it, had to do it well, I guess as a motivator. I don't think you should beat yourself up too hard. mean, yeah, there was a point where like, I don't know that I want to do it.

27:40

but you're actually doing it. And that is literally the dividing line, isn't it? Yeah. And that was, guess, the thing where it was kind of like, I almost got a second chance to do it in a way because it just came back around and gave me the push that I didn't give myself the first time. You didn't have a choice. It chose you, huh? Yeah, kind of. think so. Probably as much as anything else in my life this ever chose me. Wow.

28:08

That's pretty, that's kind of powerful, actually. And Tori was there with like, we've got to do this. And you're like, I don't know. And she's like, no, we really got to do this. And you're like, I don't know. Kind of, I think she was like, we could do this, right? Like, I didn't know that she was like, we have to do this, but she, she definitely wanted to do it first. Wow. Okay. Okay. Let's, let's hit some of these questions. Sim has a lot of questions in chat.

28:36

And I think you've answered some of them, but we're gonna, I'm gonna scan through them real fast to make sure we kind of covered it all. So, Sim was worried about TikTok relying on an algorithm that feeds toxic addictive content. So do you have any thoughts on perhaps departing from that model with Skylight differentiating on that front? you kind of answered that question, but I think this is good context, right? Like it's a lot of people are very concerned with what algorithms are feeding them.

29:06

How are you avoiding that? And they should be. And I think that that's one of the things that the whole time, basically, when after TikTok got unbanned, that people were still really interested in Skylight, that that's the big reason why, if there's some unifying reason, is that these apps are not trustworthy. They're not trustworthy enough for how much they require your trust to show you content.

29:36

in many ways, the only game in town for an experience like that. And so we've kind of nailed down on a guiding principle for ourselves, which is just satisfaction over addiction. If people come back, it's because they actually find it satisfying, not because they're addicted to it. And that when we make features, we should be choosing things.

30:01

because we think that they're actually satisfying and not just, again, that they drive engagement numbers by getting people hooked and not conflating engagement with satisfaction. uh And so with this curators thing, this was actually something that before we had the principle, we were already acting on the principle because we did like that it was not based off of this algorithm that was maximizing for watch time, which is what basically all these algorithms do. They're like.

30:29

I'm trying to maximize the odds that you watch these videos for like eight seconds. So I'm going to try to show you videos that you'd watch for eight seconds. But there's so many ways to game that and so many ways to abuse that system. And now it's even worse with AI content, which can look real. It looks like it's a real news report or something, but it's not. And people are financially incentivized to make content like that on these other platforms versus on Skylight.

30:58

with this curator system where it's like you've got this level of which there's a person involved who's maybe seeing this new content and can be more skeptical. They know they're looking at potentially AI-generated content or something, and they're sifting through to find the gems, or just sifting through new content to find the gems. And so you can trust them much more than you can trust this algorithm, essentially. And I think.

31:24

You also then get a lot more explicit control over whether or not something shows up in your feed and why.

31:32

Along the lines of that, because it's funny you say this, and I think back to uh legacy media. And they said that there's an important part of the editorial process. that's writers go and find the story, they write the story, and then the editor adds the additional context of like, did you go way off base? Are you staying within the scope? And is anybody actually going to care about this?

32:01

you have to publish the news that people are going to be invested in. you, yeah, sometimes you have to publish some news that people like, I don't know that I care about that, but there is, there's an audience out there and there's always that editorial oversight with traditional legacy media. with curated content, sounds like it's a similar model in the sense of, yeah, we're going to have people that are going to make stuff and then we're going to have other people that are going to be like, yeah, I'm going to at least weed out all the bad junk.

32:30

so we can get to the better stuff. Yeah. And I would say that that's true, except for the fact that it's community-driven instead. So anybody can be a curator, and essentially, the audience is selecting who their curators are. OK. So it's in case that wasn't clear. So it's like, you could be a curator. So it is kind of tying back to the feed, people crafting their own feeds, but without the complexity of having to actually create a feed to do that type of stuff.

33:01

Yeah, and so there's the easy version, then there's like, you can make the custom feeds. We might even try to make like an in-between, like a playlist or something. uh But it's ultimately about like, those things get surfaced. Feeds get surfaced in a way that they're really easy to discover and follow, and that the people reposting videos are surfaced in a way that makes them easy to discover and follow ah instead of just the person who authored the post.

33:30

you actually have the opportunity and kind of the context clue to follow those people and have them like kind of determine your the content that shows up on your screen. That's great. I like that because I think the number one thing I run into because I intentionally go and talk to new people on Blue Sky is there are so many people that not only do they think that Blue Sky is a Twitter clan or they think that Skylight is a TikTok clan.

33:59

the concept of feeds just kind of blows their mind, right? Like even though Twitter kind of had some of that to a certain extent, they didn't have it to way it was intentionally designed like feeds are. And your point is more like, well, there needs to be a step before feeds, right? Because before you get invested into being a power user and wanting to craft a feed, you still want to go like, no, I want to curate a user list of these people that I trust to post relatively good stuff.

34:27

or like stuff that I'm gonna like, right? And if I have, let's say 20 people on that user list, then I know that like at least half of the time I'm gonna be happy with the amount of things that they've reposted or liked or shows up in that user list feed. And then later you can get into like, oh no, I wanna get fancier. I might actually wanna create my own feed. So I'm totally on board with your idea of curators and having community members be curators and going.

34:55

Oh, my tastes are similar to this person. That's going to work out great. Yeah. think that I'm honestly really excited too, because especially for short form video, I don't know that anything like that exists. So I'm sure that there's like some people out there, some kinds of content that are essentially just like not being served by TikTok or Instagram Reels and could actually work on a platform like Skylight and find their audience and sustain an audience.

35:25

And that's a really compelling thing, too. It gives Skylight a unique advantage. It doesn't completely depend on TikTok maybe messing up at all. But also the idea that these people just have basically been screwed so far. They just have had no luck with how these platforms have been built, and they might finally get a shot. And I bet that's really satisfying content that people don't necessarily always want to watch, but they always want to have the choice to watch.

35:54

And so it would do poorly on a watch time maximizing algorithm, but would do well on a follow-based system. Yeah, it's an empowerment that we haven't seen really before, It's that you can make this as complicated as you want to, or you don't. The choice is still ultimately yours. You can let somebody else feed it to you, or you can be like, no, I kind of want to have my own say in how this works out.

36:24

And then I'm going to let these people be my curators of things, right? It's like why people prefer 30 years ago, one newspaper over another or one magazine over another, right? Because they're like, oh, I like the voice that they have and the kinds of things that they push out. And yeah, they'll introduce me to new stuff, but it's also familiar at the same time. All right. got a, I got a chat. It's been kind of really active for the last two minutes. let's let me, scan through that real fast. Okay. Well, let's see. Um,

36:53

Sim, well that's a hot button topic because I know how music licensing works. So we'll double back to that one. Let's see. um

37:07

Bailey, Bailey was like, he really likes how feeds is how he sells blue sky to people in real life, you know, just create your own timeline. So you know, he was just commenting that like, that ends up being a good incentive. uh Sim was saying, well, you know, not everybody wants to carry their own feed. So the, idea of incentivizing creators, you know, back to your curation comment and custom feed makers, totally on board with that. uh Sims like

37:37

Also, that's a important piece of the creator economy, which I don't know, I'm torn on the creator economy, right? Because like you said earlier, sometimes people are more interested in playing to the algorithm than whether or not the content is quality. And when people complain about AI slop, I'm like, man, there's been so much social media slop for such a long time. I don't even care about the AI slop because I don't care to watch most of it. Well, I think that's a worthwhile one to touch on because at least

38:07

from the, because I'm sure some questions are about like, oh, how are you going to make money or like the creator economy as a whole? I know a lot of creators have that question. To answer your part about like uh almost like is the creator economy a bad thing? And I kind of think it's probably like my guess. And so we'll have to like see if this is true is that it's bad on bad platforms with bad incentives. So. oh

38:35

I think it's more about managing what gets you paid versus whether or not somebody gets paid. So TikTok rewards certain kinds of content engagement bait and stuff that are really, I think, toxic to the end user. A lot of manipulation, grifting stuff is very successful. But I don't know that that has to be the case with social media.

39:04

One of the ways is like this curator way. We're hoping to add other things over time to like kind of work in the same direction of satisfaction over addiction and try to get to a point where it's like people are making money when they like are really providing kind of like the most real value to people. So like a good example is like uh if I follow somebody on YouTube and I've watched like plenty of their videos and then they recommend

39:34

Somebody recommended Affinity, which is an Adobe uh competitor, for video editing and stuff. And that's a really worthwhile recommendation. And it's highly likely to convert me. It's highly likely that I'm actually going to be satisfied with the product. It's not very likely that he's going to essentially manipulate me in that situation. Because he has no. oh

40:03

incentive to do that. He would just lose his audience. And so I think it's about balancing these incentives, where it's like, if you're just getting paid because you're getting views, that could be a really dangerous thing, because all it does is incentivize essentially tricking people.

40:21

Yeah, which is, you know, like, let's just pick the biggest creator on YouTube, right? Like, Mr. Beast is all of his videos are explicitly about engaging a viewer. I want to watch. I know that I'm going to want to keep watching. I'm going to watch all the way to the end, right? Like, oh, this is excitement. And he's branching on to other things, but I don't know that I would trust Mr. Beast to recommend anything to me, honestly.

40:49

One, I'm not as demographic, but two, it's just like, man, you make really fun videos to watch. That doesn't indicate that I would trust you to do anything other than make fun videos. Yeah. But then there's that aspect to it where it's like, if it's like you're not getting necessarily paid for the view, but you're getting paid if you can make a meaningful recommendation, and you have an audience, but they don't necessarily trust you in that way, I think that's actually a success story for what a monetization story would be.

41:18

that you couldn't just hawk anything and people would buy it. They're buying it because in the overall grand scheme of things, your recommendation actually does carry weight for them and probably should to some degree with some of these things, some of these people. So along those lines, have I paralleled to things that exist today and more importantly missed opportunities for those things that exist today?

41:48

Twitch is a subsidiary of Amazon and each Twitch channel is about, we do whatever thing we are. We build up some community. You know, for the core couple of people that you've got involved in your community, you actually carry a lot of sway with them, right? You've built trust with them over time. And when you're like, I don't use this product because I've had these problems in the past. I'm not trying to say it's a bad thing. I'm just saying my personal experience, but I do use this.

42:17

And then you have an Amazon affiliate link and you're transparent on that. Like I use this, I recommend it. And if you are inclined to buy it, I have an affiliate link in the bottom. Then you are, like you said, staking some monetization on the fact of like, no, I'm mostly focused on making good content and having engaging community. And then if one of the ways that they want to contribute to me doing things like this still is buying something that I recommend and both parties kind of feel good about it. Right. Because yeah.

42:47

The content creator's like, I didn't get there just to shill stuff. I'm making recommendations on something I use or have used that I really like. And the person's like, I want to support my person. And of course, I want to buy a thing that they recommend because that carries value to me. So I'm totally into it. Yeah. Yeah, and I think that that's the kind of thing that we're really hoping to optimize for and incentivize, that that's the way that creators are able to monetize on Skylight um and not so much

43:17

like engagement bait or anything like that. that makes me think of a question. And I don't think you're going to have the answer to this. might. Have you thought about with Skylight putting affiliate links in there? So I was editing a video early Sunday morning. And I put in the URL for the Twitch channel that the person normally streams from. And Skylight.

43:45

editor instantly converted that into a link and I was like, oh, that's cool. But then later when I post the video, I couldn't like really tap on the link for it to do anything. But it seems to me like their opportunity there in skylight is you can go, oh, this is the thing that I'm using. You put it in as a link. It's your affiliate link to Amazon. And then you can go, oh, here's this person who keeps pushing videos I like, and they happen to mention that this is the thing that they have.

44:12

and I can tap their affiliate link and I can support them and get the thing that I was like, oh, I totally want those pairs of socks or something. Is that the kind of thing that you're talking about and falling in the realm of what you're thinking about for Skylight? Yep. I think so. Yeah, that's exactly kind of what we're thinking about. And the feature to have a link that you can click on the video, I think that was just oversight on my part, to be honest. It should already have been working.

44:41

That'll be there in Skylight V2. Oh, okay. Yeah, it was one of those things where I'm editing a video while I'm there and I've been up for 24 hours at this point. I was like, don't know if I did it right or not. care. I just want to make sure I get this video published. Yeah, it should definitely be easier than that. Okay, let's see. Got some conversations about friction, better way to browse and subscribe to feeds. Yep, everybody.

45:10

Everywhere in the world is going to agree to that statement. uh Controlling distribution. I think a lot of people realize that. mean, there's been a lot of creators in the creator space, especially on YouTube. Like, I mean, when I think of Nebula and Curiosity Stream, it's about those people going like, I am tired of being at the whim of YouTube and ad monetization. I've got to take some control over my content to figure out how to publish it. Sim was very excited that

45:39

the of selling excite, the concept of creative selling content isn't bad, but we're incentivizing terrible content right now on platforms. So Simms, big fan of what you said. And I have no idea what Bailey is saying. This is a three minute highlight for Beast Guy, but I'll go through it later. right. um So we've talked about a lot of product.

46:08

And Bailey was one of the few people that had a pointed question about the technology. uh we'll kind of go right to that, right? So Bailey, it's a two-parter. So the first part was how hard was it to move from client side O-Log to backend O-Log? I kind of already have some inside baseball on that. And then the second one is how did you achieve that move?

46:33

And what makes the actual XRPC request to the PDS? Is it the client or Skylight's backend? Skylight's backend. I can say that. And it wasn't too hard to actually make the move, other than that there's some incompatibility between the app proto packages and Cloudflare. We use some settings in fetch that they don't support, and it just crashes and doesn't log a very easy to.

47:01

debug error. That was kind the only problem I had with it. And I could just say I used durable object. So every phone gets a kind of fingerprint to it. And then there's a durable object in Cloudflare that just matches that device. And it runs the OAuth client node package, which does all of the stuff behind the scenes. That's what they made it for, to make it a really easy way to adopt. And then all I have to do is uh

47:29

proxy all the XRPC requests through that durable object. Oh, OK. All right. And I'll probably release an example of that once Skylight V2 is out, because it's a pretty nice architecture that the one reason why I went with the durable objects and why I'm so big on Cloudflare is because when we launched Skylight V1, basically all of our servers got overloaded. And it's one of these things with social app where it's like,

47:58

If things start taking off, you're going to have a ton of people using it. A lot of apps you could just probably get away with just a single server or something. But I don't know that that's true in social anymore.

48:11

No, I mean, I think about the waves of people getting on the blue sky and how, you know, the team was very much like, we have the best problem in the world, which is people are killing us. We have, we have the signup rate that's almost something we weren't planning on for years. And our system, fortunately, was designed to handle it. But boy, were we really nervous. We're all sitting around looking at dashboards going like, is it going to crash? What are we going to do to spin something up?

48:40

So yeah, I totally get that. The term I've heard before is the hug of death, right? Which is people are very excited about a thing, and they're like, this sounds great. But if you have an entire crowd mob you to give you a hug, it's very hard to breathe. Yeah, and that was exactly what happened. then there's always this specter of like, know, TikTok could just get banned or it could get sold, and a bunch of people will just show up on our doorstep.

49:08

the day that that happens. Any times there's any news about TikTok and a potential ban, we get thousands of more people. Really? Yeah. And so it's so hard to predict, too, that it was kind of like I had to come up with something that would almost indefinitely scale. And that kind of overhead is actually probably one of the biggest problems for building Skylight.

49:35

in particular, it's just the potential that like millions of people could show up at any given time.

49:42

Yeah, you know, it's funny, my experience with things having to scale rapidly is from traditional retail. And the thing that they were always worried about is because it was a was a spike you could plan for right? Because you go, we're going to buy a Super Bowl commercial. And then our website's going to get destroyed because everybody sees the commercial and

50:08

This wasn't so long ago that people didn't all have phones, but people had smartphones and they'd go, well, I'm just going to go check out that website right now. And, and that was a classic case of, oh now that I'm interested, I'm going to be part of hundreds of thousands, if not millions of people surging to hit that site. And then it's like, our normal traffic is like tens of thousands of people. And when you suddenly have to go to tens of millions of people, that's brutal. Yeah. Yeah. It's, it's like, uh, it's pretty surreal.

50:37

And there's nothing you could do. It's already happened. It's already too late. You just try to do whatever hacky thing, low-hanging things you can. And you just know that you're probably going to have lost most of them, or had most of them slip through your fingers, without you having really put your best foot forward, which is a lot of what happened when we released V1.

51:07

enough insight? And I don't know that this might be a difficult question answer because there's two of you working all the crazy hours all the time. Right. But do you have enough insight that you can see the difference between people who go, oh, wait, I'm already on blue sky. I didn't realize I could skylight. And then those people who go directly to skylight. Right. Like, do you see a difference in those two types of traffic or you're really saying that every surge is explicitly just skylight?

51:36

I'm saying that every search is Skylight, but it's almost always a search from TikTok and not from Blue Sky. Really? I think that we have a much larger footprint on TikTok, which is kind of by it was our goal. We really thought of this as like, we're trying to get TikTok short form video app users.

52:03

over to the app protocol, not necessarily trying to build a product for other people who are already on that protocol. Got it. And not that they're mutually exclusive, but it's really tempting to just focus on the app protocol audience, because they already have an account. They're already familiar with lot of these concepts, and you can get to use your app. But the more valuable thing to have done is to actually make it appealing to this other group of people. How do we get people from outside of?

52:32

blue sky onto that protocol, think is really important. Now, have you have you and Tori even gone as far as to put ads on the TikTok and Instagram reels, letting people know that there's a thing or did their advertiser agreements explicitly exclude you because they view you as competing platform? I don't think we we've done that. I'm not 100 percent sure, but I'm I'm more than 50 percent sure, I guess, that we haven't done that. And I think the reason was just

53:00

It wouldn't be organic growth if we did. And we want to prove the point that people want to use this is actually a big part of why we're doing so much outreach and stuff.

53:15

So with every video that Tori's posting on Skylight slash Blue Sky, she's posting another two times, three times as many on TikTok so people can discover it and go like, oh, I should check this other thing out. I think she basically, it's almost one to one. There are some that aren't, but it's mostly one to one. At the beginning, might have been post, we probably posted more on TikTok. There's some that are like where,

53:45

we're making videos about the app protocol for somebody who doesn't know anything about that protocol. And those don't really make as much sense to post on Blue Sky. So some of those don't go through. But for the most part, they're pretty much the same. think that the big thing was just the TikTok users are really the ones who care most about the TikTok ban and the problems with TikTok. And a lot of people on Blue Sky aren't really interested, I think, in a short form video app. Oh, OK.

54:13

Yeah, mean, honestly, wouldn't know because I am an atypical user by far. OK, let's see. Nick's got a question. ah Have any thoughts on edge services to distribute content processing? Relays and a growing number of PDS instances have pretty interesting impact things. um Yeah, I'm pretty big on the edge services stuff, especially if we can just get the right paradigms down. I would love to have if.

54:42

I don't think we're ever going to make a PDS, but if we did, think it would be one where we try to get it to live on the edge.

54:51

Yeah, I could actually talk a little bit more about that because I think that's natural question for an app protocol app right now. It's like, would you make a PDS? I think that for us, because Blue Sky is a PDS themselves, if we end up being this second app, I think there's this mythical second app that has to develop the app protocol.

55:20

And I think it'd be really important for the app to be very different from what Blue Sky offers and to kind of prove the point that like, we don't have to be the ones who own your account. And so it's like, you could sign up with a Blue Sky account or you could sign up with a Black Sky account or like another PDS provider. And we like list all those, but because none of them are us, we're not really biased towards you think that most of the, let's call it,

55:50

Popular at views are all within the same realm of standard social media that people That you're saying that the next big thing isn't necessarily going to be within that social media space It's going to be something that is social media, but not a traditional thing, right? So like you said you got the blue sky You've got yourself. You've got things like spark, which is on the horizon Sky swipe, which is you know a viewer. Um

56:19

Then you've got your other cool things that are branched, like you've got your front page FYI, you've got ah smoke signals. Those are all things that are slightly different. They're not standard social media as people think of them, as most people think of them. So you're thinking that the next big second app is going to be something outside of the realm that already exists? um I think that the app that makes sense

56:50

That makes sense. The second app is one that just really takes advantage of the connection with Blue Sky, which is why we do it as much as we possibly can. Because that's really one of the unique advantages that we have on that protocol is that we can do this sharing thing, and we can leverage the fact that they had to bootstrap this social network. And then we don't really have to. We don't really have a cold start problem in the way that a normal social media startup would.

57:20

The second we turn the app on, there is a video there with 4,000 likes. And that's incredibly important, because that's probably the biggest problem to solve for any social media app. And that we can just skip right past that is huge. So I think that's necessary almost to prove that there's also more here in the app protocol, because

57:48

If it becomes a trend that new apps are succeeding on the app protocol, then there'll be lots of investment. It'll be much easier to build them. But somebody has to do that first. Somebody has to prove the point that there's unique advantages to building on that protocol.

58:06

OK, all right. Last question for me. ah Are you going to the NYC Hack event in three weeks? Yes. Sweet. I'll see you there. All right. right. Any, since we're like with the last five minutes, any things that you want to stand on a soapbox about? Do you want to thank some people? Do you want to give people some heads up? The last five minutes are all yours. I just want to thank Tori for being the best.

58:35

co-founder guy could ask for. uh

58:41

I hope she's listening, but she might be in a meeting.

58:47

Don't worry, that'll be the first clip I'll publish. I'll make sure to tag her. I'll be like, Tori, just want to let you know, Reed couldn't think of anything to say other than thanks for Tori. Thank God for Tori. I really think of you as a dynamic duo because I get it, having worked in technology for 30 years, right? Sometimes you're just like, man, I don't even have time to write a freaking report to explain what I'm doing. I'm so busy trying to do stuff.

59:16

And then she's out there going, I'll be the face, and I will do all the talking, and I will field all the questions. And you're like, thank god, because I can't answer how I made something work when it just broke on me, and I have to make something new to replace it. And she always pushed me in the right direction. And I pushed her in the right direction. And we got a little yin-yang thing going on. That's nice. That's real nice. Anybody else in the?

59:45

you want to talk about shout outs, things that are coming. Like you mentioned Skylight version 2 is on the horizon. What else? I'm really excited to see Dame's Anasota app. I just think Dame's a really cool guy. And I'm really interested in what he puts out. You didn't ask to be part of the beta? No, I got my own stuff to work on. I'm always like, I'm seeing it on the e-reader was like,

01:00:14

I was like, oh, man, he wants to make it for an e-reader. OK. I don't Something's going come up with something really interesting that I would feel like I'd never have thought of. I feel like Dame's the guy who'd do that. OK, so you're waiting for a first full bloon release before you get into it and give it a go. That or just more time. If maybe I get Skylight into a good spot.

01:00:44

I could take a breather and check it out. Okay. Well, both Natalie and Bailey are both like, it's a pretty great app. So the chat is very much hyping it up. That'll be my reward for finishing Skylight V2 then. Nice. Nice. We'll all hold you to it. There are a couple of people here and we'll all be like, Reed, yeah, release two. You know what time it is. You gotta go check out Minnesota. All right.

01:01:13

Any other parting words?

01:01:17

No, I think we covered so much of this. You got a lot uh to edit through already. That I do. All right, well, Reed, thank you so much. Like I said, I'm excited that we'll get to meet up in New York. We'll see Nick there. Nick's very excited about it. ah I'm excited, Same name as your favorite posters poster. He is.

01:01:47

You know, it is interesting with Dane because when somebody made the comment and it might have been him about there's social media that's PBE versus PvP and I am very much a PvE kind of guy and he is a very much a PvP. Yeah, that's funny. I never thought of it that way. I could have to noodle on that. Yeah. All right. Well, Reed, thank you so much for your time. uh We went right up against the hour, which was great. I got lots of uh audio to edit and

01:02:16

fun animations to create. Thank you for your time. Thanks. Hope to be back in a couple of months or something. Definitely. A couple of months it is. Go. Bye. All right, bye.