Discuss Podcast Brand Safety AI's Impact on De-Monetizing Creators | New Media Show

Todd and Rob in the afternoon.

Hey, afternoon delight with Todd and Rob. Oh yeah. Oh yeah. And here we are on the New Media Show, and I am back after being out for 2 weeks. Yeah.

So, uh, back in the studio, back in the saddle again, as they say.

Yeah, you know, and it's been like I've been in and out of here. Uh, the week before last I was closing the office in Columbus. So, uh, the, uh, the office in Columbus is no more, has been turned over to the landlord.

Saw that.

Yeah, we got, we got our, you know, got that money back that we don't have to spend no more on, on rent. And it can be—

are you sad about that? And on, on one hand, nope, nope, nope.

No.

Okay.

If you're— why? You know, since COVID there's been nobody in there. You know, I've been paying—

because I know that you were excited about getting it when you—

yeah, but you know, years ago, COVID changed everything, you know.

Oh yeah.

So we got 2 out of 5 years of good use out of it, you know.

So you moved in just in time for everybody to hit the road, right?

Well, I came back in May of '19 and then, you know, I was back and forth between— and then Boom, COVID hit in February of '20. And since then, you know, we've had— we've used it for some team retreats and some other stuff. We go in there, you know, matter of fact, we go in there and like, oh, we didn't take the trash out from the last time we were here, you know, that kind of stuff. But no, so that's, that's all turned over.

And then I have like a 3-year lease or 5-year lease. 5-year lease, right?

Yeah. That's kind of the way it works in commercial real estate. And then I got turned on to an AI conference.

Yeah, I heard about that. Yeah.

And so that's— I just, you know, pulled the trigger on that real quick because it was in Cleveland. And that's a cool spot.

Yeah.

So, you know, 2.5 hours over there and— or 3, I think, or whatever the drive is. And then so I spent Wednesday, Thursday, Friday last week and, you know, get my brain exploded. But, um, yeah, so here we are. I'm back, back in the saddle. Meanwhile, you, you know, you've been busy without me, so trying to fill in, right?

Trying to keep this, this show on the road, per se.

And, and we have had, um, we are lit and live, and we have had some, some Satoshis come in. People have made some comments while I was out.

Well, that was the whole topic of the last episode of the show.

Yeah. So Scott said when you had Lisa on, said, wow, no mention of Podcasting 2 with Lisa. I thought this is a new media show. So 10,420 sats. It was just a Twit Infomercial episode. And then last week, you know, R. Nash says, from the days of MP3, OGG, AAC, and WMA, 3,333 sats. And then someone asked again, where do Leo and Lisa stand on B4V? And that was another question by RWNash3333.

Well, and Lisa answered it online.

Oh, really?

And it was a resounding no.

Oh, well, interesting.

It was pretty clear on there. They don't, they just, they're not buying into the crypto stuff. So, right.

Yeah. Well, you know. It's not about just crypto.

I know it's not, but, but it does have ties to it with the Satoshis, right?

Well, you know, again, value for value and value for value comes in many ways, time, talent, and treasure.

So, oh yeah, yeah, there's no question. I mean, I mean, if you want to hear a complete deep dive on it, check out last week's episode with Sam.

I already, all there. I'm already in. I know, I know everything you guys talked about. So nothing new per se.

No, that's true. That's true. I think what it was is more of an exposure and exploring, um, what the PodFan, uh, yeah, product was and how that all works and how that's a terrific example of what could, you know, the industry could evolve to that kind of a model over time.

Sure.

Right.

So, you know, lots, lots going on there, but, uh, Yeah, so, and for me, it's been, uh, just kind of this whirlwind because I got back and one of the sessions they had there— matter of fact, um, I'm not sharing. So there's a whole bunch of stuff that we found that people are gonna have to dig hard to find some of this stuff. I'm not sharing on the show some of the tools I found. I'm, I'm just not. It's a competitive advantage at this point, and I don't want to reveal all that was learned. So I'll just— I guess it was good. That's all I'm going to say. I will say this, I will say this. There is what's going to happen for knowledge workers, and if you're a knowledge worker, you're basically someone that's in marketing, HR, someone that's, you know, anyone that's in knowledge work. Life is going to change completely, and there's going to be those that think and are creative, and there's going to be subject matter experts, and then in the middle is going to be this term called thunking. Thinkers, thunking, and SMEs. Today we're doing a lot of thunking. You think a little bit, all right, let me do that proposal. All right, let me think a little bit. Let me write this project document. All right, let me think a little bit. Let's go analyze this spreadsheet. That's the thunking, the middle part. The thunking is going away.

It's the transitional stage, right?

Well, in 2 to 3 years, the thunking is going away.

So because people are going to have to go from how they've been working like this today to how they're going to work like this in this thunking phase, like you're talking about, it's kind of like, are you going to move there or are you going to stay?

Well, and if you're not the SME, if you're not the special— the subject matter expert, because guess what? The thunking makes mistakes. The future thunking is not— the output isn't always going to be right. It's going to have biases. There's going to have wrong information. So you're going to need the subject matter expert to clean up but the thinking got wrong, right? And then you're going to have to spend more time on creative and thinking. So—

and that actually popped up in the podcasting news today about brand suitability. Do you see this?

Oh yeah.

It's just a little bit of an example of, of what you're talking about.

And just exactly why I thought it was a bad idea to begin with.

Well, yeah, I mean, does it speak more to the biases that exist in AI? Yeah, the question, regardless of the topic of brand suitability monitoring, so that's kind of a separate discussion here, but this is the dangers of this.

There's, there's great, there's some great studies by some great companies that have come out and talked about the biases, not only black-white biases, male-female biases, And I'm saying that in general terms. So people, you know, so the—

these bias, bias can manifest itself in many, many, many, many different ways, right? And it can be interpreted bias. And that's part of what the bigger agenda is here. And this is what I've been hoping for is, is that these AI engines can apply context. What it seems like is going on here is the brand suitability and brand safety folks are going out there saying that they're, they're all about context, right? It's all about digging into the meaning of words and, and rating and ranking things based on what they mean in the context of a phrase, or contextually or culturally what they mean, right? But what this is starting to show is, is that maybe the AI is not as good as we think it is in deciphering this stuff.

It's the model they're using has not been trained well.

Well, yeah, that's exactly what I'm trying to say, Todd, is that it's just not, it's just not there yet.

Well, again, again, it's dependent upon the people that have done the work and have put the guardrails and put—

are there guardrails in any of these platforms? I think it's a good question.

There are, you know, there, there, there are. And if you're going to be—

who's, who's setting those guardrails?

Well, you, if you are developing your own solution You said—

I get that, Todd, but there's no standard guardrails here that exist in any of this.

Well, there are some, and don't get me wrong, it's not perfect by any means. Because here's the situation that's going to happen. Is— and a lot of it has to do— and if we're boring you, I'm sorry, audience, but there's some context you have to understand about large language models. What, how those large language models are essentially created is, here's a question, go collect a whole bunch of data and give me an answer. And they do this tens of thousands of times and they go out on the interwebs and they look at publicly available information and they index it and they try to, the model tries to reason and tries to come up with the context of. And this is real simplified, it's more complicated than this. But when you get done, when you get done with these question and answers and how you're building the model, the model is locked tighter than a— there's no way to pull data out, there's no way to put data in. You have a, a model of information. It doesn't contain verbatim text of books and so forth. It's taking the concepts and all this information and come up with the a model that then when you ask it, here are 3 resumes, who's the best candidate for a job? And let's say those candidates were complete. Let's say those resumes were nearly equal. And the difference between them was maybe there was 2 male resumes and 1 female resume. Well, as the model has been built, it has went out, looked at thousands and tens of thousands of resumes, maybe from a specific job title of that type. And because of the history of the internet and everything else, it may find more information about men in those resumes. So, so if the guardrails are not put in place correctly to remove those biases, it's going to—

in this case, you know, white men, right? Not as many black men.

It's going to pick the best candidate is a male. And there's gonna be these biases that are gonna be built in. So this is where whoever is designing the software that's pulling data out of a model, and if you don't have some sort of guardrails, it's going to go off the rails. And so this—

Well, it's being trained by data, right?

Yeah, and it's been trained.

And I guess, At the end of the day, it seems to me that if society and if the information that exists that it's being trained on has certain leanings or biases or directions that are inadvertent, they're probably not done intentionally.

Well, it's also, again, the model is built by people and people are coming up with the questions and the queries and the things that the model is built upon. It's very simplistic in its nature. But again, in the end, no matter where we're at, because right now we're using the most least capable AI we ever will, or language model. If we, if we think about that, we're using the least capable today. There's nothing will be more dumb than it is today.

Well, it's only reflecting what we give it, right?

That's true.

So what we're expecting it to do is know about cultural change.

Well, again, we'll see.

And that's not necessarily part of the dataset, potentially.

I'm not a scientist on building large language models. It can probably talk for a year and a half. But in the end, if you're a company employing any type or getting to the point where you're going to be working in this space and using tools, you need someone that is very— you're gonna have to have a very high-end creative and design part of your team, and you're going to have to have an absolute subject matter expert to review the output of any— just kind of as you do today in business. What happens today, you come up with an idea, you do a, you do a project document, you turn it in for review, it's signed off on, and it's shoved off. It could have mistakes in it as well. But I think it'll become more critical as time goes on. But anyway, that's basically all I really want to talk about on the conference. Everyone else can have to pay for the pleasure to get the knowledge.

Well, let's just back up and get really clear on, you know, my understanding of how the AI, typical AI language models work is that they learn They're basically an algorithm that was written to learn and process information that it's given.

Well, it, it goes and scours information that—

well, that's exactly what I'm saying. It's taking information that it gathers from whatever the source is.

Thousands of sources. Yeah.

Right. And, and it formulates kind of its approach to this. I mean, I, I, I do queries of ChatGPT all the time. And I, I see certain biases that are in their results.

But the thing is, it's not a full, full spectrum output, right?

It's, it's, it's only, it's only going to tell you what it's been trained to tell you.

So I would say if you're using it in a raw form to ask a question and get an answer, you're using it absolutely wrong because that's going to lead to trouble. You, if in a business sense, you better feed it something. Okay, as an example, I had an article to write that was due, and I gave it 26 things that I said, this is fact, this is fact. I gave it a statement, such and such is a fact, such as that. And I said, take all my facts according to Todd take all my facts and then create me an article that is this many words long. And it did. And then I spent an hour going in and shuffling and editing and making sure things are right and it made sense before I moved on. It definitely saved me 2 or 3 hours worth of work. But I never ask a raw question without providing it some sort of input data that I say, this is, this is the source of truth. I don't ever say, what is 2 times 2? I don't ask it that. I tell it 2 times 2 equals 4. And I give it the information that I want it to use. And then I go make sure it didn't go grab some external source of information. I think it's very, very dangerous to ask a question and ask for a raw output. Yeah, you can make it ask for a topic. All right, I'd like to— I'm going to do a podcast episode on X. Give me some— give me some questions or give me some thoughts on what I should talk about with this topic. Probably pretty good for that, but it's going to sound a little messy. It's going to be general. It's going to be general.

It still sounds messy.

Um, but the key is Most people are not, do not understand how to write prompts. So the prompt writing is going to be critical if you're going to succeed using these tools successfully. You just can't write it a one sentence prompt. You got to take time and think about what you want to accomplish. You have to write a complex prompt using your creative brain to give it as much descri— it's like talking to an intern. If you've got an intern that has never ever done a task, you have to tell the intern, okay, what we're going to do is we're going to take this data, we're going to analyze it, and I want you to write a summary using the data from lines 82, 86, and 87, whatever. You have to tell the intern exactly what data you want to use. And I want it in this format and I want about so long and I don't want to use any external references. You have to write the— you have to tell the system to write. You have to treat it like an intern. And if you keep that in mind, an intern that knows nothing, it's a day one on the job, and I think you'll be okay. But again, if you're not feeding it anything and just asking for a raw answer, I think you're asking for trouble because it will make stuff up. Yeah. In a big way.

Well, it has this concept of hallucinations that happen too.

Right, right, right.

It doesn't always come up with factual information. So, But to bring the conversation back a little bit to the brand suitability stuff, you know, this, this context is, is costing, you know, some podcasters money.

Well, then they need to choose—

they're being flagged for things that are, that are not correct.

Then they need to choose not to work with that vendor.

Well, I thought it was interesting that Sounder AI was kind of the the tool that was behind this disclosure about certain other tools like Sounder AI that are doing the brand suitability that are not understanding. Just give an example, Urban One kind of was being, I guess, discriminated against based on a cultural difference, right? By using the term bomb.

Yeah, like the bomb, right?

That means in in certain segments of culture means it's cool, right? But into other segments of culture, it means an explosive device because, because the thing, their model is so weak that it doesn't understand context. But I'll be frank with you, I've talked to a lot of these brand suitability platforms, and they will tell you that that's, that's the reason that they exist is to understand context.

Well, that.

And it depends, but obviously it's missing on this.

Well, it's been pointed out again. Are they— what model are they? There are literally— now you can go out on the web and there's marketplaces for them, literally 1,000 different models built for different things. And if they're using a model that someone else built, they, you know, they have no idea. They have no idea. If they built their own model, then they screwed up. Simple. If they're using a third-party model, they're using someone else's model to— and they don't have the right controls in there. This is hard to get right right now.

Yeah, but I, I thought it was interesting, the more competitive dynamic that's going on here. Sounders saying they're not doing this, but somebody else is, but there's really only a couple other players, right?

Of course, of course they're going to say that.

Well, I mean, they could have these problems too.

Rob, not millions. Billions of dollars are on the line. Billions of dollars are on the line. Who gets this right? We're not, we're not, and ultimately trillions of dollars. Now, the biggest development was Meta releasing Llama 2 open to the open market. That That has sent shockwaves throughout the entire ecosystem. So, you know, again, in privacy and data, data can, you know, data rights, and it's just the topics can go on for hours. And I'm trying not to let it become obsessed by it. But all the only advice I can give is don't sign no long contracts with nobody. If you're going to use a tool, go month to month because next month something's going to come out that's going to replace a tool that we think is fantastic now. Nothing is going to— this is going to do this for the next however long. It's just going to jump from, from then to then.

What's a little bit disturbing about this news that Sander AI was associated with is that It's, it's totally trying to play off a little bit of a, of a race, uh, card here to some degree, um, pointing out that only Black creators are being discriminated against here when maybe it's more shows are being discriminated against based on this AI not having proper understanding of context and culture in all aspects of society.

If you, if you're being forced to use this, I'm sorry, you're screwed.

Well, there's shows that this is being used on that don't even know are being evaluated like this.

So that's a problem too, right?

Well, so there's inherent issues in all this, and this even goes beyond moderation. This is around using these tools as a weapon to decrease their ability to monetize, right? Which is a whole nother level of—

We talked about that. Yeah, we talked about this on the show. Beforehand, of course.

Right, right. It's beyond just kind of moderation, right? Because this is financial blocking.

If you're a creator and you have to warn your guests, hey, don't say the bomb on the show because we're going to get demonetized. That was the bomb. Right. You know, or whatever other, you know, that's a word I've used before. That's the bomb. I've used that before. You know, so yeah, it's you— this, this control of this and this, this untruth and of these companies untrusting content is ruining— it's going to ruin podcasting.

I would imagine all these platforms don't even know that these problems exist until they get exposed like this, right? I mean, It's, it's, I mean, how would you know?

It's going to ruin, ruin, you know, and guess what? Okay, you, if you chose to go down that route and live that route of monetization and you have to watch your words, what, why, why do this? You know, why, why, why would you be— I, I refuse to be censored. I'm careful in my own way of what I say. Mm-hmm. But I don't want to even have to have the thought that someone's looking at my content behind my back and saying, oh my God, that show's non-monetizable or unsafe and give it a rating just because some stupid large language model can't understand slang or cultural references or whatever it may be.

Yeah, it said that the standard keyword block list of this research that came out, um, claimed that 92% of Urban One's podcast content was potentially problematic and it couldn't be advertised on.

That's insane. Absolutely insane. Someone needs to be fired.

Uh, it's, you know, we're, we're putting too much trust into these.

Absolutely. You can't trust any of this. That's why I'm saying if you're going to employ this as a company, you have to have a subject matter expert that's looking at every output that you're getting from these platforms.

As an example, is this a legal liability, um, to, to those tools that Urban One could file, like, you know, a lawsuit against them?

It depends if they actually lost money because of it.

Yeah, you know, it clearly could be seen as damages, right? Yeah, there was damages incurred in the misunderstanding of the cultural references, right? Yeah, this is a Pandora's box that, you know.

Oh, and there's going to be lots more of these. There's going to be lots more of these Pandora boxes for sure. And talk about Pandora's box. Did you see that? Sounds Profitable today announced their own newsletter.

I thought they had their own newsletter.

Well, they announced a new— it's called, uh, I think it's called The Download.

And I thought they had that for a long time.

Well, there was a post that came out today. I got the email this morning, and as I read that, you, you have to read it. And I am sure, I am quite sure. Don't eat me, James Cridland. Someone is very irritated. When I read through their description of what that was, there was some direct shots in there. Like, we're not going to commercialize. There was some language in there that made me go, hmm. Oh, you know. Um, very interesting commentary. I will just— it was interesting how it was written.

I'll just, uh, so is it a, like, a daily newsletter? Okay, so that would directly compete with Pod News.

It says, what makes The Download different? Three keystones of The Download are respect, objectivity, and ownership. Um, I hope you notice, sounds proper with current newsletters and completely ad-free with our Pride Edition of The Download graciously sponsored We continue that ad and clutter-free focus for the daily edition.

Uh, right, that's the key right there.

The download Monday through Thursday with Spreaker from iHeart providing 4 unique data points that will be able to trend week to week.

As a reminder, um, and that sounds like a podcast, Todd.

Well, as it's a newsletter, as a reminder, neither partner is given additional preference or influence Okay, but Spreaker doesn't, doesn't distribute it, distribute newsletters. This is not Spreaker, this is Sounds.

No, no, but you just read the word Spreaker in the context of talking about this.

Yeah, Spreaker, that's a podcast hosting. Spreaker and iHeart are providing unique data points.

Data points. Okay, yeah, yeah, okay, yeah, got it, uh, got it.

Okay, so Anyway, it was just, it was just interesting how it was kind of written. And, you know, they were talking about how it's going to be impartial and, and so forth. It, but it, it just, to me, it, you know, I read things maybe I shouldn't between the lines.

So, well, I'm sure there's some people get mad at competitive interest here, right, between these, these platforms. And there's a bunch of other podcasting newsletters out there too. But yeah, but yeah, I mean, I think, you know, if you think about it, Sounds Profitable started in a partnership with PodNews, right?

Yeah. It says on the company level, Tom and I believe it's important to disclose that we've never received outside investment. Neither of us have or plan to hold board seats, advisory equity, advisory roles with any company. So there again, I think that's another dig. We don't offer any sort of work on commission in our personal investment portfolio managed by outside firms, which is, that's, that's an interesting disclosure. So, yeah, so they're trying to be very, very transparent on what it is and what it isn't and whether it will be influenced or whether it will not. And it's, it's a, it's just, it's an interesting write-up. And, and I, I read it twice.

So awesome.

But anyway, they're going Monday through Thursday on it.

Oh, Monday through Thursday. Okay. Yep. So they're skipping Friday and taking Friday off.

Yeah, that's, that's smart. No one does hardly anything on Friday anyway, right? We switched to half-day Fridays. We, the Blueberry team, now leaves at noon on Friday. We do 36-hour work weeks now.

Wow.

Okay. Yep.

It was, uh, because, you know, it's just one of those progressive of you, right?

Well, it's just one of those things, you know, it's like who gets really a lot of work done on Friday afternoon? If there's something critical, I'll do a meeting, or, you know, if there's something that comes up, we will. But by, by practice, the team's out by noon. So yeah.

So I guess it's, it's becoming more public that, uh, TikTok is going to be joining podcasting in some capacity. Now that's, that's, that's been a rumor and a topic that's been around for a very, very, very long time.

Well, now that there's emails are on the street and there's actually an example, I guess I can talk about it. I, I've known about this for a long time.

Yeah, it's been a couple years now.

Well, how they were going to implement it has only been really finalized for the last couple of months, and I think the approach— I don't know if I can say how much I can say But the initial test, I think their approach is interesting. Let's just say this, you have to create a post, you have to create a basically, you know, some sort of TikTok video. And then within, they get a special, they get special access to link to the media file in the, uh, in the TikTok app. So what happens is, is you do a, you do a social TikTok post and then you link— there's a link to the media file in the app that allows you then to, um, as the, as the video is playing for the promotion, you can then click in and actually listen to the full episode. So basically you build a promotion for the episode in TikTok, and then you're able to listen to the episode right in TikTok.

Yeah, I'm looking at a screenshot of their, of their interface. And it does say go to upload icon at the top right-hand corner of the app. It says, click the, the link podcast tab. And it says, enter the link to your RSS feed and then and select an episode, I guess. Right. So it's not very clear if it's talking about video or if it's talking about audio.

You have to do— it's audio. You do a video TikTok and then you link to your audio podcast. Okay. And then you can listen to the audio podcast in TikTok as long as you don't— basically, if you flip off though, you won't be able to listen to it. But you can come back to it. So there are some things going on there. Now, I can't probably say more than what I'm allowed to say based upon what I know right now. But basically what you publicly see is what I'm going to publicly talk about.

So they're, they're saying that you can select the videos to link to your podcast.

So you basically have to have a TikTok done to link to an episode.

Right. So it's probably going to show up in the menu on the right to click to get to your podcast in some way.

Yeah. So basically you, you are going to create a TikTok post and then you will click in their interface to link to the episode that you want to be.

Well, it's the RSS feed and I'm assuming it. So you have to associate the episode to a particular video. That's what you have to do.

That's correct.

Right. And then what's the experience when you click on the, on the actual podcast link? I wonder.

It plays.

Does it go to a separate?

No, it plays. The audio just plays? Plays within the app. And I don't, I haven't seen, but my understanding is if you stay on the, stay on that thread. You—

it'll play within the app, but I think it'll probably take over the screen, so you won't be watching a video anymore. Someone's going to have to show us, uh, I'll be watching, watching a video, you know, what'll take over as an audio.

Yeah, right. So it's essentially a way to promote your episode in TikTok, right?

Yeah.

Okay. So, and I think that's probably acceptable from an implementation standpoint.

Well, it's a pass-through, I'm assuming.

Or is it? I, you know, I can't comment on that today.

I could see some similarities with what YouTube is doing here. So, yeah, definitely.

Yeah. So it's— but it's more— it's— but the thing is, you still have to create— it's— there is going to show— it's not going to show up on TikTok as an episode. You have to create a promotional— you have to create a TikTok— I don't know what they call it— an episode post, whatever they call it.

Well, or, or associated with a video you have published is what it probably says here.

Yeah, yeah, right, right. So here's the thing, I think what they want is feedback. I am 90—

the pathway for that.

What's the pathway? I'm giving 99% sure TikTok is listening to the community. Yeah. And if people have feedback, feel free to send me an email and I'll be happy to pass it forward to the TikTok, the ByteDance representative.

So, okay. Uh, let's see here. Yeah. So it's coming. There's no date for when this is going to roll out.

Um, my understanding is limited beta at this point with select partners. Right.

These folks have been talking to podcast hosts for a long time.

A long time, right?

I'm not saying who, but just, well, a long time they've been talking.

They've been talking with us and, you know, we've been giving input. Yeah. So, you know, but TikTok's gonna do what TikTok's gonna do.

And there's increasing talk about some— something might be coming or happening around the same thing, uh, with X. So we'll see.

Yeah, we'll see. I've been hearing about this for, for a while now, and I think what it really boils down to is, in my opinion, what TikTok's trying to do in this approach is allow a person that has a TikTok channel to build exposure to their recent episodes through the platform and give them a viable way to make people build awareness to a show. And if they— and this is where I think they're not going to get too much pushback from the community because they're using their platform as a way for a content creator to promote their podcast episode that's delivered via an RSS feed. Now, they don't give them the ability to subscribe to the show that in, in TikTok.

There's going to be, my understanding, well, who knows, or follow it, or follow it, or like it.

Well, you can follow the TikTok creator And maybe that'll come back up in your feed again if the logarithms are right. But I think those that are in this test are probably going to have some pretty distinct advantages here because they're going to want— I'm sure I would. This is my suspicion is that TikTok is going to surface that content more to see how this plays out. Mm-hmm. So being on—

because right now you can hurt, right?

But being on the beta, you will have a distinct advantage for your— not only your TikTok account, but for your content, because, you know, TikTok's trying to figure this out.

Yeah, I think it's the bouncing between video and audio that's going to be interesting to see how they kind of do that.

The biggest challenge is, um, not a lot of— believe it or not, not a lot of podcasters have successful TikTok accounts. So there's a distinct opportunity there.

It's hard to focus on everything these days, Todd. I think that's— I agree, that's, that's, that's the bigger discussion the industry needs to have. And, and certainly AI can potentially play a part here.

Well, you can use tools to help, to help create content for TikTok, for sure.

Existing content.

Yeah, like Memento.fm, the one that I use for sending stuff over to TikTok, you know, is a great way to do this. But still, even with Memento, the way it works is I get the TikTok video ready, sends it to TikTok, then I have to go into TikTok and approve the post. So from that instance, if I was part of the beta, I would be able to link my media file and, um, and publish it. And so, you know, the tools will help you in, in getting exposure. So this is what it really is. I think it's, it's smart because you build exposure to the original content without doing some— without calling it podcasting, you know, without bastardizing the, you know, without— that they didn't build a directory and they didn't build a follow and subscribe function. They built a promotion function.

Which rides along with existing content they already have on the platform. Right.

Yeah.

So that's kind of what YouTube is trying to do too. Yeah. To some degree, I think if you think about their strategy as well. So it's, it's like, it's a little bit like a companion strategy going, going forward. It's not a dedicated podcast app.

But you know, it's every traditional— is the way people keep talking about YouTube, you would think that that's the second coming of Christ. You know, it's just, it's just such a huge platform, Todd.

I mean, there's just so much, so many users.

Yes, yes, so many users. But again, talk to average podcasters and they're getting no traction on YouTube. Average podcasters that are pushing stuff on YouTube, they're getting no traction. Zero. 10, 15— they're not getting traction over there and they're not getting discovered like everyone claims. I, I think it's a house of cards, to be honest with you, because you're lost in the shuffle of all the other content that's out there. There is so much content on YouTube. Do you think Discovery is bad? Good, good luck. Unless you are specifically looking for a specific show and specific content, you will not find it if you do a search. Who do they surface? They surface big shows. They search— they, they surface the shows that are following the YouTube algorithm. They're not surfacing small podcasters' content.

It's a recommendation engine platform, which is different than most of the other podcasting platforms. I do believe that— I think we've heard some news out of Spotify that they're kind of heading that direction as well. So I think what they're— Spotify is trying to mirror, I believe, and build a YouTube competitor is what they're trying to build. Yeah. Okay. And, and that's going to, that's going to have those elements in it where there's going to be suggestions and algorithms that evaluate what's presented to what user, right?

Probably the big shows. Again, the average podcaster is not going to be found, and then your content's going to be monetized around it, of which you'll get not a dime.

I would say that that's probably coming more generally to podcasting too, is, is that they're going to be driven more by algorithms.

Well, again, I go forward, I pay for premium, but as soon as I go off premium, I log out of my account, I go over to my show or our show on YouTube and I press play, I'm subjected to 60 seconds of advertising that I get zero pennies for before the episode starts.

So, so have you signed up for, for YouTube monetization yet, Todd?

We don't, haven't reached the number. It's astronomically high number of views you have to get. Go ahead, try to get those view numbers to sign up for YouTube monetization.

It's 3,000 hours.

3,000 hours, right, in a 12-month span, right?

Okay, so I just, I just crossed over the threshold to monetization for my personal YouTube channel just this past week.

Well, that's good. You know that you're in like a 5% category then that have been able to do that. So again, it's— unless you have some content on there that's getting significant views.

I've had a few. I think my total views is like 750,000.

So, well, that's, that's good, you know, for a lifetime. But the, you know, yeah, it's been there for 16 years. So, right. So again, and but then, then now you're out. What I want to hear from you later is, okay, how much did you make last month from that?

Well, that's what we're gonna find out. Yeah. And I'm also creating more content. I'm publishing more content, you know, like we're pushing this, this show over to my, my YouTube channel now. So this is an experiment. And I've got a couple other shows that I have over there as well. So I'm actively creating content over there.

So we'll see if you get a $6 or something like that. I actually, you know, I, I read that people were getting like 0.00.

I don't think you even get a check until you reach $100 though, Todd.

Oh, so it might be a while.

It could be. I'm not, I'm not really sure that I totally understand how the, how the monetization engine actually monetizes. I mean, like what share do I get per view?

And I heard it's like 6/10 of a cent or something like that.

Okay. I wasn't, I'm not really sure yet.

Like CPMs of like $6 or something like that. $3 to $6 is what I heard the CPMs were. Oh, okay. So, I mean, I don't know if you get a portion of that or how that works, you know?

Yeah, I don't know either yet.

So if you're getting, if you're getting 3,000 views a month or something, then, you know, maybe you'll get $20 if you're lucky.

I suppose. Yeah. So Todd, I did find it a little, little amusing and interesting that the number of episodes being published on Open RSS by Spotify for Podcasters is to the lowest level, um, ever.

Okay, say Open RSS.

Yeah, I mean, oh, episodes being published to Spotify for Podcasters.

I'm not surprised.

Is at the lowest level. Not surprised. There's ever been. And, you know, that's the formerly Anchor platform, so this has kind of fallen out of their rebranding of Anchor, right? Which has proven to be a pretty strong engine for new podcasts and things like that. I mean, granted, it doesn't have those podcasts typically don't have a lot of traction, but this is, this is specifically talking about the number of episodes being published to, to Spotify for Podcasters is way off.

I'd like to see, I'd like to see the delta. I'd like to see what's being published in the Spotify that is the ratio that is not making it into Open RSS.

Yeah, but it still hosts 1 out of every 5 new episodes being published into the medium. And Spotify's Megaphone platform posted another record high for episodes being published. The biggest increases were seen by Omni Studio.

Well, they're going after shows in a big, big way., to monetize the big shows. So Megaphone has a lot of big shows over there.

Yeah, and it's an expensive platform to host on too, right?

So, you know, it's, it's one of those situations where it's a case of the haves and the have-nots, and the have-nots are shows that are your average, you know, I don't want to say it that way, but it's basically that it's almost wholeheartedly ignoring most indie content creators.

Yeah, I don't think you— yeah, I don't know what the total number of shows that are hosted on Megaphone is, but I think it's, it's been, I don't know, 40, 50,000 maybe max.

But so it is a good, it's a good number. Yeah, it is what it is. And, um, meanwhile, it's, you know, I look at our July results and the green line was above the red line.

So I guess that's the That is a good sign.

You know, that's, that's the key. Right.

And then it said over the last week, this is off of PodNews. Thank you, James. And James is going to be joining us on stage for our live media show at Podcast Movement coming up on the, I believe it's the 23rd of August coming up later this month. Yep. And so I guess over the last week, there's been 195,804 podcasts published. At least one episode. So just in this last week.

Well, let's go over and actually look at the, the system, the overall numbers, right? 113,000 last 3 days, 232,000 last 10 days, 336,296 in the last 30 days, which is really like way down. 423,060 days, 480 5,000 in the last 90 days. These are shows that have published a new episode. That 336 number is shocking.

Yeah, that's like half what it used to be, right?

Yeah. So, and, and I probably can— I have to find it here. I gotta load Slack up.

Well, we're— I think we're going through a, a massive reset of this industry, um, you know, I hate to use that term, but I don't think we're going through a reset. I think what we're going through, well, it feels like it, it feels like everything is kind of, you know, shifting. Companies are cutting back. Um, there's, there was an, you know, a post in PodNews that came out, um, that was talking about, you know, the effect of corporate dollars on podcasting. Right. So, you know, was there a lot of investment that was made in podcasting that was, that in hindsight wasn't wise? And I think that that's what this article points out. It was, uh, an article written by Rebecca Sanez, um, in, in PodNews. And it was talking about kind of a good example is like the Harry and Meghan, um, show that the Spotify was involved in and how that, that really turned into a flop and And, um, and, you know, you just— there was a lot of investment, I guess, that has turned out to be not wise. And that's why we're seeing a lot of pullback by Spotify right now, right? And other companies.

It's not just Spotify, but the indie space and the listener space remains strong. And, and matter of fact, there is huge gains in download numbers. We're seeing across the board shows are growing.

Well, audience is still— oh, absolutely. I mean, I agree with you 100%. I'm just, you know, it's still tough time even for indies though.

I mean, I think we're seeing overall numbers kind of not doing as well as listener numbers. Absolutely not. Listener numbers are—

any, any content creators have been the beneficiary I agree, but there's a smaller pool of creators because— right, right, right, right— this pullback, right? And that's shifting audiences. The growing audiences are, are basically big shows are getting much bigger now, and, and smaller shows are growing as well, bigger too. Yeah, right. The ones that are actually still publishing, right, are getting, getting the fruits of the, of, of the market as The competition, you can see, you can see where the line hit about, about May of this year.

It's been falling off ever since May, about, about, well, no, about March, about March. It's just continually dropped off. Now I see it's interesting here. July for us, we were up and it reflects here a bit of a bump.

Up as new creators?

New creators. Yeah, up in new creators. So, you know, so I— it was interesting to see, you know, it's, it's been, it's been pretty flat, you know. Yeah, it does. And, you know, I was, you know, I'm thinking, okay, what, are we going to see a 2% or am I going to get a 3%, you know, growth this year, or am I going to be zero? Um, you know, and right now I'm a little bit ahead of the curve, so, but July was a great month, at least for my company.

Interesting. Yeah, that happened, happened with you, but maybe not across the board in the industry.

I don't know. Well, I think some of the stuff that we've been doing and the investment we've done on all the stuff that we've been adding is starting to pay off. You know, starting, we continue to add, you know, we're adding lots of features and, um, So, you know, we're seeing, you know, I wouldn't say hundreds, but we're seeing, you know, significant numbers of shows starting to migrate. But there was some announcement, there was some discussion on one of the forums about Podcorn seeing a huge drop, um, this past month in the number of shows, or, uh, and again It was basically using a redirect. The numbers shows using a Podcorn redirect fell through the basement for Podcorn. Oh, um, so, uh, John Spurlock said, looking at the change in mix of hosting companies using Podcorn from May to July, looks like Buzzsprout went from 53.49 to 11.48%. So must be Buzzsprout was using Podcorn for monetization. Um, and so Buzzsprout went way down, where RSS.com using Podcorn went up. Of course, if one goes way down, the others are going to get a bump. But, uh, a number of Libsyn shows using Podcorn for monetization as well. So there was It's quite a, quite an interesting number here. Um, so I, I, it's interesting that they, they were— I wonder if they switched, uh, to a different ingestion partner or if volume dropped off like a rock. And programmatic, our programmatic is booming. We're not seeing any decline in programmatic whatsoever.

So it's good in, in just overall usage, or are you talking about CPMs or CPMs usage, everything, fill rates?

So SoundStack continues to do a good job for us and keeping the, keeping the ads rolling.

But so is SoundStack using any kind of brand suitability stuff at all in their platform?

You'll have to ask SoundStack. Yeah, yeah, I'm not sure. To my knowledge, no, but I could be wrong.

I know that those, those platforms are out after everybody, so yeah, well, you're gonna have to decide, you know, if you want content that's being filtered.

If your programmatic drops off Blueberry, let me know. We'll look into it.

Yeah. Yeah. Cause I'm, I'm pretty sure that that filtering is happening at the buying level, right? Potentially. Yep. So it's not going to happen at the host level. So, right. You may not even know that you're being filtered against. You just don't know.

So, and it, um, and I'm sure there's— yeah, I don't know. Again, we don't guarantee fill rates or anything. It's, you know, it's—

well, you can't. No, I mean, there's no way that you have any way of controlling anything.

I can kind of look at the average when it comes back, but I don't— I, you know, I could probably look show by show by show by show and do, you know, comparative analysis and see if one dropped from one month to the next, but payment rates look to be about the same historically as, as, as, you know, when shows get up and get on, you know, it takes about 60 days really to get on performance. And, and once they're in there, then you kind of each month you can almost kind of predict what a show is going to generate revenue.

Yeah, I thought it was remarkable, this Edison Research report that came out about kids podcast listeners.

Yeah, I didn't read that yet.

Yeah, it says 29% of kids aged 6 to 12 in the US are monthly podcast listeners. To what? Oh, that's a good question. It said 87% of them have shared something they learned from a podcast with others.

Now, I don't know how they know that for sure, but that's in polling, I'm sure.

Yeah, that was a question that was asked.

But how do they How do they poll 6 to 12-year-olds? It's not like I'm going to let my 6 to 12-year-old on the phone. So they're probably getting this from parents. Yeah.

Parents use podcasts to foster family connections, said among parents of children, monthly podcast listeners. So it does look like that they spoke to the parents about their child's activities. Right. So it says, um, kids podcast, uh, to be enjoyed by the whole family. It said 71% said it is important for kids podcasts to give parents and children topics to discuss together. So there's a bunch of detail which I don't have in front of me.

So by the way, by the way, Rob, did you read the New Media Show's blog post summary for the last episode? Did you— have you been reading those?

No, I didn't actually.

Yeah, you should go, you should look at what, what the summaries have been.

So is that on our, our page?

Yeah. So as an example, here's the last one. I don't know if I want to read it all in entirety, but it says Rob Greenlee and Sethi have an in-depth discussion about new monetization models for podcasting, including micropayments, the value model, and using Satoshis. They talk about platforms like Podfans. So you provided me the first paragraph, right? Okay.

Then you use that as the, as the trigger. Now it actually pulled that from, pulled this from the episode content, right?

This, and then the rest of this was pulled from the transcript. Sam's given an overview of the value concept. Podcasts set a value for content. Listeners can pay that amount or adjust it up or down depending on what they feel their content is worth. This facilitates micropayments. So I should have added in there, this was for Podfans. So there's missing context right there. So this, Sam explains how his Podfans influence value for value using set micropayments. Listeners pay by the minute. Duh, duh, duh, duh. This rather than a flat subscription fee, this gives us just more control. They just come to challenges around getting people used to digital wallets and transacting in Satoshis, talking about how the process can still be complex. Sam argues that with time exposure, Satoshi micropayments can become ubiquitous to sending an email. They talk about how you and him explore how values relate to Bitcoin and the controversy around cryptocurrencies. They about friction in government. They shift to how podcasting can better market these new standards, like the new namespace talks about the podcast standard projects needing leadership, discuss where big platforms like Twitter could potentially accelerate. So it goes on and on here for another 2 or 3 paragraphs. Do you think that's a pretty good summary of the last episode?

Yeah. Oh, that sounds, sounds good. So how did, how did it know? What section of that was me versus Seth?

Did you have to tell it? Nope, I didn't.

Really? You didn't tell it?

No. So it just recognized my voice? It didn't even have a voice. It had the text. There was no audio used in this whatsoever. I submitted the text, not the SRT, the text version of the transcript.

So how did it know it was my, my comments versus Sam's comments? Or did it not even do that?

It did. I didn't have to edit any of the names here and change out. So what I did is I got this summary from Claude, Claude.ai.

Claude. Did Claude know, know different voices? So let me play around with Descript, and let me actually have to tag Descript when it generates a—

all right, let me go back. Let's find the podcast, uh, New Media Trends. Okay, um, let me find the query I gave him. Okay, here was the query I wrote: provide a title for this podcast transcript between Rob Greenlee and Sam Sethi. Also provide a summary of the podcast along with a 2-word SEO keyword in 143 characters, secondary SEO summary for Yoast. I didn't use the title that it gave it. I used the one you gave me. Got it. And so I did tell Claude.ai who the two participants were. And got it. And so I guess it's sorted through who Who was who. Okay. And free. Cloud. Cloud. Free. Free for anyone to use.

Any text. Yeah. So who owns Claude?

Is it its own company? That's another company. Yeah. Okay. But it was kind of cool just dropping the text of that. So what I did, because the way our system works, is I don't get the transcript until after we publish. So I put the first paragraph in there, let it publish, got the transcript, and then added that additional summary like 10 minutes later to the episode. So I think the summary was pretty doggone good, maybe with the exception that I probably should have put some context in that Sam was talking about pod pants at the very beginning. Giving an overview of the value. Probably missed that in that first paragraph that I should have said he was talking about podcast, but I think it did pretty good. Now this one, uh, the one was done by— I did a little different summary for the one with Lisa, and it didn't give me as much stuff for Lisa's episode. So again, it's about It's about writing the— but I think this is— I think these show notes are much better.

Did you specify the amount of words that was generated?

No. And basically it kept giving me more summary. And I just— because I said continue, continue, continue. Zero or more. And it got to a point where it started repeating itself. So, you know, about the third portion of the summary, it kind of like, I need it. I'm like, okay, well, it's, it's trying to give. So I just basically took the first two snippets of summary that it gave me and did minor editing. But, you know, I don't know if it sounds like a human, but I think it sounds pretty good. Yeah, that's good. So I didn't create an image for this episode though. I didn't do anything there. I've been playing around with Midjourney and it's been mixed results.

So you think it's getting better or do you think it's— Midjourney. There's a big upgrade coming, right?

I think Midjourney is extraordinary. You, and again, writing the, writing the, for a better word, right, writing this, the, the description is better. So like this one here, this was an image from the Midjourney created, and, and I basically, the title of this, SEC wanted Coinbase crypto delistings. That was the title of the episode., and I asked it to create an image of the SEC building in kind of a dark twilight. And so this is kind of what it gave me. Now, it did not get the SEC building exactly right, but the SEC building, as I looked it up on Google, is similar to this. Someone that doesn't know wouldn't know this is the wrong person. Would know the difference. Right. Right. And then the second half of the article was talking about how the FBI discovered that it was using outlawed software. And what I asked it to produce, and this is an image of, was supposed to be a picture of an FBI agent. Basically spying on an American. So this is kind of what it came up. So this is not perfect, but, you know, it's got the flag on one side and the guy's looking down at a mobile device and with a tie on, kind of looks like, you know, he could be a typical FBI, but it didn't really do a perfect job. And so I've been having kind of fun with the the Midjourney piece and coming up with album art. And that's another thing too. Apple's watching for shows that are using episode art. So if you're not creating episode art right now, you better get to it.

So what does that mean that they're watching for it?

Well, I got an email from someone over there. Well, I guess I can say who it was. Ted sent me an email and said, hey, your album art is only 300 by 300. He said, you, you need to have your album art be 1400 by 1400 minimum. Minimum for your episode art. Yeah. Right, right. So because always before I didn't want— I made them smaller so that we didn't— it wouldn't affect the page loads for Google and all that. So that's why I had done it smaller beforehand.

So on the episode art, are they still looking for 3000 by 3000? 1400 by 1400 minimum. Minimum.

So you could go 3,000 by 3,000. 3,000 by 3,000, you could.

But I think you have to keep your, your physical file size down to what, 500K or something like that?

Yeah, and if 1,400 by 1,400, like an image like the one that's on this post here, is about 160K, and that's not square anyway.

That's, that's not even the right format.

Well, this is square. This is, uh, This image here is 300 by 300.

Yeah.

Oh, it is.

This one on the screen, it looks, looks vertical or rectangular.

Yeah. It's correct. Maybe it's just because the building's offset a little bit. I didn't get it centered right, but.

No, but the actual image on the screen looks like it's a vertical image, not a, not a square image.

Yeah. But it's, it's, yeah, it's, it's square though.

Optical illusion, Rob. Or it could be just the compression that's happening because of the screen.

Potentially.

Yeah, I use Photoshop. I'm seeing here because I do see us in my, my Zoom, we're kind of compressed.

Oh, okay. So the image that Midjourney comes out with is 1024 by 1024 is the default image size. Right. Right. Then you have to upscale it. So, and what I do with the art for the the actual video is I upscale that to 1920 by 1920, then I take a segment of it that's 1920 wide by 1080 deep to get the right, the right ratio for, for the video for it to be 16 by 9.

Well, Todd, I'm, I'm looking at the, the YouTube kind of streaming page on my my account inside the YouTube Studio, and it's showing the, the amount of views that I'm having and the average view duration. Okay. And it says so far that just to my stream that there's been 27 views, and, um, and it's— and it looks like the average view duration is about a minute.

Yeah.

So let me— there's a lot of people passing through checking out the show.

Let's see here. Uh, average viewer duration on us right now is 18 minutes and 29 seconds. On what platform?

On YouTube. On YouTube.

Oh, that's interesting. Yeah, average view duration. But there's not a lot of people on here.

Off of, off of your—

yeah, your distribution. Yeah, my version of YouTube.

Yeah. Right.

So it's interesting. Yeah. In 18 minutes, uh, it looks like there's 4 currently live on the YouTube. So that number's down. It's probably because I haven't been live on YouTube for a couple of weeks.

It also says, it says here, streams current bitrate is, uh, 3,149 kilobits. Yeah, that's about right. It says it's higher than the recommended or something like that.

Yeah. And it's based on the setting that I have here for the stream output because I'm pushing a couple of streams. So I give it about 3.5K per stream.

Oh, you do? Okay. Yeah. Interesting. Yeah. Yeah. It's interesting to see kind of the in and out of viewers too, where they check it out, where they leave.

Yeah, so for me, watch time hours in the last 28 days is only 83 hours.

So I would have— or just this, this show, or—

yeah, for my channel, whole network. Yeah, for my channel, GNC and NMS. So that tells you that's, you know, a long ways to go. I gotta triple, quad, almost quadruple the amount of people watching in order to get the watch time hours up. Um, but I've been, you know, the experiment that I've been doing that I, on YouTube, that I, I'm just, I thought I was on to something, but then, so here's an example. Um, you probably can't see this, but I started using this art thinking that would help attract people to the show. The SpaceX one was super successful with about 300 views. But if I look at, if I go through and look at all the imagery art, we do better when I just don't use any Midjourney art for the, for the show. People are not reacting well to most of these Midjourney images that I'm putting in on YouTube for the, for the, for the image, a selected image that we do better when we just have our faces on the screen, which is curious. And I thought that the art would have caught someone's eye and caused them to click more than just seeing our faces and clicking. But that does not seem to be the case.

Well, there's no context there. That's, I think that's the, that's the challenge with those art pieces is that they, they don't clearly communicate to anyone unless they read the subtitles.

Right. So maybe there needs to be some text on top of those graphics then.

Maybe. Yeah. If you look at the shows on, on YouTube that do well, um, that's kind of what they do is that there's like a, shocking look on somebody's face and then, oh yeah, text, you know, across the—

yeah, the actual image itself. But most, you know, just— yeah, and it's a lot of, a lot of clickbait.

Yeah, there's a lot of kind of like teasing things that they— yeah, that the video may not really be entirely about.

But yeah, it's— I hate clickbait. Some shows use it on YouTube, use it so much it drives me crazy. It's like, oh, we have a We have a problem.

Well, you never know what you're going to get. You know, it's like, oh, the world's going to collapse, you know, or it's going to be, you know, it's in the world's ending in 24 hours.

I hate to say it, but that's what people get sucked in on. That's what they— maybe that's what we need to start doing is clickbait podcast episode.

Podcasts will die on exactly September 1st.

So, hey, that's what we'll do with the audience. So come up with a couple of clickbait titles and we'll do some tests with the new media show and we'll use some clickbait titles to see how people react.

So what is the, or what should the title of this episode be? You, you think, is it about the scary AI in podcasting?

Probably should be. Yeah, I think so. That'd probably be a good title. AI causing, you know, you have to— what would be the clickbait?

AI causing demonetization in podcasting.

Oh, but I have to get more— you have to get more divisive than that, Rob. You have to, um, you have to say AI creators. Yep, AI discriminates on Black podcast creators.

That's the clickbait.

That's it right there, right? But I just get— I just get— I just get icky thinking about writing a title like that, though.

Our average view duration just went up when we talked about that.

So, but I think that, I think I just feel icky thinking about doing a clickbait title like that, though. Yeah, I know.

Yeah. But it's true. Yeah. Well, people, people get attracted to things that decisive, provocative, right? That's true. You know, inspire them to think, oh, I want to learn more about that, right? So yeah, so there are— this is a mental game is what it is. And I do think it's coming more and more to podcasting. I think is this kind of concept of clickbait.

Let me read you a quote for something. I'm going to read you something that's going to really make you think here. And if this could have come out of my mouth— oh, it could have, it could have, but it didn't necessarily— didn't. And I think you're going to appreciate this. At this conference I went to, there was a recurring theme talking about original content, podcasts, videos, and live events. I mean, podcasts were said so many times I wanted to stand up and say, hey, By the way, all of you here, I'm a podcast host. You want to start a podcast? Come on over, right? I really wanted to do that.

What was the context of the comments that were related to podcasts?

Okay, so listen to this. This wraps it up in a net, and I copied this verbatim. There's a shift in perspective where more value is placed on human viewpoints through podcasts, videos, and live events. There's a shift in perspective where more value is placed on human viewpoints through podcasts, videos, and live events in the context of an AI content-generated world.

Who do you trust? The trust factor. I'm doing a whole show on that.

But, but I didn't say this. This came from a very respected individual who was he up on stage or she up on stage, a keynote. Yep. There's a shift in perspective where more value is placed on human viewpoints through podcasts, videos, and live events. So here's the thing we got to think about. There are companies now that used to put out 50 blog posts a week. They're up to 1,500 blog posts a week through AI-generated content. Wow. Okay. Google has said, we don't mind AI-generated content as long as it's good AI-generated content, edited AI content, validated AI content. So a company that's putting out 1,500 posts a month just trying to be a machine and get enough clicks and data views, their brand value is going to go down because no longer are they using humans. They're not, they doesn't have a human's viewpoint. It's just, gah, right? Right. They're thunking out 1,500 worthless post in just a numbers game, right? So when people are reading stuff now, they're not going to know what to trust. So podcasts, videos, and live events where you can see a real person moving their lips, talking about life and giving a real human perspective— I think we're on the cusp, Rob, of a new golden era of podcasting. Because of this and this alone. So it is a great time to be a podcaster. It is a great time to be a podcaster. I'm not saying that from a marketing standpoint and come host at Blueberry. I'm not saying that. I'm just saying this is, this is the time. Make hay while the sun shines. Yeah, because diametrical change is coming within the next 1 to 2 years.

So how do, how do you think content creators that are doing that are in a better position going forward?

Well, it's, it's simple. We're not putting an input in and thunking something out that we're presenting that has been regurgitated by a large language model. We are We are not affected by the large language model.

We are—

We're unique. We're positioning our voice and giving a perspective that is human.

So that's the differentiation, right?

Right.

We're human, not a machine. Yeah. That's how we're going to stand out in the future, Todd, is that we're now competing with robots.

No, we are.

We already are.

Yeah, I know we are. But original content is going to rule the day, right? Stuff that's been put on news sites, sports sites, all of that is going to be nary a human having a touch on it. It's going to give a robot's opinion of the trade that was made between the Yankees and the Detroit Tigers. I don't know sports, but you know what I mean. Yeah. Oh yeah. There's going to be very little human commentary involved. So it looks like you're getting texted, Rob.

I am. I thought I would post a, just a quick comment to, to my YouTube page. Oh, okay. Just to spark the algorithm a little bit.

Oh, that's what it takes to spark the logarithm is to Is to post on the YouTube page?

Just to show a little activity.

Well, you have a bot do that. They got bots for that now. They have a bot for that. Hey everybody, thank you here. I'm glad you're here.

Yeah, they do. So yeah, so you just, you know, that's us live, live on the web.

So you're kicking the, you're kicking the ilago. That's kind of makes, that makes me laugh. We did get a comment earlier from Sean that says Leo hates Adam Kareem, will never support V4V. And Beauty Bubble said, what is moving this, this needle downward? Uh, economy. And people are— I think there's a, a complete shift. I think that people are not watching as much regular TV.

Oh yeah, I think their mainstream media is doing this.

Yeah, I think that, uh, I, I— people are being discouraged. They don't understand this is hard work. Budgets have been cut. You can't monetize the network, you know. You're getting, you're getting ad loads every 7 minutes on popular shows. People are sick of being sold to every 6 and a half minutes for 30 seconds. I, you know, Yeah, right. And they're just gonna run themselves into the pavement. They keep doing that, the listeners are going to leave. Are they getting pretty good at hitting a 30-second advance button?

Yeah, or they just go to the next thing, right?

They just hit skip. Yeah, skip, skip. Yep, 30 seconds, fast forward 30 seconds past the ad. And be annoyed doing it. But when does that annoyance run out and people leave?

Yeah, I mean, I know on, on YouTube, but you know, this is a big factor of, um, average duration, right, of viewers, right? That goes into the logo, right? So, you know, given that factor is a big factor to YouTube as well, is the amount of content that's actually watched or listened to.

We're at 23 minutes and 5 seconds for average view duration on GNC right now. So, but again, not that many people are watching. So those that are watching or hanging with us, we thank you for staying with us all the way through here.

Yeah, but I've been getting on the replay of this on, on my channel, um, you know, a few of the episodes that I, I did through StreamYard have had, more than a couple hundred views already, but yeah, how much did they actually watch is the, is the bigger question. So you add the couple hundred that are on my channel, however many are on your channel. And then if you think about all the other platforms that we're pushing this out to, to whether it be Twitter or LinkedIn, yeah, we're not even looking at see how we're doing on Twitter. Twitter's actually done very well for us.

Or X, or X, right? Yeah.

Well, I'm going to, so when we're doing this show, um, on your platform, I don't believe we're pushing it out to LinkedIn yet.

Yes, we are. It's live on LinkedIn right now. Yep.

We are. Okay. Yep. Okay. Yeah. Cause I know when I was doing it on StreamYard, I was pushing it out to my own personal LinkedIn account.

Yeah. So, so as an example, today we've had, oh, it's horrible today. We've had very few viewers on Twitter. Maybe when it ends, we will get more, but usually, uh, I'll come back to Twitter a few days later and As an example, um, the show we did on July 30th, the tech show I did on July 31st, got 360 views, but it didn't say how long they listened.

I don't know. Right. So yeah, I'll occasionally go in there and I'll, I'll, I'll repost it, right? Yeah. And then add more context to the post, which always helps.

And it's been so far today, our has, ours has gotten 32 views, which is a little bit lower than normal. On Twitter. So, X, X, I'm not gonna— it's still Twitter to me. It can be X, X, whatever we want to call it. It's still Twitter. It's going to take a while to break a habit.

Yeah, I don't think it's being pushed to my Twitter account.

It's just being pushed to yours. No, because I have to authenticate this unless you give me your username password and I can't authenticate it. Right, right. Yeah, yeah, that's true.

That's the challenge. But I could go to your page.

Yeah, and you could retweet it, right? Yep. So I, I guess we're, you know, it's an interesting time to be a content creator at this point, be a podcaster. Pour on the coals, ladies and gentlemen. Don't, don't— what are you seeing with you? You know, what is everyone seeing with their stats? I want to hear from the listeners. What are you guys seeing? Are you seeing your numbers go up? Uh, is it flat? What are you doing? Right. I love, love to hear what you're, what you're experiencing. I see all these people that went out and bought the new RODECaster Duo, and, uh, that's a cool, that's a cool thing. I've had no need. I've got my original RODECaster sitting here that's rocking and rolling. I have no, no desire to, uh Dump more money into the system on another toy? I don't know. I've got stuff that costs a lot of money that, as a matter of fact, I should start putting some of it up for sale. I got a, I got a premix that I should sell.

I'm perfectly happy with my, my Rode Pro, my first, first version. Yeah, me too. There's no, there's no real significant reason. I know, I think Dave Jackson upgraded his system because he just wanted to make more space on his desk. That's funny. He was selling his— even he was selling his, his RØDECaster 2.

Oh, as well, because he wanted to get that Duo, which is a much smaller— oh, just the 2-channel. And then he's going to need at some point more additional channels, and he's going to be mad he sold it, right?

Um, yeah, he just wanted more room on his desk is what he told me.

Well, if you look If people could see my desk, yeah, that would, that would save me 4 inches of stuff on the desk. But in the scheme of things, laptop, laptop, monitor, TriCaster, monitor, monitor.

I've got a big table.

I've got actually my desk is actually a dining room table is what it is. Oh, that's funny. I have an L-shaped desk with a cabinet on one side and I have I have a Mac Mini boxes stacked up to bring the one laptop up higher that's doing the live stream. Hey, if you have been listening live today, where have you been? I see you guys sending us Satoshis by the minute, um, but I guess we're not loved anymore on the Value for Value.

I think we've been out of, out of the pocket for value for value. I now have an LB account.

Oh, send that to me and I'll put you in the split for the show. So is it Rob at GetLB?

What's your— do you remember?

I believe it's, uh, I believe it's Rob Greenlee. Well, send me, send me the info and I'll add you, uh, I'll add you to the Satoshi split.

Okay. Yeah. I mean, if you can, you know, Sam gave me already, I got like, uh, 1,300 Satoshis in my account.

Oh, that's about 30 cents. I know, believe me, I don't feel rich right now because of it, so that's all right. Well, this show in the last, uh, probably 3 months have done a little over a million Satoshis, 1.2 million Satoshis, which is equivalent to like a dollar or something. No, that's, that's about, uh, I think it's $100, but maybe like $100 and about $140, I think. Oh, okay. So, and it's, you know, it's like stuff is coming in right now, 28 sats at a time. People are listening to the show on the apps and they're streaming us sats by the minute. That's where we get most of our, our earnings as people are, are using the new podcast apps at podcastapps.com and streaming us sats by the minute. Anywhere between, like, there's someone was doing 94 sats a minute. Another person was doing, wow, 500 sats a minute. Wow, that's impressive. Mitch, thank you for your 500 sats a minute. That's, that's a, that's a baller boost. 28 sats a minute from Dave Jackson. Chad was doing 141 sats a minute. Brian was doing 47 sats a minute. I'm just looking back, looking back through here, people that are streaming this, listening to the show live and then sending sats by the minute. But yeah, 24, 27, that 500 one, that, that's, that was a buck 25 a minute. Oh, you got your picture of your RODECaster 2.

No, this is the Duo. Oh, the Duo.

Yeah, yeah, yeah.

Yeah. So I just wanted to throw those up. This is the new smaller, more compact version of the Procaster from RØDE. So I think I saw this in Vegas is where I saw this. It's got pretty cool software in it, but it is a, it does have 4 sliders there. So I think it does have the ability to have 4 channels, Todd.

So, uh, what do you call it? Uh, The, um, Adam Curry's in love with that, that, uh, small one because he's, he's made it in his travel kit.

Oh yeah, I mean, it is pretty small.

Yeah, you're right, it's nice. What happened to you here?

Hang on. I don't know what, what did happen.

All right, there we go. There you go. All right, we're already over. Um, I'm Todd at blueberry.com, @geeknews on Twitter.

I'm on Twitter as well, @RobGreenlee, and I can be reached at rob.greenlee@gmail.com. So that's a great way to reach me and check out my new Trust Factor show that I'm doing now. So, and then I'm also involved in a project with Robert Scoble out of the Silicon Valley area. We're working on a project together called Unaligned AI. So I'm I'm doing some interviews with some AI startup founders and we're gonna document the, the development of this, um, AI stuff, especially around media. And then also, um, you know, um, kind of talk about these issues of how these AI platforms are not really aligned very well with humans quite yet.

Yeah. You're gonna need to do about 50 interviews a week to keep up. I don't disagree with you, Todd. And it's changing so fast you can't even breathe.

So yeah, I was just in New York City yesterday, starting doing, doing some conversations with some AI startups in New York. So yeah, so there's going to be more content coming out about that from Robert and I.

Right. Very, very cool. You get to handle the East Coast, he's handling the West Coast.

Is that how you're shaking it up? Splitting it up, right? Well, so I'm, I'm focused more on media, and he's, he's more focused on kind of like the LLMs and, and what's happening in the Silicon Valley.

I talked to a startup in Columbus this morning that it's not just the coast, the money is far and wide.

It's true, it's true, but there is a heavier concentration, that's probably media in New York compared to Silicon Valley, so Right.

Have fun. All right. Oh, I'm glad you want to do it. I don't want to keep up with it.

You don't want to keep up with it?

It sounds like you're all in on it, Todd, actually. Well, I am, but I'm— I can't let it— I have a company to run and stuff to do. I can't be spending 8 hours— you could spend 8 hours a day on this and not keep up. So I'm— well, I'm doing the—

I just get newsletters. Is it a worthy effort to try and, uh, keep up with what's happening with AI? Is it, you know, is it a smart, you know, career and business move to be knowledgeable about this stuff?

If there's the other question, right?

If you're not, you're done, right? So the answer to that question is yes.

So, so if you're not, and if you're, if you're not, if you're listening and you haven't touched anything yet, uh, time to get busy. Time to get, uh, time to get chopping here. Sean says a RODECaster Pro 2 rules, should upgrade from the first model. Yeah. So, all right, everyone, thanks for being here. Thanks for watching us or listening to us. And if you've been live, make sure you've made comments in whatever channel that you have listened on. And again, drop us an email if you have comments. We'll be back with you next Wednesday, same channels. And I'll say that plural. And now we'll take this live pod— live show and turn it into a— oh, I heard that slip. And we'll turn it into a podcast. We'll turn this live show into a podcast. All right, right, all right, all right, everyone, thanks for being here. We'll see you next time. Bye-bye.