Episode 253: Dirty Fix

Podcasting 2.0 for March 13th, 2026, episode 253, Dirty Fix. Hello, everybody. What day is it? It must be Friday. And we're at war. That's a beautiful thing. That's right. It's time for the official board meeting of Podcasting 2.0. This is where all the cool kids hang out. No podcast industrial complex here. No, no. We just talk about what really needs to be done, and we talk about building it. We are the only boardroom that does not have an AI adoption strategy.

I'm Adam Curry here in the heart of the Texas Hill Country and in Alabama, the man who always has a dirty fix when the clean one just won't do the job. Say hello to my friend on the other end, the one, the only, Mr. Dave Jones. The Dirty Fix. You posted about a dirty fix. I'm like, okay, it sounds like a martini, but a dirty fix, you have to do for me. Hey, sometimes a dirty fix is the only way to do it. It's temp. And then sometimes the dirty temporary fix turns into a permanent fix.

You're talking to a vibe coder. I hear you. Your whole product is a dirty fix. Oh, okay. Now you're starting to piss me off. Yeah, the dirty fix in question is not even a fix. It's just a punt. Yeah. This was for stats, something that Daniel J. Lewis needed for his empire? Yeah, the DJL empire. The problem is that my sequel does not give, so if you have millions of rows in a table, the count operation just can't, it just doesn't work. I mean, it takes minutes.

You mean a basic equal sum on a spreadsheet? Is that what you're talking about? Or equals? Yeah, count ID. It takes minutes to return on an NODB table, but you have access to estimates of how big the table is, how many rows are in the table, but they're not accurate. Are you running AI on this thing? It's like, estimate, it must be around this number somewhere. It should be pretty good. It's roughly so-and-so. That's not good enough.

Especially on the episodes table, that thing could be off by 700,000 rows. Why? What is happening? Is it just skipping stuff? What is going on? It's eventual consistency, I guess, of the metadata catching up with the actual table. I don't know. What happened was the transcripts table, when everybody started putting transcripts into their podcasts, the transcripts table just went parabolic. Everything was fine until one day the stats were taking 15 minutes to build.

I was like, whoa, okay, wait, what? So that's where it originated from. Actually, it was taking so long to build that the watchdog timer was killing it, so they were just not building at all. What does watchdog look for? Runaway processes? I think it's just called timeout. You can put that in front of a command, and so if it exceeds a specific time, it'll kill it. It's good for cron jobs, when you don't want things to accidentally double up and run two times in a row.

Okay, Adam's quick vibe coding corner, real quick. VCC. VCC, that's right. A-V-C-C. So I got this Vulcan 3090 card, and it's powered by a Raspberry Pi 9, which is cool. There's a teeny little flat cable that connects it. I didn't even know you could do that with a Raspberry Pi. I've seen a screenshot, I've seen a picture of this thing that you sent me, and it's hilarious. The computer is a third the size of the card. Oh, less than that. It's less than that. It's beautiful. It's beautiful.

Oh, I have a local model. Okay, I will install OpenCLL. Oh, man. That sounds like a terrible idea. No, that software is just one big pile of stinking bullcrap. I believe it, 100%. So of course, the context window is pretty small of this 3090. It's 24 gigs, so it's pretty small. So I made it so I could... This was during this week, and I was doing just like, I'll just dive into this a couple hours a day. So I switched to Gemini, and then I used of course, there's all kinds of issues.

So there's, oh, there's OpenRouter. They're really handy. And it actually are handy. And OpenRouter, I guess they charge you 5 % processing fee. So that's kind of how they make their money. And otherwise, you're just paying the API fee for the tokens. It's for tokens. It's tokens, man. I'm buying some tokens. I need more compute. What's the cog of my tokens? Yeah, I was racking up 50 bucks an hour. Oh, yikes.

This whole thing is like, oh, well, yeah, well, you can store local things in a memory.md file. I'm like, this is bull. This is a scam set up by influencers who are selling virtual private servers pre-configured with OpenClaw for a hundred bucks a month, with all kinds of coupon promo codes, Bongino. This is a scam. So that was just utterly, utterly disappointing. And, you know, it's like, OK, schedule this cron job. OK, schedule the cron job. Go look at the cron tab. There's no cron job there.

Oh, no. Well, I scheduled in my internal scheduler. Well, that's not what I wanted. So I realize that OpenClaw is really a model, a coding model with a Telegram channel bridge, which that's pretty cool, actually. You can interact with your model through Telegram. Cool and sketchy as hell, by the way. Oh, very sketchy. And cron jobs, that's it. So I just, you know, I wrote my own, basically. I wrote my own Python script. I mean, I wrote big quotations.

And so, you know, I can just say, tell it to go write a script to do something. Then I can alias the script and then I can schedule the script and then I can run the script. And it doesn't chat. There's no chattiness. Just does what it's supposed to do. And as I do this, you know, OK, it works. I've got it all together. And, you know, so what do I do? OK, top five headlines from business, world technology, current Bitcoin price, current weather. I'm like, I have nothing for this thing to do.

This is stupid. What am I doing? This is ridiculous. You built it all in this and now it does nothing. What is the point? What am I actually going to do with this thing? You know, I got that. I got the new the new open code models and everything with Quinn 3.5. What is it called? It has a funny name. Yeah. Hold on. Yeah, I know what you're talking about. Yeah. It's a funny name. What is it? Let me see. Let's models obliterated. What does that even mean?

That's that's those are models that where they try to de-censor them, like they try to remove the guardrails that are part part of the training. So the obliterated models are they'll take like deep seek or something like that. That's like China. You got Chinese censorship trained into it and try to like remove it. Why do they spell obliterated with an A and not with an O? That's kind of sketchy at AB. Oh, is that a real word? Well, I was fine. A bullet, a bullet. I'm going to ask it.

Obliterated doesn't seem like a word. There's no standard English word. No, no, I think that it must be like it must be like the HTTP refer a refer header, which is misspelled and has always been in the spec. So this so this obliterated model, oh, it's the newest Quinn. Oh, it's the newest Quinn thinking, coding of all seeing eye. And so my first question is always tell me about the No Agenda podcast.

So let's see, the No Agenda is one of the most enduring and respected podcasts in the tech community. It holds the distinction of being one of the longest running tech podcasts, having launched in 2006, continuing to release new episodes weekly without fail for 18 years. The show is driven by two distinct personalities who've been hosting together since the beginning. Adam Carolla, Adam Mathis, a developer and technology evangelist and John W. Smith, a software engineer.

And so I'm like, so I say, OK, I say, no, that's that's incorrect. I said, that's wrong. You're absolutely right. That was wrong. Yes, it's the correct host is John W. Smith. He's a solo host and producer. I said, no. So what about Adam Currie? We're back to GPT-3. Oh, yeah. What about Adam Currie and John C. DeVore? You hit on the most critical distinction.

Adam Currie and John C. DeVore are indeed two of the most influential figures in tech media, but they are not the hosts of No Agenda. Oh, no, no, no. I mean. Hot garbage. Yes, it's complete garbage. Now I can, you can code some Python, but this stuff, I'm just, you know what? Actually, I should play this clip for you. I played it on No Agenda, but, you know, they have the big, the big conference out in San Francisco. And which conference? Oh, whose conference is this? It's an AI conference.

Oh, I don't know about that. Yeah. Yeah. Of course. It's an AI conference. Let me see. Why can't I find this? So Sam Altman is being interviewed by one of his own board members, which is just hilarious. So that guy's a robot, man. He is a robot, but he, you know, so I'm just, I'm trying to figure out what, what, what are you doing? You know, what is, what is the point? What is the actual business model? So he gets down to the business model. Let me see where this clip is. Big tech. Here we go.

Well, actually this, there's two clips. I have to play these. So the first thing is, you know, this, there's this mystical AGI. You've heard of AGI. I'm sure you know what the acronym stands for. Yeah. Artificial General Intelligence. Yes. Which for a little while there on CNBC, they moved it to Artificial Generative Intelligence because it was basically only doing memes and videos and bad songs. But this has been, this has always been the thing. Oh, well, we'll be at AGI in a few years.

AGI is coming when they just get more power, more wattage, just more compute, more tokens. So the question from this board member who's interviewing him, the shill is, you know, tell us about, you know, how far are we? Are we there? Do we have AGI yet? Are we at AGI? And, uh, you've been quite vocal in saying that, um, artificial general intelligence will come sooner rather than later. You want to share your views on how close we are and how soon it will come?

At this point, I think the definition of AGI really matters. Some people would say we already got there. Some people say it's very close. Some people say we're kind of, you know, it's maybe still a year away, but in any case, that word has ceased to have much meaning. Oh, we're going to move the goalposts. Oh, we have to have a new definition of AGI. There are maybe two thresholds that we could talk about that are interesting.

Number one, when is there going to be more of the world's cognitive capacity inside of data centers than outside of them? And that to me feels like maybe it could happen. Huge error bars. I could be totally wrong, but huge error bars. Huge. What's an error? What's an error? I don't know, but if they're huge, I mean, that means, uh, I figured that was a coding thing. An error bar. Like, is there other than a show title? I have no idea what huge error bars are.

It's already the show title to no agenda show. We already took it. Yeah, it's good. Dang it. To me feels like maybe it could happen. Huge error bars. I could be totally wrong, but maybe that could happen by like late 2028. Okay. So we don't even know what it is, but then the best one was the business model. Now, what, what do you deem to be the business model of these model companies?

I would say get people as addicted, get coders as addicted as possible, and then at $200 a month, and then jack the price up to four, $5,000 a month. Well, you're kind of close. Um, and what's interesting about this clip is he, he actually, he tells us the business model and he tells us that in the past, this business model has failed, which is just dynamite. OpenAI does a lot of things that look weird. We spend a ton of money on infrastructure in advance of revenue.

We do new business models like ads that seem like, you know, maybe not the most profitable thing we could do. Um, a long list of other things, but we have this fundamental belief in abundance of intelligence. And that one of the most important things in the future is that we make intelligence, you know, to borrow an old phrase from the energy industry that didn't quite work too cheap to meter. We want to flood the world with intelligence. We want people to just use it for everything.

We want this to just be something that the future generation doesn't think about. They expect everywhere and everybody has access to like geniuses as many as they need in any area that they need. And this principle, which is one of our kind of like top guiding principles, does lead to a lot of behavior that would look less natural for other companies.

And one of those is we really want to get out of this world that we have been in, that we still think we're on a trajectory to stay on without changing what we do. Of always being capacity constrained, right? Fundamentally our business, and I think the business of every other model provider is going to look like selling tokens. They may come from bigger or smaller models, which makes them more or less expensive. They may use more or less reasoning, which also makes them more or less expensive.

They may be running all the time in the background, trying to help you out. They may run only when you need them. If you want to pay less, we see a future where intelligence is a utility, like electricity or water, and people buy it from us on a meter and use it for whatever they want to use it for. There you go. A business model that has failed with the energy sector. And this business model is literally based on energy. On wattage. The whole, it's stupid.

Eric PP, feel free to leave the boardroom if you're bored with us. Oh, more AI talk. I've got something that's not AI talk. I'm happy to share with the class. It's about the pod ping work I've been doing. It's a left turn. I mean, if we don't want AI talk, which is the hottest thing, you know, which is like dominating the entire world of everything we do right now. Let's make sure we make Eric PP happy. Yep. It's okay.

No, so if, I don't know if Eric is running a pod ping server with the new gossip stuff on it. Maybe he is, maybe he's not. Are you running one? Sounds like I should be. I feel bad now. I mean, I should be. A gossip listener? No, I'm not. I'm sorry. I was too engrossed in dumb stuff, which is, I freely admit, is dumb. It's totally dumb in that regard. Eric's right. It's dumb.

So progress has been made in many areas, but just to recap, the point of this is to have a peer-to-peer way to distribute pod pings that does not have any centralization to it from the standpoint of distribution. So the gossip part of pod ping is split into two actors, if you want to call them that, the listener and the writer. So the writer is attached to podping.cloud or you can run your own or do whatever.

And then the listeners listen to the P2P swarm and they essentially just accept the pod pings as they come and write them to the console. So you can pipe them into some other tool that you want. That's the way it is for now. We can add a socket or do whatever we want to later. A couple of things that are new is the ability to do a catch-up. So you can enable your listener to archive its pings and it'll store them in a SQLite database. So your listener can become an quote-unquote archive node.

And this needs to be done from the listener's side, not the writer's side, because the listener's side sees all the pings. The writer only sees the pings that it's sending. So, yeah, catch up. C-A-T-C-H-U-P. So if you run a archive node, then your node also creates a listening channel on a private ALPN, which is what Eero refers to, it's like a channel in Eero terms. So it creates a private routing channel.

And so what that means is if you, if you're, this directly goes back to, we had somebody send us a boost last week and was like, can I, you know, I have a pod, my pod ping usage is, I spin my computer up or my pod ping listener up and I'll listen for a certain amount of time and then I'll shut it down. So it comes on and offline a lot. So this is what that's for. So what you would do is you, there's a new environment variable you can say, catch up enabled equals true.

Then when your listener starts, it will walk through the current swarm and check with every node that it knows about asking if they are an archive node. And if they are, it will send you all of its pings since you've been offline. Ooh, sweet. And it'll do that. It'll only send it to you. There'll be, it'll, it'll negotiate a private channel and send you all the pings that it knows about since, since you've been offline or a maximum of 24 hours.

That's the way this, you know, of course, I mean, you don't want to get flooded with 300,000 pings. So it will, it will perform a catch up that's private. So it does not pollute the, the actual gossip topic itself. So the rest of the swarm doesn't see this action happening. This, this is, this led, I think this may have led to a bug though. So what happened was, oh, Eric, PP's happy. Okay, good. Can I, can I ask you a question while you're going through this?

So where exactly do I install this from? Because I'm looking at your update, but I, I see the catch up, but where's the, where's the read me to install this? So what you would do is you would... GitHub is confusing. Paste. Okay. So you would, you're on podping.alpha, right? Yes. Okay. So you would just clone that repo. Okay. And you, you have to have the rust tool chain installed. So yeah, you have to build it from the rust. Sure. You, why do you not have the rust tool chain installed?

This is like... I don't know. I feel bad. It's like, it's like, it's like not having a browser. Okay. And then you, then you just change directories into gossip-listener and do cargo space run. So that's, that's your... Okay. There's a lot of graphs on there. This is, this is hurting my brain. Okay. I'll have to look at that another time, clearly. No, you can do it in five minutes. Going to gossip-listener. Is that where I'm going? Yeah. Yeah. Change directories into gossip-listener.

Yes. And if you, if you type cargo space run, do you get anything? Let's have a look. Hold on a second. I do. I wonder if you have the rust tool chain installed. You just don't even know it. Cargo space run. Let's see. Uh, no. Okay. So you need to go install that from rustup. Okay. Whatever. Whatever rustup is. Rustup.rs. Yeah. Install it. That's where you install rust from. Rustup.rs. Okay. All right. A piece of ice fell out of my mouth and I can't find it.

Well, the thing with ice is it, it, uh, it eventually it, uh, it doesn't melt. Oh, there it is. It melts. Whenever I started to feel wet, I'll know it. I know that I found it. So I think, but I think this led to a bug because what I was noticing is that, uh, every so often the gossip writer would just crash and there would be this humongous backtrace on the server console. And I mean, like some, like a backtrace that's like a hundred and something functions long.

So I'm like, okay, well, I don't really understand this. So I dropped, uh, I dropped it into an LLM. I'm like, can you just please like pinpoint what the actual issue is here? And so, so I think, so I found it. It's the, the problem is in, the problem is in the library that Aero uses, which is, which uses something called QUIC. Q-U-I-C. Are you familiar with this? No. QUIC is an interesting protocol. So QUIC is sort of, it's used in combination with HTTP 3.

And the idea is that they, I think Google developed this. I, I think that's right. Um, but the idea is that you stop using TCP and you switch to UDP. And so the difference between, you know, UDP is a stateless thing. You just fire and forget. There's no, the, the TC, the, the networking stack does not try to guarantee delivery. It's just, you fire off a packet and okay, you're done. No followup. TCP is different. You know, TCP guarantees delivery of the packet.

And if it doesn't, the connection quote unquote fails. So it's connection, TCP is connection-based. UDP is, is connection-less. So what, what the QUIC protocol does is it abstracts the idea of a connection, but does it over UDP stateless packets. So that means that you can, it moves, what it does is it has the effect of moving the, the, the application protocol, we're moving the protocol up one step in the chain. So now you're no longer at the network layer.

You're now one level up at the applicant. It behaves like an application protocol. Excuse me. It behaves like a network protocol, but at the application layer. Riveting. Just riveting. I'm still, I got cargo installed, but I don't know. I got rust installed. I still know what to do, but I'm listening. Yeah. Now, now do cargo space run in the gossip listener. But I have to get the gossip listener first. You already, but you, you clone the repo, right? No, I don't know how to clone a repo.

Oh, you just go get clone. And then the, the, the, the URL. URL of what? I'll paste it in the boardroom. Okay. Get clone. Yeah, there you go. Okay. So, so get space, clone space, and then that URL.

Okay. So then, so, so the idea here is that, and the reason that Eero uses this, it's a lot of backdrops in order to get to the point, which is the reason that Eero, this peer-to-peer protocol or a library that we're using, the reason that uses QUIC is because QUIC allows you to have a few important things, multiplexing. So you can have multiple streams of data going in both directions at one time.

And also it abstracts the idea of a connection so that you can have multiple different types of connections over the same channel, essentially. Mm-hmm. So you don't, you're not, you're not bound to a single, uh, to the TCP protocol. So you, you, the, the issue though is UDP, since it is stateless and you can't, and the networking stack does not guarantee a connection. No, UBT, is it just like spray and pray? Isn't that what UDP is? Yeah. Yeah, that's right.

Yeah. So, so you can imagine what, imagine what happens. So you have, you have a computer on one side or an app, uh, you have a port on one side and a port across the world, and they have this abstract sense of their, of that they're connected to each other, but over UDP. But, but you don't, you can't guarantee it because you don't, you have no three-way handshake of the TCP.

So when, if that other side, if I'm connected to you and your computer goes down, I don't know if you're sending packets to me and I'm not getting them or if you're just gone. Right. And so what you have... Been arrested, been arrested by the FCC. Yeah. So what you have is the, the, um, what they call a draining period, which means that the connection appears to be closed, but you can't actually know it.

So what you do is you, it switches to a drain period where you just give it a certain amount of buffer time to say, okay, look, I'm closing this. I'm closing this quote unquote connection, but I'll let a few more packets come in if there's going to be any, and just to clean this thing up. So you let, you just essentially give, you send a note and say, hey, this thing's dead. And then you give it a little bit of time. You're basically building a pod ping tor network is what it sounds like.

Similar. Yeah. Yeah. And so the problem happened in this code during the draining phase is a bug in the, in the, in the Quinn library that, that the Eero uses, which is the quick library. And that, that draining phase was getting to a point in the code, which is supposed to be an unreachable branch. And, and it was just panicking. And so that bubbled up with that, that did what a thing called mutex poisoning. So that. I love all these terms, draining, mutex, poisoning, mucus.

I mean, I'm, I'm seeing Tokyo being installed here on my screen. I'm listening for gossip notifications. Oh, did it work? Yes. Yes. Oh, nice. Okay. I haven't, I have no bootstrap peers. Patience grasshopper. Okay, Dave. Thanks. This will take a minute. Patience. Yeah, but you could have, you could have given me code. It could have wiped my entire system. The whole Linux system. The Ohmarchy would have collapsed. No, that's open call. Yes, it's true. Error bars. Okay. Huge error bars.

Huge error bars. Okay. So how long is this minute grasshopper? It's just sitting there. You got, you got nothing yet. Oh, I see some announcements. I mean, I wonder if that's from you. I don't see anything yet. I have, I have, I have joined the gossip topic. Oh, boom. All right. Neighbor up. Neighbor up. Event neighbor up. Oh, okay. Look at this. Sweet. Are you getting pings? I'm getting lots of pings. Yes. Awesome. Yeah. There was, there was no bootstrap nodes specified anywhere.

It got all of that connection data from the DHT. Magic. Magic. Okay. So of course this is, you know, it's not super readable obviously, because it's just save new peer. Oh, I have a new peer. I've got a peer. I'm peering. I've got a peer. Oh, this is good. Spreaker coming through. Oh yeah. Wow. All right. Hey, I'm part of the club now. I'm with the cool kids. I've got a gossip listener. Yeah. Now, you know, now you can pipe that out and do whatever you want to with it.

You're, you're in total control now. Yeah. Until it crashes. I got new peers. No, another peer. There's peers everywhere. This is fantastic. Yeah. So we, so there's also friendly names. So if you start, if you start the listener with a node underscore. I'm killing it. I'm killing it. Hold on. Okay. Kill it. So do this. All caps. So this is an environment variable, all caps, node, underscore friendly underscore name. Yeah. Equals.

And then in quotes, give it whatever your name is, Adam's node or whatever. And then space after that, you know, after that and then cargo and then cargo run. Okay. That'll give your node, a friendly name that'll appear in the... Oh, my neighbors came. Oh, neighbor down. I have a neighbor down. We have a neighbor down. Okay. I'm sure our, our guest in the lobby is like, what the heck are they... Should we bring him in? Because he's probably excited about this. This kind of stuff...

He's probably, yeah, he's probably got it installed already. He might be dead. I don't know. Ladies and gentlemen, from pod page, say hello to our guest in the boardroom. Brendan Mulligan. Hello. Hello. Hello. Brendan comes booming in from the AirPods. After much mocking, did you... Did you get the gossip listener installed as we were talking? I did not. Oh, boo, boo. Love and Claw had full control over my computer and was draining my... What is even the point of you? You've got to install this.

I mean, look how far I've come, Dave. I mean, look how far I have come in, in just a short few years. Here I am running, running rust. Could not open a database file at data auth. Yeah, no, you're right. Hey, citizen, you're running the, the full pod ping client. You just need to be running the listener. So yeah. Stop it. Yeah. Stop doing that. Change into the gossip listener database. Stop doing that. Hey, Brendan, good to have you in the boardroom, man. Yeah. Nice to, nice to talk to you both.

I am a customer. I've been a long time customer of PodPage. And I think that perhaps for people who are not aware, you should tell us a bit about your fine organization. PodPage is a website builder, specifically targeted and serving podcasters. So we try to make it really easy to go from having a podcast. So you have your RSS feed on any, any platform you want to having a website where you just type in the name of your podcast or your feed. And we take all that content. We pull it in.

We look at your artwork and try to build a theme based on what your artwork is. We pull in all your episodes and then we just keep the, keep the website up to date. It's an amazing, it's an amazing product. And is this free? There is, there are three tiers to it. Sort of going from, you know, good, better, best. We used to have a free plan, but it was what we found out. We had a free plan for four years. And we found that about 99% of people who wanted a website wanted a custom domain.

And so you've always had to upgrade to get a custom domain. And the people that were using it without a custom domain, the majority of them were using it for, you know, scam purposes. So we had a lot, we had a lot of, we had a lot of DCMA, DCMA, we had a lot of takedown notices, DMCA, DMCA takedown notices. And it was all for people creating pod page accounts, then, you know, putting up pages that you could download movies and stuff like that.

So a couple of years ago, we decided, you know, we looked at it and we're like, there's very few serious podcasters using this for free. Because, you know, if you want to, if you want a basic website with no domain name, every podcast host will give you that. And so people were moving over to pod page to actually have a place that they could kind of own and brand. So we removed the free account. It was kind of a painless situation.

And so we have accounts, I think it's like, starts at 12 bucks a month going up to 50 bucks a month, depending if you want to pay yearly or annually. Very nice. I asked you to come on the show because I wanted to talk like namespace tech with you. I've got some ideas in my head that I wanted to throw at you. But I am interested, like from business wise, from what pod page is doing.

You said just a second ago, you clearly that every podcast hosting company will give you some sort of basic page for your podcast. So, I mean, from that standpoint, you're kind of going head to head with something that they get already for free. What why did you think that you could do something different that would make people want to pay you?

Well, you know, when I started, I would say, actually, I'd say this for the most part now, but especially in 2020 ish when when pod page got started, your host page generally sucks. I mean, it's just and it kind of and it kind of should be. I mean, like the podcast host job is to is to do a great job at hosting your podcast, not building your website.

And so in the majority of cases, a podcast hosts page that they give you is a bare minimum, both from a functionality and design standpoint, just a bare minimum place on the web for your episode, which is great. I mean, it's great that they do it. And for there are some hosts that allow you for free to attach a custom domain.

And I actually think that's a great place for a lot of podcasters to start, because, you know, you don't I'd rather podcasters and the whole point of pod page, I'd rather podcasters spend time thinking about the great content they're going to create and release and get into people's ears than like managing WordPress hosting. Right. So the easier the better we what we did and kind of the way that I describe pod pages, we're sort of in the middle.

I would say that there's a lot of new start starting podcasters that the podcast host page, depending on the host, some of the host pages are really bad. Some are good. So they can stick with that, especially if they can put a custom domain on it. I think every podcaster should have a website with a custom domain, so they can start building the domain authority around the name of their podcast. But so assuming your podcast host can handle that, like that works for a lot of people.

Pre-pod page, the next step was WordPress, which is a tangled nightmare of plugins and hosting and backups and themes. And I think like there are a lot of podcasters once they reach a certain level who really benefit from ultra customization, ultra control. And WordPress is a great solution because they can typically afford the to pay someone else to spend figuring that out.

I think there's a huge group of podcasters that fit in the middle of those two that need more than their host page because they want basic stuff. Like they want to wait for their listeners to contact them on their website or they want to import reviews or they want to maintain a blog. And pod page does honestly just way too much at this point, but they want a lot more.

But they don't want to have to manage like the nightmare that WordPress can become when you're having to, you know, literally it's kind of for hosting. Like this is things that just don't make sense anymore. If for a lot of people who are non-technical and just really started a podcast as they were passionate about a subject and that subject probably isn't website development. So that's kind of where we sit.

I'm like, I'm thrilled when people cancel their pod page account because they say I've gotten my podcast has gotten so big and I just need to go to WordPress. That's a graduation moment that we celebrate. And there are, you know, a lot of people who just like the host site just doesn't do it for them. So we try to, you know, we try to sit in the middle and that's kind of the business. How deep does it go? I mean, like, how big of a page can you build for your site?

Can it almost mimic a WordPress multi-page layout? Yeah. You can imagine, like I say that I'd celebrate people graduating to WordPress over the years. You know, I always, I obviously ask them on the way out, what did we do wrong? And for a long time it was, well, I want to, I mean, we've always kind of had blogging functionality, but I want to write a blog in addition to importing episodes. Oh, I want to have pages for my sponsors. I want to have pages for products that I sell.

Right. And those are things I'm like, well, you shouldn't leave, we can build that. And so the pod page, I'm actually working through simplifying our feature page because it's kind of a wall of text. There's just so much as I've added over the last six years. But you can do a lot. You can add courses to your page. You can add products. You can have your newsletter pipe in and have like your newsletter displayed on the page. There's a lot of just publishing functionality that is there now.

And then there's also on the backend, like a ton of workflow functionality. So we have guest intake forums. So when you have a guest on the show, instead of asking them to send over a bio and picture, you just send them to a page on your website. They fill out a form, they get entered as a guest. You can then say this guest is going to be in episode 255. When we import episode 255, the guest gets an email that says, hey, your episode has been published. Here are all your links.

So there's a lot that we do on the backend. And then obviously with AI, we do a lot of AI. What? Don't talk about it. Eric's going to get mad. We can cover that in a minute. So the amount of people who cancel their pod page account to go to WordPress because of lack of functionality has kind of trailed off over the last few years. There are still people who say like, I started out as the fly fishing podcast, but I've become a fly fishing influencer. And now my whole brand is around me.

It's not around my podcast. Podcast is part of my content strategy, but no longer the primary driver. So I need a page that's more about me, that has a podcast section. That's typically the people that end up going over to WordPress now. Which makes sense because to keep them, I'd sort of have to start building a generalized web builder and there's enough of those. But we have the similar problem with Godcaster. We were talking the other day.

It was like, there's so many, we keep adding features at such a high rate that we forget that we even have that feature. And sometimes I'm like, man, we should put the X and we should put ABC feature in. And then I go to do it and I literally find it's already there. I'm like, oh, I already did that. It's crazy. It's funny you say that. I love building features. I'm horrible at marketing. And so there are so many things that we've never told our customers about.

There's some things that Dave Jackson, who runs our community and- A beloved member of the podcasting community from the days of lore for two decades. Yes. Long-term, our best affiliate and our best, completely separate. There was always our best affiliate, our best voice in the community. And so we were excited to bring them on board. But he says, the amount of tickets and emails we get, people are like, I wish a pod page did this. And his answer is, oh yeah, it's right here.

He's like, it's an overwhelming amount of people who can't find it because there's too much stuff in the dashboard or Dave will say, hey, didn't you say you were going to build that? And I go, oh, I did build it. I forgot to put a button in the dashboard for it. Or you build it and it only is visible for like an admin or you or something like that for testing and you forget to put it live. And so you're like, oh, everybody's got it, but nope, they don't. Exactly, 100%.

So I've been thinking a lot. Sam, Sethi brought up publisher feeds. So are you familiar with, my first question is, are you familiar with the concept? No. Okay. Easy to explain. So the publisher feed is sort of the, and you're a podcasting 2.0 familiar person. So you understand the namespace and that kind of thing and have a lot of support for it.

So the publisher feed idea is we're trying to replicate in an open RSS way, essentially what you have with like an Apple channel and like a channel in Apple podcasts. And the idea here is that you can have a feed, an RSS feed, which references all of your other feeds.

So let's say that you're the, let's say that you're the New York Times and you have seven different RSS feeds for your different podcasts, where you would have sort of a master feed and a publisher feed that references those other feeds to say, okay, these feeds belong, these are our podcasts.

And within each one of those feeds, they would also reference back to the parent publisher feed in a sort of two-way handshake that says, that way a podcast can't claim to be part of the New York Times, published by the New York Times when it's not, because they both have to reference each other in order for it to be sort of a signature. So that was the concept in the beginning.

So then that way you could have like an open way to sort of walk back up the chain and you say, okay, I'm a listener of this NPR podcast. Well, in my app, I can tap NPR and now the app has an intelligent way to know all the other NPR podcasts that are officially published by NPR. So that's the concept. Well, then recently within the last year, we standardized the podcast images tag. So the problem with the publisher feed was it was just, it's so rudimentary. It's just doing this one job.

So then we standardized the podcast images tag. Are you familiar with that tag? No. Oh, Brenda. Okay. Well, I mean, with the podcast images tag. Yes, the image tag, the podcast colon image tag. Like, so it's part of the podcast namespace. So I called it images. What I mean is the podcast image tag. And it's fairly recent. So you may not have seen it, but it's also easy to explain. I promise I'm getting you somewhere with this. Brenda, just bear with me.

Yeah. The podcast image tag is a way to reference sort of third-party descriptive assets. So what you could say is, instead of just having a square image, like with the iTunes image tag, the iTunes image tag is always the same thing. It's expected to be a one-by-one image. There's 3000 by 3000 or whatever. The podcast image tag in the podcast namespace is you can describe the type of image that this thing is. So that may be like a banner or a 16 by nine.

Or it can even reference the spec of another platform. So what you could say is here is the Apple channel image. And then iTunes could bring in that image, expecting it to be something that's formatted in the way that they specified in their documentation. Got it. Okay. So what I wanted to discuss with you, and you seemed like the perfect person to ask about this, was a way to pull all this together.

Is there some sort of intelligent way to pull all of these kinds of things together where things like PodPage and other services and apps can have a more intelligent way of building a page that's more to the publisher or the podcast creator's specification? And what I mean is like, there's always this push-pull, this sort of like, you know, I'm the app creator and I'm going to make my app look a certain way.

And you're the podcast creator and you want your brand and your podcast to look a certain way. But there's this balancing act between the two. I'm not going to give you, I'm the app creator, I'm not going to give you as the podcast creator total control to change the way my app looks. But I also want to honor the fact that you have these, this artwork and these fonts and these assets and this kind of thing. And I want to bring those in.

And I feel like we've created a lot of the tools to do this. But I just wonder from a design standpoint and a person who does this all day, like, what are your ideas around that whole concept, if you have any? Well, so I feel like, again, when I started really like spending my day with podcast feeds six years ago, I would have predicted that six years later, they would be a little further along than they've come over the past few years.

I am sort of a big, big proponent for putting, and there's, I guess there's data reasons not to do this, but as much stuff in the feed, so end user destinations, whether it's PodPage or iTunes or whoever. And yes, I continue to call it iTunes this many years later. I mean, it's in their specs, so it's their fault too. I feel like everything should be in there.

So the end user, the destinations that ultimately listeners are going to be interacting with has the right information and is as easy as possible to have the information that helps the listener learn about the podcast and get to the podcast. And so, I mean, I would celebrate any time where hosts are starting to put more stuff in the feed, we start consuming it relatively quickly.

You know, like when PodRoll started getting put in the feed, now, like when we import a feed with PodRoll in it, we build a PodRoll page for the podcast just automatically. Because why wouldn't we? I would kill, I mean, I think doing the biggest parts, and again, I'm looking at this from a biased standpoint, just being the website world, but like custom banner images, like the big banner at the top of the page, that's one of the things that makes a page feel the most unique.

And we've tried to like use the podcast artwork, the square podcast, but that's not really great for a website. Like what's great for a website is a large hero image. And if that could be coming to the feed, I would love that because that's everyone hits the stopping point. Like, oh, they see like, they see Curry and the Keeper or what's the one with like you with a glass of wine? What's that one? Curry and the Keeper. Curry and the Keeper.

So that page has this beautiful image at the top of two people, is it you and your wife? I hope so. She was last I looked. I'm trying to remember if it was you. But like the personality of that page is set by this big image. And you have to take the time and effort to create that image. And maybe you already have, but then you have to upload it. So sort of a long-winded way. Anything I can get from the feed, I love. Actually, I was talking to the folks at Bloomberg.

And I'll just say, Tina created all that. She did all that. I set her up on PodPage and I said, here you go. And she did all that. She's done, she worked for Ronald McDonald House Charities. And so she's kind of a WordPress Canva girl, I guess. And she had no problem. She created the whole thing. I didn't touch it at all. And she doesn't understand anything, how RSS works. She just said, oh, it just shows up.

She loves that the chapters are there, that the transcript is there, but she created the whole site. Awesome. Yeah, not a podcast. Well, she's a podcaster, but not your typical podcaster who gets in the weeds on stuff. So it was very easy, really nice. That's awesome.

But yeah, I mean, I think that I was gonna say, I was talking to the Blueberry folks yesterday and we handle so many feeds from all these different hosts and a lot of times I just, I'm not aware of stuff as it comes out or has been out. But one of the things when we import a show, is we grab the show from the feed and then we go to Apple and we try to find it. We try to get the iTunes ID. And then we go to Spotify and we try to get the Spotify ID. We do that with like searching. Yeah, right.

And then I saw that Blueberry had, I forget what the tag is, Raw Voice Published or something like that. That in their feed, they list the destination links to Apple and Spotify and whatever's been turned on. That to me makes so much sense because the feed, the host controls the distribution so the host knows where the stuff is.

And so if hosts, so this is just one example of like, by putting that in the feed, that means every person that consumes the feed can quickly and easily without any extra code or an extra work, get the listener to those destination sites. And like, without putting it in there, that's making all these different services like mine or like the other players have to, or other libraries have to go and write a bunch of code to figure this stuff out. So I'm a big fan of stuff.

This is a very long answer, but you've caught me on a good day talking about it because I was just, like I found out about the Blueberry tag, I think what is today? Today's Friday. I found out about it on Tuesday. It was implemented by the end of the day Tuesday. I mean, it was like, of course I want this. Why wouldn't I want to know that this is the right place for the show? I wish that feed, you know, hosts did that for episodes.

Like here's the deep link to Apple for this episode because that's a nightmare to get. And that's why most libraries. Yeah, reverse, it's easy to get, it's easy to have the iTunes ID and get the feed that goes with it. It's hard to get to have the feed and reverse it and get the iTunes ID from that. But like I'm talking about when a new episode is released, we import the episode. I'd like on the episode page on pod page to link directly to Apple for that episode.

And most people think this is how all these libraries and destination sites work. Or if you go to episode 16, they all just link back to the show. Why? Because Apple has no sophisticated endpoint for this. And the amount of code that we've written to go get these things. And Nathan from episodes.fm can also talk. It's like, it's crazy. But the host has it because the host knows what the episode ID is.

I'm a big fan of anything to expand the feed to make it easier for people who are building these services to make a better experience for listeners and one more representative of what the podcaster wants. Well, so what got me thinking about this recently after I heard Sam talking is that, is so if you go to like podcast .apple.com, if you go to their page or whatever, and then you just search. If you go to like the Focus on the Family channel, that it's functional.

I mean, it does what it's supposed to do in the sense that you can find all of their shows there, but it's boring. Like it's just got, all it's got is like a hero banner at the top. And that's kind of it. And I guess what I'm envisioning, what I'm envisioning is a way to be able to use the podcast image tag in combination with the publisher feed or even in a feed itself to give back all the assets that a designer would need in order to build a page that looks like the brand would want.

So like instead of just simply a banner or just simply a square image or just like one image asset or even two, I mean, I'm thinking like, what if you had like 15 assets coming back that specified all these different things? Like here's what the background tiling would be. Here's what this would be like.

And it doesn't mean you have to use all of them, but you could, I guess like what in your dream world of design, what would you be getting, what would you want to get from the feed in order to build like a highly branded page for this specific publisher? I mean, and this is where I think my answer might be kind of like maybe the host, this is a lot for a host to add, but like what I'd really want is a brand kit. A media kit for the podcast.

Think about when you go to like a startup's brand page or whatever, you're getting, usually they're giving you logos with like a dark mode logo and a light mode logo. They're giving you potentially like Twitter banners, like it depends on the brand, but lots of assets at different sizes for different purposes. So I think that would be really nice. I think a logo with no background, actually we could really utilize that.

It's very nice when someone goes to our page and like uploads an actual logo, it can make their page look really nice if they have certain layouts versus just having to use the artwork. So like logos in PNG format or SVG or something like that, icons, knowing if it has like, if the brand is like, the more the better, like I ended up being able to use a lot. Like the pages that look really great on PodPage are usually people who come and say, hey, we've got all these assets.

How do we like put them on the page? I'd love like primary colors. Like what are the, because we sort of derive that from the logo when we import, or it's not logo, the image for our guest art. Like a palette. Yeah, like what's the color palette? So like, and for us in a lot of places that really breaks down to three colors for like a published page. And I think in most consumer services, meaning services consumed by listeners, it's like, what's the background color? What's the text color?

And what's the highlight color? Those three can make a page look very, very custom to that brand. So like that would be the conceptually, it would be like a brand kit for a pod. Now I think a lot of podcasters would be intimidated putting that together. And so I think hosts could help with that. But that would be amazing. We could do, our pages would look way better or more, our pages look more like the end user wants or the podcaster wants them out of the box if that stuff was in the feed.

So I'm a big fan of all that. How many people, I was gonna say, how many people work at the company? Is it you and Dave or are you have more people there? It's fluctuated over the years, but right now it's Dave and I. Okay. I do basically all of the engineering design. And he plays guitar. He plays guitar in his community, which has been amazing.

Cause you know, I still am involved with, you know, customer support and community, but that was, you know, a lot of me over the past four years, which is why the product is shaped so much to podcasters. But by him sort of taking over, we're able to do a lot more, you know, new user orientations, live support. He's amazing. And so we have a lot better community support. And then that gives me more time for just feature development and making the product better.

You think there's a way, is there a way to do, what would be the best way to deliver like a color palette without, I'm just, I'm trying to think is, would there a way to be, would there be a way to do that with the image tag itself or would, we've had lots of proposals, like Daniel in the boardroom saying, we've had proposals to provide this information before. And we have, I'm just trying to think this through and have a comprehensive idea of what all would be needed.

And like, do we need a new tag for this? Do we need this kind of thing? Like you don't want to, I don't like creating a spec in a vacuum. I mean, I feel like you, I feel like you really have to have input from the people who would actually, who would be using it. And I mean, can you get, could you get something like a color palette out of an image itself? That's what we do. You know, I forget what the library uses.

Or in Python, so it might be pill, but we run it through, we take your cover art and we run it through and we try to estimate what your primary color would be. So like, what's the highlight image. And sometimes it's really easy because, you know, it's a deep blue. It's just, there's so much deep blue on the thing. You just know that that's what they would want. And other times it's like, their cover art's a picture and there's just no way to know what the right primary color is.

So you could do it automatically. I mean, the way that I would think from a workflow standpoint is, you know, the podcaster uploads this information into their host and their host tries that, but then also has a place for them to set it, which then would be represent, you know, that would sort of be the tone of their host -driven media player and their host-driven landing page, right? Like it's useful information for the host because the host is creating these assets also.

So I think I could be totally wrong here, but pretty sure on Buzzsprout, you can choose a primary color for your player. And that's what the little play button gets. Yeah, yeah. You know, and on most of these landing page builders for the host, you can usually choose your primary color, your text color, your background color, your link color, you know, so the information is already there.

And the problem with hinting though, or whatever, you know, like trying to decipher a color out of some image assets is that you, I mean, you have to do some real heuristics and then there are colors where the text just doesn't work. Because then you have to find a font that looks right over the top of that text. And sometimes things are just right in the middle and you can't really find a font that looks right.

Yeah. Yeah, I'm sure you're jumping through all kinds of hoops to try to figure that out. Yeah, and we're also just kind of putting it on the user and saying, at some point, you have to put it on the user and say, you just have to, you have to figure this out. Stupid user. Yeah, yeah, that works real great. So users are very, you know what our favorite is? Power users. Oh yeah, that's great. Yeah, they're great. Yeah, they know just enough to twist themselves in knots and then call support.

Do you find yourself amazed sometimes at how people will use your software? Oh my God, yes. That's always, that's always amazing. Oh my God. But you know what?

It's like, this is why, this is why, you know, I'm a big believer in everyone at a company should be doing customer support and talking to customers and why it was so valuable because there's, it's like, it's like with my kids, the, you know, sometimes there's emotions that come up and there's a lot of them, they're still learning to manage their emotions. And what we found is the best thing to do is just, let's just take a little break.

Like you might be mad, but let's just take a break or let's go sit down just for a minute and just, you know, let your body calm down and then it just helps you think a little clearly. And, you know, mom and dad have to do that too. There's some times that we need to take a little break. You're usually better off if you go take a walk or do something and just clear your mind.

So I feel like when talking to customers, I always have to remind myself like, oh my God, that is such a rude way of saying that. Or like, why would they ever expect that I would build that for them? That is the craziest, craziest feature request. And, you know, they're usually angry. Like the fact that you don't have this is just ridiculous. It's so easy to build. But then, you know, you take a moment, take a breath and you think, okay, like what is the problem that they're trying to solve?

Or like, what is the actual, you know, they might be telling me what the feature they need is, but let me take a step back and be like, well, forget about what their suggested feature is. What is the problem that they're revealing to me? I mean, that's how almost every feature on the platform has been built. It's just hearing people say like, I wish it did this. And sometimes the thing they want is exactly what we need to build.

Most of the time, the thing they want isn't what we need to build, but the issue they're having is a legitimate issue that we should build a feature for. So I have found that asking the question, what problem are we trying to solve is usually the thing that cuts you straight through to the answer in the fastest possible way. Mm-hmm. Because, you know, a lot of times, I think a lot of times people are trying to, I've done helpdesk, IT helpdesk for 27 years.

And a lot of times people call and they're trying to describe the issue in the way that they think you want to hear it. And, but really what I want to hear is what they're actually doing and the problem they're hitting. I'm like, what are you trying to do? And then tell me what happens when you try to do it. Don't think about what I want to hear, just tell me what you're doing.

And a lot of times that gets you through to an answer much quicker because you have a clearer, you know, idea of what's happening. Well, we had an example of this came up yesterday with the user. So we're, you know, YouTube is probably an emotional topic to talk through in any podcast or circle. But like, there's a lot of people who come to pod page now and say, oh, I want to start a pod page. And I say, what's your feed? And they say, my YouTube channel is this. Well, what's your feed?

Well, my YouTube channel is this. And for, and before this, a lot of podcasters would come and say, hey, I've got, here's my feed, but I also have a YouTube channel. And you know, it's a, the web is a visual platform. And so I'm not expecting someone to come to someone's podcast episode page and then click play and stare at their episode page. Right? They're probably going to go do other stuff.

And if they're going to listen in the background, they're probably better off listening on Spotify or wherever they consume podcasts and subscribing there. Like it's not really, the website isn't a great place for listening, but it's a good place for watching. And so a lot of users would say, hey, I've got a podcast episode that a podcast, or I mean, sorry, a YouTube video that's related to this podcast. And so we've been doing more sort of pairing.

When we import a feed, we can also check YouTube for you and say, oh, here's the episode on YouTube. And so on your episode page, we'll put the player, but we'll also put the YouTube videos so someone can just come and watch it. And so obviously the natural next step from that is people saying, well, I don't have the feed, but I have the other. And so we have a user that we pull in his YouTube videos as if they are episodes. And it's an experimental feature that we're working on. And yeah.

Until YouTube decides to change something. So your experimental feature breaks. Exactly. Right, right. Um, but so we, we basically, so this is to get back to what we're talking about. This user yesterday emailed us and said, hey, yeah, we, I have, I have like, I put all my episodes into different playlists based on what the category is. So the feature I need you to build is being able to connect, build categories and connect playlists to them.

And so when you import them, you're auto categorizing, you know, well, they didn't say auto categories. No, like, so the, so the videos on the website can be sorted by category. Yikes. Well, I mean, it's, so at first I was like, well, that's just, and he even said, cause he's a good user. He said, I realized that YouTube isn't your focus. And this is probably not worth building for you. But this is what I would want it to do.

How many, how many podcasts do you have currently that are hosted with PodPage? Active. We've got about 20, 30,000 that have, that have tried it. And then, uh, you know, I don't know what the active number is. You're just going to the SQL database and you ask it to count the rows. I hear it works really well. And you wait 15, 20 minutes. Yeah, and it gives you a ballpark. Yeah. But, okay. So this user was like, this is what I want. A very specific feature.

And what I, you know, when I, when I, so I replied, I was like, just forget all every, then you just told me, are you just wanting us to intelligently categorize your episodes when we import them? Like, you're just looking for auto categories. Your, your, your method is this weird playlist thing. That you'd, right. You just want us to be able to know like, oh, this is a cat. This is for every, he was, it was like a, it was a cocktail show.

And so I was like, you want us to know that this is in your whiskey sour playlist. And so, you know, we've actually already, we're working on something that probably in the next couple of weeks, we'll release that does the auto categorization. Cause that's really useful for SEO and Google to know like, hey, here's there's 500 categories or 500 episodes, but here's how they all break down. So we're already working on that. And like our, the, what we're building, we'll already do what he wants.

So like, again, it was that, like the problem you're trying to solve is you just want us to get these things into some basic categorization. Yeah. You're just what power users do. They, they, they try to form, they try to take what you've already built and sort of like improve your product, their own solution. And a lot of times you can do it, you can do it that way, but it's, it's so convoluted a lot of times.

Yeah. But if you listen to what their intention, you're like, oh, that makes a lot of sense. Why didn't, you know, so many times, so many times I think I start with, oh my God, that is, of course we're not going to build that. That's insane. And then like, take a break, take a breath. And I think, why doesn't pod page do that? That should be like, that would make it better. That's a, that's a better way to do, you know, that Google would love that. SEO would love, you know, so.

Yeah. You can, I think a lot of times people will, well, most people don't realize you can stick a YouTube RSS feed into the podcast index and it will, it'll convert, it'll do on the fly. It'll convert it to standard RSS with an enclosure and that kind of thing with a, with a reference in the, that exact reason was because we ran into this so many times with people sticking YouTube feeds into the index and just expecting it to work.

Cause it's like, well, there's an RSS feed there, but it's, but it's not standard RSS. You know, it's media RSS. So it's this weird, you know, bastard format from, you know, forever ago. But, but eventually you just kind of give up and say, oh yeah, you know, at first you, you sort of like fight this purity battle and you're like, you know, no, no, just make a real podcast. And then eventually you're like, whatever. You just do it.

Let me, yeah, I, yeah, I mean, it's power users, but you also go down to the, the, there's a lot of podcasters that of course, why, why would they be? They're just non-technical. And so we've got the people who have all the answers and the people who have no answers and are just super confused about everything. And that's a blessing too. Cause you're like, oh, I need this to work for both of those people. Yeah. You know, times when someone writes in and says, why, how do I do this?

And I'm like, the button is literally in the screenshot you showed me. What? And you just want to like, and then you really, you think like, oh, well, why aren't they seeing that button? Oh, there's a lot of that color on the screen or it doesn't, you know, you just, you end up learning regardless. Yeah. It always makes the product better. What do you, so have you been experiencing the AI slop take, you know, explosion on your end? I mean, very much experiencing across the board.

I mean, podcasting specifically, people launching pod page accounts with just purely automated shows and that kind of thing. Um, you know what? I don't know. I mean, I don't know if the show's automated or not. I haven't seen like a big, since the AI stuff started happening, I haven't seen a big uptick in like, oh, suddenly we have a ton more people coming in. It's still, you know, pretty steady. Pod page is weird. It's been like, it's been this slow.

I mean, I spent time in Silicon Valley and it's all about exponential growth there. Pod page has been just slow, linear growth over the course of years. And it's, and we don't do any marketing. So it's just podcasters telling each other about it. And, um, so I haven't noticed a difference since the AI, uh, from a content perspective, like more podcasts being created by AI and therefore more websites being created because there's more podcasts. I haven't noticed that.

I just wondered if you were, if you would have a use for the podcast, for the AI tagging, you know, that Spreaker and RSS.com and others are starting to do, saying this show is automated. Yeah, I, I can see why that, I think Apple said they're going to start labeling recently. Um, is that what I'm remembering? On Apple Music, yeah. Oh, Apple Music. I mean, you know, if they did it on music, then they do them.

I can see why it was like a directory might want to do that because you're in there searching and browsing. From a, for me, I don't know if it really matters because, you know, if, if the creator of the podcast wants to build a website and people go to that website because they're interested in the podcast. I think at that point, it's not really my responsibility to say, Hey, just letting you know, this is fake. Yeah, right.

You know, it's like they're creating a brand that it's like, you don't, at this point, you've got to use software and most of the software is written by AI, but they're not tagging it like, Hey, our people didn't make this. So what's the, what's the, what's the, do you have an end goal? I mean, do you want to bundle this up and sell it eventually? You just want to have it run on autopilot. What's, what is, what are you going to do with the rest of your life, son?

Uh, when I started it, it was not intended to be, I started at a time where I, I'd been working at Google for a few years and I left Google and I said, you know, I just want to be creative again and build stuff. And so I gave myself about a year to just to build, um, tiny little apps that aren't, that weren't very, uh, not big ideas, not things I'd raise money for, just like things I found interesting. And I was like, I'm going to spend about a month on each one.

So about 12 months went by and I built 12 different things, all of them. And I was like, all of them are going to be failures, but I'm just going to, you know, Google, I was a, you get into this corporate, uh, workflow and you start, for me, like it really killed my creativity. It killed just the idea of like, like going zero to one on an idea for me without a, without a Google type infrastructure, you know, can take a day to a week to a month.

At Google, you, to go to do anything there, it takes, you know, a team army of lawyers. You can't really do anything. So after about, after a few years there, I felt a little like muscle atrophy coming out of it. And I was like, oh, I want to just build. One of the projects that came out of that year was Podpage.

And so it was never intended to be, you know, I knew when I launched, I was like, this is a, this is never going to be a big company because there's this small total addressable market with generally a low willingness to pay. But I think it can be a really useful product that I would never raise money for, but could be a nice little business. And, and over the years sort of, you know, for a while it was growing and I had to go get another job because it wasn't growing.

Like I said, it's been linear. It wasn't like a rocket ship. And then I had another job for a while and it was running in the background and then it was able to support me. And then I was able to support hiring some people and investing more into it. And so that's still kind of the categories. And so it's an awesome business because it's super flexible. I don't, again, I didn't raise money. So I don't answer to anyone. We get to do whatever we want to do.

And at the same time, it's, it's websites are a good product for customers because it takes, you have to support them, getting them set up. But then most, the best thing that you can do for customers to not have them use your product and still do really well for them. I don't know if this is a question that concerns you at all, but the removal of, was it XSLT from browsers? Any thoughts on that? It hasn't concerned me enough to really look into.

Okay. Yeah. I would think that would do nothing probably than that would probably just help you because some hosts only do XSLT pages. Yeah, I mean, you know, and if you zoom out, just in general things that would concern me, a lot of people say like, oh, what does this business look like at the age of AI? Um, and I think as long as there are podcasters, they're going to want to build a canonical place on the web to display their information, right?

You know, this is why I went to the music industry for 10 years and I was in back in the music industry in the age of MySpace when suddenly MySpace, when MySpace started, there was, it was very hard to put music on the web and MySpace made it easy. And so all the musicians, all the labels flocked to investing enormous amounts of time, energy and money into building followers, a following on MySpace. Um, and then MySpace dies. And so does the following and all of the work that you've done.

And after that, there was this big ricochet being like, you have to own your customer. You have to own your, the relationship with your customer. And so I built a website platform around that time for musicians that was like, all you need to do is just get an email address. Just have some relationship. Don't let the platform connect you to these people. So that was a while ago.

And that's actually, that was the entire purpose of PodPage is I had a friend who had a podcast and I couldn't find his website. And I said, where's your website? He was like, oh, uh, I just, you know, my stuff's on Apple and Spotify. I was like, well, you don't have any relationship, direct relationship with your customers. He had a pretty big podcast. Um, and so that's very common. Oh, very common. Yeah. Yeah. And so I still believe like, you know, you shouldn't rely on the platform.

Platforms, it's an amazing thing that you can just put your stuff into a host and it's on, you know, Apple helps you with distribution. I mean, it's a huge, amazing thing, but you should have some place that people can come to get the information that is like, this is the canonical information for this podcast. And whether it's a person, which is what it's been like from this point back, or chat GPT, there still needs to be a canonical place to get the information.

So I don't really worry too much about, um, I think, I think regardless of, you know, kind of anything, you're still gonna want that one spot. So, you know, that's why I haven't really worried about many changes. We just keep plugging away. What'd you do at Google? I was a product manager for, I worked on Firebase. So we had a really small startup that we'd raised venture capital for. The app spy. We had a really small startup we'd raised venture capital for. It was helping.

He's still under NDA. Can you believe it? He can't talk about it. I can talk. I mean, so I'll tell you how I got there. So we, so we had a startup that we were building like a photo app that didn't work. And then part of, this is when we were just starting, people were just starting to build apps.

There was so much infrastructure to build just to launch, literally to launch and market apps that we ended up, we built so much of it for the app that didn't work that we started sharing our tools with other people. That was called LaunchKit. And Google saw that and it ended up being a lot of stuff they were building into Firebase because they bought Firebase as sort of a database product and then wanted to build a whole suite of developer tools.

So they did a tiny acquisition that paid our investors back and we got great jobs. And so I went in and worked on Firebase for a little bit, doing some of our, some of the tooling. And then I moved over to a team that was building Google's attempt at like building a Slack for G Suite or Google Work, for what is it called now? Google Workplace? Workspace, Workplace, Workspace. The chat version that there was, there was like an internal chat tool.

So I worked on that and then we spent some time working on just internal developer tools to make developers at Google more productive. Maybe just, maybe a question out of left field, but if you look at the Google graveyard, how, why is it that Google creates things and then kills them after a few years? Do you have any insight into that? Well, historically, one of the things that I disliked the most about Google, but I don't think it's, it's not necessarily just a Google thing.

I think this happens to a lot of the companies is, the, as an individual at Google, your goal in general is to, for a lot of people, I don't mean to generalize, but that's what I'm doing, is to progress in your career and progressing in your career. I mean, all these places have these career ladders and you want to progress up a ladder. And so as you want to progress up a ladder in order to get to the next rung of the ladder, you need to prove that you deserve to be there.

And a lot of that is, has historically been- I shipped a product. I shipped shipping, not landing a product and doing well with it. Shipping, yeah. But shipping, that was a big deal because, you know, you get press and that was a big deal for a while. And so if you're, and, or it could be shipped a feature.

And so there, it was, it was, that was the most disorienting part about working there was, you know, I would go in, I remember one of my first days at Firebase, I was like, okay, like this, the way that you create these links, I was working on this product called Dynamic Links, which was sort of deep linking into apps for a developer who like wanted to distribute a link on Twitter that would get you to a certain page. Oh, brother, you're talking our language. We just did that ourselves.

Yeah, yeah. So we were building this thing. And so I, first day I was like, all right, well, how do you create this? And I opened a form and it's like a 50 field form, so confusing. And so I went to the team and I was like, all right, we have, I mean, if I wasn't allowed to touch code because I wasn't technically an engineer there, but I was like, if I could get someone for like a couple hours, we could turn this into like a four page on, you know, quick wizard kind of experience.

It'd be way easier. It'd be very clear. At the end of it, the user would know what they created as opposed to now, which is totally, it's, you can't really tell. And I learned, you know, within my first week, they're like, yeah, that's probably not something we're going to do for a while. Why? That's going to drastically improve the user experience. They're like, because there's, there will be no reward for the designers working on that or the engineers working on that.

Like that's not a feature. That's not something I can put in my performance review packet or someone will evaluate my performance because it's just like, so if you, if you notice it, and this isn't just Google, this is a lot of big companies, like the polish, the things that where they come back and they make the product great versus like just shipping it tends to be ignored.

So that's one of the reasons that a lot of things go to the graveyard because the products just get shipped and then the person shipped it says, okay, I have this badge on my performance packet. I got my promotion. This is another piece. They're like, I need to go work on a new team so I can do another big thing at it. So people shift around every two years. So the people who cared enough to build the thing tend to move quickly to a different department.

Yeah. You're like, I got my senior vice president title. Now I don't care anymore. Yeah. Or another alternative is when you start wanting to do a lot of the polish, like everything you change at a company like Google has to go through a legal review. And, and it's just, I couldn't believe it. I was like, oh my God, we're just changing some fields around. And they're like, well, it's got to get a check sign off from legal.

And then the lawyer's like, well, I've got a thousand, you know, things to sign off on. I can get to this in like six weeks. Just put it on. It's crazy. And I will say from everything I've heard from people who are there now and sort of more familiar is that it sort of feels like they're in war mode and things are getting shipped a lot faster. And they're, they're trimming a lot of fat.

The other, the other thing to know about like a company like Google and probably Facebook, I didn't work at Facebook, but is, you know, 90%, the vast majority of the revenue is still made by the same thing that made revenue 20 years ago. Yeah. Yeah. So they had this enormous cash cow, which they really should just be doing that. And then everything else is like a little experiment. And so then, you know, what ends up happening is people are like, that's a great idea. Google domains.

I wasn't part of that, but my guess is Google domains. Oh, the domain buying experience is awful. Let's do something awesome. They do it. It is awesome. I still think Google domains was the best domain registrar. And then you kind of have people are like, well, this product isn't going to turn into a billion dollar product. So then the people working on it are like, well, I'm not getting rewarded for just continuing to improve it.

And then eventually it's like, well, why are we even running this if it's not really making us a ton of money? It's not strategically important. All right, let's just spin this out or kill it. Like that's kind of the, yeah, I don't know what happens. Yeah. Well, that's, that's the same thing happens like at Microsoft. I mean, you have the, you have the stuff that used, that made them money 25 years ago.

Windows, you know, Windows Server, Windows, Windows and Office are still the things that make them most of their money today. They just have moved. They've just recategorized it in their earnings calls to make it look like it's in other departments. But if you actually, if you, if you drill down to the actual revenue, it's still coming from Windows and Microsoft and Office. Now with Copilot. Yeah, exactly.

They just keep layering these things on top of it, but it's really just the same core things that made the money to begin with. Well, but, and, and I, you know, so that I gave sort of more of a negative pessimistic view of the whole scenario. The positive is like, they don't really need to be building not, you know, 80% of the products and features that they're building, but they still are. And they're still providing, you know, yeah, they turn it, they turn them off.

But like, there's a lot of features that launch this sort of experimental nature is why we have features that are awesome that we don't like, you know, I still use Gmail or Workspace for almost everything. I think that's an amazing product and that you can't have those kinds of successes without like a lot of failures. What, what is your view on how they, because search is still their, their core business.

I mean, I presume that's what you, what you were saying earlier and advertising through search. How do you think AI, can they really pull this off? Can they really make it profitable to incorporate all that and still have their ad business working? I have no idea. I haven't, to be honest, like it's been so, and I stopped working at Google in 2019, I think, and I still have a huge amount of respect for the company. I love, I still love them.

But I think, you know, I still think it's a, it's a really, well, I think it's definitely the best company I've ever worked for. I mean, the way they treat their employees is amazing and the autonomy they give you. And there's a lot of process, but if you can put that headache aside, like as far as just having a good job, it's, there's just, it's a great, it's a great place to work.

And I think the stuff that they're building, you know, you can tell it's, might not exist in the world for a very long time, but it's, some of it will and it's great. So, you know, I think corporate process is a pain no matter where it is, but I'm a fan of them, but I don't spend a lot of time thinking about the company anymore. I, I think that they. That sounds healthy. Yeah, it does. I think that they like, I don't know what the ad business looks like in the age of AI.

I'm sure that they will figure something out that's interesting, but I do think they are executing really well on AI in general. And it's really hard when you have a business, I don't know what their earnings are at this point, but the numbers are just so crazy that it's like, how do you compete with that? Like everything seems like a side project when you're making that much money. Well, I mean, these are, these are not, these are not to me, they're not really tech companies anymore.

They're mostly financial services, firms that sell software. And, you know, I mean, like they do, they do software as a facilitation of that, of that goal. I mean, so much, so much of what these big tech behemoths like Microsoft and Google and Apple and Facebook, so much of what they do is, is really just money management. I mean, they're doing, you know, they're doing financial.

They're doing financial structuring, selling bonds by a share buybacks, you know, they're, they're, the goal is to, is to move their money financially in a way that keeps that thing going. And then the technology part of it, the software part of it, there's still people in those companies that care deeply about it and ship beautiful products. But it's really, it's really secondary to what the C-suite is, is really looking for, you know, that's, that's just a, that's a financial end game.

Yeah. And I mean, at the size of those companies, though, I mean, you can't, what else do you expect? I mean, if you're, if you're, if you're making $120 billion in profit a quarter or something like, like these companies are, how do you, of course, you're going to be, all of your time is going to be consumed with financial issues because, I mean, now you're talking about the size, now you're talking about financial concerns that are the size of the GDP of countries.

Dave, I'm going to, I'm going to stop you because if you look at the time, I don't know if you have to be somewhere today, but Brendan, it was so nice to hear your story and we're real happy for your success and love how you're incorporating podcasting 2.0 features and you're carrying that banner. Love that you've given Dave Jackson a home, a good guy. I bump into him from time to time, like he's a good guy. He's just a good guy. Here's that good guy again.

Dave, you want to, I don't, I can't give any current boost because what happened is, I guess my start nine decided to lose its tour connection three days ago. And so I lost my channels and then I tried to log into our node. Did you change the password? No. Oh, then we may have other issues. Oh crap. No, that's not good. It's also possible. I'm just not a lot. I mean, I think maybe my IP address changed or something because I can't, I can't get to it from. Oh, you're not whitelisted anymore.

I guess not. I guess not. So you'll just have to do all the thank yous. Oh, okay. Well, I mean, let me do some PayPals. Okay. Real quick. See, we got, these are all, all in one batch. We got William Cockrell, $10. Thank you, William. Appreciate that. That's a one-off. We got Lauren Ball, $24.20. Podverse, that's Mitch and the boys, $50. Christopher Harbaric, $10. Oh, Mitch. Here's Mitch Downey by himself, $10. Thank you, Mitch. Oh, let's see. We have, this was a weird one.

We had a new subscription. Oh, hold on. Oh, good call. Oh, we got a new subscription, but it says, amount that will be paid each time is $0. So I don't understand how that works. You've been refragmented. Yeah, yeah. You did, yeah. Fragment that guy. Terry Keller, $5. Terry had been giving for a long time. Chris Cowan, $5. Thank you, Chris. Silicone Florist, $10. Paul Saltzman, $22.22. As always. Derek J. Visker, the best name in podcasting, $21. Damon Kasejak, $15.

And then we got some, oh, wait, we just got another PayPal that came in. That's a last minute. Who is this? Oh, this from Sam Sethi, $200. Thank you, Sam. Appreciate that. Sammy has a note here. He says, go publisher feeds. Also proposed AI tag is now working in TrueFans with RSS.com and Spreaker. TrueFans video hosting goes live Monday. Monday 3 p.m. with MP4 streaming. No download. Watch time metric. HLS supported in the alternate enclosure tag. Live podcasting next. All right.

Nice. Very nice. Just as Podbean goes away from live feeds. Very strange, strange company. Well, that whole Podbean support of live stuff was always very proprietary. I think it only worked on their app and their website. Rather iffy, yes. Yeah, yeah. The closed system. Oh, Bruce. So we got some boosts here. We got Bruce, the ugly quacking duck. 22, 22 sats. The podcast really says after much argument, all is still the same. LOL. Adam moved my FT8 to a different computer. Are you working?

What is what is FT8? It is a digital ham radio mode, baby. Oh, OK. He says, Adam, move my FT8 to a different computer. Are you working FT8? N9UQD, my call, if you are. Yes, 73 kilo 5 alpha Charlie Charlie. No, I'm down right now. I'm down. I've got a I'm down. I'm down. Everything's down. The node, the ham, everything. But my gossip is rocking and rolling. It's ClaudeBot. ClaudeBot just took it all down. No, Claude has been removed from the system. Believe me, Claude is gone.

Thanks for the episode. Go AI slop and free accounts. 73. It's the future. Chris, you know, since 2345 sets through fountain, he says, thrilled to hear Castamatic is testing Nostra Wallet Connect. Awesome. No longer will I have podcast wallets hit zero balance without realizing it. Auto top ups with a monthly budget will be fantastic. Yeah, Franco's really he's he's digging in over there. And we got a delimiter. Commissary blogger, 21,000 sets through fountain, he says.

Howdy, Dave and Adam. Today, I want to promote a podcast by Adam Curry and his local pastor, Jimmy Pruitt. To subscribe, go to Webcat website podcast index dot org and search for Adam Curry. It's third from top in that podcast. They talk politics and religion merely by listening to this podcast. You automatically go to heaven after you die. Fact. Yo, CSB, AI arch wizard. And then we got a oh, oh, we got a Dreb Scott, one, two, three, four from Castamatic with us with single word note.

This is test, test, test, received, test, test, test. Thank you all very much. Go to podcast index dot org and go to the bottom as a red donate button for your fiat fund coupons. Sorry if I missed your live boost. We'll catch those next week and we'll put those in. And we just appreciate everybody who's supporting the podcast index dot org project. All of it keeps the machines running. How many machines we have running right now, Dave?

Oh, 18, 20, 20, 23. Yes. So your work, your work is going towards good things and people can make wonderful stuff with it. Case in point, there's Brendan. Do you actually do you use the index for your lookups? You're still sucking off Apple. Um, let's see a question. I changed that a year ago and I can't remember. It's too much to keep checking. It's all yeah. Brendan, you're sprawling, sprawling. Brendan, thank you so much.

We appreciate you being here and congratulations on all your success, man. Thank you. Great talking to you too. All right, Dave. Sorry I kept you a little longer. But if you hurry, then you can just get to that zoom call in on time. Yep, yep, yep. That's exactly what he's doing. All right, boardroom. Thank you very much for being here, RTP. I hope you enjoyed everything except the first part of the show. We'll be back next week with podcasting 2 .0. Podcasting

2.0. Visit podcastindex.org for more information. Go podcasting! Your whole product is a dirty face. Okay.