Creative Flux

Bilal Tahir (00:00)
hello, we're back.

Pierson Marks (00:01)
Well,

we're back and guess what?

Today is episode 20. That means we did it.

Bilal Tahir (00:05)
20, we did it. What was

the stats you told me were less than 1 % or we're in the top 1 % now or something? Wow.

Pierson Marks (00:13)
tough 1 % of podcasters. So ⁓

99 % of podcasters never get to episode 20, and we made it. It's pretty sick. No, with one, only one week we skipped, ⁓ and that was just because it was so busy and it was just like impossible and the quality would have been poor if we did it at ⁓ conference. I think it's good. ⁓

Bilal Tahir (00:21)
We did it, we did it.

Yeah, I think that's pretty good, you know?

Yeah, cause there's

something about doing something consistently is actually surprisingly hard. You know, I mean, cause so much stuff happens and stuff, you know, and you know, you always, feel like I really struggled with that. Like I have the initial phase, I'm excited about something, a project and dig, and then you hit the hurdle where it gets a little more gnarly and then you, other stuff comes up and then, you know, you kind of leave it. And I feel like the difference between someone, you know, who does amazing stuff versus, you know, not is like they push through that.

Pierson Marks (00:39)
It is.

Bilal Tahir (01:03)
plateau and then just keep grinding you know so

Pierson Marks (01:04)
For for sure.

I completely agree. It's always really interesting, especially where we live, know, in San Francisco and the Twitterverse. It's like, it's a ADHD overload, you know? It's like somebody's doing this, somebody's doing that. Like, there's a $61 million, like...

seed round, a $20 million pre-seed, they're doing this and this and this and everybody feels like so spread out. And it's just like, this is awesome, it's super fast, but it's very spiky. And I think that building a successful business over the long term, in any domain, it comes with week after week, consistent grind, focus on like...

Bilal Tahir (01:33)
Hmm.

Pierson Marks (01:42)
one thing and just go really, really deep in one thing, which can get distracting when you're in a world where everybody's working on all these cool things all the time. Like, man, I wanna go do this, and I wanna go do that. yeah, so.

Bilal Tahir (01:45)
Right.

Yeah, maybe

in the future once AI speeds everything up, we can have our cake and eat it too. Just get through a project and then move on to something else, know? Just streamline. Yeah.

Pierson Marks (02:01)
Totally, I completely agree. But well, for everybody

that's listening today, this is Creative Flux. I'm Pearson and this is Bilal. And we talk about generative media, video, audio, and all of the above. Anything that does with media and AI in general. Sometimes we get to stocks and that's unlikely, but today. ⁓

Bilal Tahir (02:11)
I'm Bilal, nice to meet you guys.

Well, it's

been a brutal market last week. I feel like people are like right now, is the bubble popped? Is everything over? I don't know. I feel like we got some legs. We'll see.

Pierson Marks (02:34)
Right.

Totally. Well, I think so. I think it would be cool. So there's a few things that we wanted to cover. Episode 20 is pretty big. Last week we ended on LTX. So LTX for anybody that doesn't know, it's a video model by who's the, do you know the foundation model provider? I forget the name. I'm out of ahead.

Bilal Tahir (02:40)
Oof.

that's a good question. I

can't think of them off the top of my head, but they are really cool because they are like putting out these open source models that are just not getting in as much hype as ByteDance or Sora or all these other people. they started, I remember we covered them initially with LTX, their first model. it was, what was it? ⁓ They had the first, like, I think once...

Pierson Marks (03:12)
Right.

Bilal Tahir (03:16)
basically one cent or two to four cent, five second video, which is insane. I think VO3 at the time was like 75 cents. So insane, right? And I mean, obviously the quality wasn't...

Pierson Marks (03:22)
It's huge.

Bilal Tahir (03:30)
That's good, but you could get something super fast, almost at the speed of an image. And I remember we talked through this workflow I had where I literally would generate these images. And then once I like the image, I would generate the LTX video just as fast. And then if I like the LTX video, then I would go to Sora or V03. So it's almost like this cheap way of making sure I like a scene. But now they've...

released version two of their model. It's called LTX V2 and it's insane. It's now up to 20 second output duration, which ⁓ is unheard of. I mean, I think the max I've seen is 10 seconds for, know, maybe eight seconds for Vio 3.1, 10 seconds for Sora. So 20 seconds is insane, which I think it might be the GB 200 or something, you know, could be.

Pierson Marks (04:06)
All right.

Bilal Tahir (04:13)
some architecture, well actually thinking they're Chinese labs, so actually they wouldn't have access to that. So it's kind crazy they did that. ⁓ And it's also native audio, which is only Vio 3.1 and so on. So you can actually have native audio in there and it sounds pretty good. So I've been playing now, it's been super cool ⁓ what you can make with that.

Pierson Marks (04:20)
Right, that's it. Yeah, that's typical.

Totally. Yeah, so this was announced.

Right,

no, it is super cool, because this was announced on the 23rd of October, so like a week before Halloween. And we'll get into this, because I think there were some things that were generated for Halloween, like the Frankenstein video that we saw.

Bilal Tahir (04:45)
that was amazing. Yeah.

Pierson Marks (04:47)
That one was cool.

but yeah, it's like LTX is one of those, I just looked it up. It's the company or the organizations called Light Tricks, makes sense, LTX. And they have audio, video together at 50 frames per second, generating 4K. So I mean, at native 4K resolution, 50 frames per second with synchronized audio, pretty cool, it's cheap. And is it on foul? Do you know?

Bilal Tahir (04:55)
Yeah, live tracks.

Sorry, was it? It is, yeah, yeah, it's on file and they have, I think they have a multiple one. They have a pro version, which is six cents per second for a 1080p. But the other thing resolution, 1080p, insane. mean, usually people, they do, we go from 480p to 720 and 1080p is like.

Pierson Marks (05:12)
Is it on foul? Do you know? Okay.

Bilal Tahir (05:30)
really expensive here they started six cents for the pro version 1080p and then go up all the way up to 4k 2160p 24 cents per second which is you know i mean basically very reasonable and then for the fast version is you divide that by basically take 25 % off so four cents a second for 1080p but 4k video

Pierson Marks (05:35)
Right.

It's cool.

Bilal Tahir (05:50)
4k video 16 cents per second. you know basically under a dollar you can get a five second 4k video. That's insane. You know I think this this lab needs more love. You know people just don't know about it. It's crazy.

Pierson Marks (05:57)
Totally.

Yeah.

It's interesting, we've had this discussion before, what dimension do these labs differentiate themselves on? One could be price, one could be quality, one could be style. think the style differentiation's gonna be the key because I think you'll see price go down to zero. Slowly you'll see longer and longer generations for cheaper and all models will have to compete on price. And then all models will also get really good.

you can safely assume, I think, not safely, but like you could assume that the models get really cheap and really good. And so what is the differentiator between a model? It's like the understanding, the native understanding of user intent. And you come to a model because you go to, let's say, mid-journeys model, because you know all of mid-journeys models have that sort of cinematic, realistic aesthetic. I think it's gonna be the same thing with video models where you're gonna go to LTX.

Bilal Tahir (06:34)
Mm-hmm.

Right.

Yeah.

Pierson Marks (06:56)
for a certain type of style, Sora for a certain type of style, and don't have to be explicit in your prompting because you just know that that model's gonna fill in those gaps. And I think that's like where you're gonna see differentiation because quality and price are gonna just be commoditized, I think.

Bilal Tahir (06:59)
Right.

I agree 100%. I

mean, mid journey is a great example. it's been, I mean, you know, they charge premium fee, but people pay for it because that's of it. And it's funny because I've resisted mid journey for the longest time. And I'm like actually really close to subscribing because I feel like the images you get are just insane. And if anyone who's played around with the UI, have, they make it much easier for you to discover styles as well. Cause you know, it's one thing a model can do stuff, but you know, you kind of have to

Pierson Marks (07:17)
Mm-hmm. All right.

Really.

Mm-hmm.

Bilal Tahir (07:39)
have an idea of what's possible. And with Mid Journey, you have these, I think it's called SRFs, SRF style references where they literally have these amazing palettes and you can just sort through them. You can find the thing you're looking for and then generate a consistent ⁓ story from that. And so if you, it's not a coincidence, then if you look at all these, we're talking about the Frankenstein video or all these.

Pierson Marks (07:42)
That's rough. Yeah.

Totally.

Bilal Tahir (08:01)
pretty sure all the images were generated by Mid Journey. And I know Mid Journey has a video model that's pretty decent as well, at least for animation. But I think most people, what they do is they generate the images in Mid Journey and then go to Vio3 or Sora. And that's the combo that's been working really well.

Pierson Marks (08:03)
Right. Totally.

Right.

Right.

Well, I mean, and also this is like a good segue, I think, into this concept of like workflows, because we've talked about this in the past. Like the common workflow is you go and you generate an image. You use that image as the first frame. You may generate also the last frame image. So you have like the first frame, the last frame, and then you bring those into something like Vio 3.1 or any video model that accepts a first and last frame. You let the video model kind of fill in the gaps. Like, Hey, this is the first frame.

fill in the gap to get to the second frame. And it's a really powerful workflow. And so this concept of workflows, mean...

Bilal Tahir (08:46)
Hmm.

Pierson Marks (08:51)
Why I wanted to bring this up is because you're to see model capabilities improve. And I think that model capability is going to get commoditized except for some few dimensions. But where the value I think is in these workflows in like piecing together different capabilities from image models, from video models, from audio models, from also all these other things like non AI sort of capabilities. Like everybody's so focused on AI and obviously

Bilal Tahir (09:04)
Thank you.

Pierson Marks (09:18)
this is an AI generative media talk, but something that happened last week was FAL released a auto captioning endpoint. And so they said, hey, if you want to make TikTok style captions on your video, call this endpoint with any video and we'll add the captions in the bottom. And so you're like, that's super cool. How do they do that? Well, obviously they're not, and it's super cheap. It's like three cents per.

Bilal Tahir (09:25)
Hmm.

Pierson Marks (09:41)
minute I think of audio. So it's really cheap. You add captions to your video. How are they doing it? Well, they're obviously not taking your video and using Sora 3 or Sora or VO3 and putting captions on that. They're taking your video, extracting the audio, going to transcribe that audio, and then use some deterministic sort of presets and captions and putting them on.

Bilal Tahir (09:43)
That's right. Thank you.

Right.

Right.

Right. Probably FFmpeg.

Yeah. Yeah. Right.

Pierson Marks (10:04)
Exactly. And so I think you're going to see

more of that where it's like, it's a hybrid between the AI capabilities and kind of stuff that we already are capable of doing today, but maybe we could do it better when you pair it with AI, like generating SVGs, animated SVGs. Like, can you, for example, it's something we talked about in the past, like you could use an LLM to generate the code to write the SVG that has the animation. And so you don't actually have to generate the video.

Bilal Tahir (10:22)
Hmm.

Hmm.

Pierson Marks (10:33)
you could just generate

Bilal Tahir (10:33)
Right.

Pierson Marks (10:34)
the code that represents the video and have motion graphics. And so I think that you're gonna see like more of these piecemeal pairings. And I think you'll see a marketplace or an app store type thing on FAL on these places where maybe you're not a model provider, but you're a workflow provider where you piece together these workflows, you build them and you sell them as an endpoint hosted on FAL and these places.

Bilal Tahir (10:37)
Right.

100%.

mean, the layering is just like, gives you that polish that distinguishes it so much. Like you can do stuff like add your own background music or layer effects, sound effects and stuff, and just go to different models that will generate like a lightning, thunder strike. I mean, rather than zero shorting that, you focus on getting the scene right, then.

Pierson Marks (11:11)
Mm-hmm.

Bilal Tahir (11:15)
put it through another model, get the sound effect. I've also seen other, like one use case that I thought was pretty cool was I'll often generate 480p video, because it's cheap and fast. And then rather than, I could just go for the 1080p and just pay more. But I've discovered I can just do the 480p, then go to a video upscaler and there are end points for video upscaler, and then upscale that 4x. And it actually can sometimes come out cheaper and better. mean, ⁓ so these kind of like, you know workflows as you're

Pierson Marks (11:37)
Mm.

Really?

Bilal Tahir (11:43)
we're saying, stitching together these different models, or even not even models, algorithms, FFMPEG, endpoints, et cetera, are so powerful. And we're still figuring out what's the best combo of all these.

Pierson Marks (11:51)
Totally.

For

sure. I think there's a lot of leverage right now to just figure out what exists, what has existed for last 10 years. Like FFmpeg. FFmpeg's a horrible library, honestly. It's complicated, it's hard to run, you have to use these weird terminal commands to run it, and we face challenges together. But just to even know what's possible with that, and then thinking about how do you pair that with AI?

Bilal Tahir (12:13)
Yeah.

Pierson Marks (12:19)
It's a lot of especially if you're a marketer or you're a creative you like people aren't doing this stuff And there there's a lot of room for opportunity. There's this other thing like when I was doing our launch we had our launch week last week and Over the course of the week the the thumbnails for YouTube Hopefully it got a little bit better. I was just playing around with it It got faster for me to create with the last one that I posted

Bilal Tahir (12:42)
Now I'm loving the

thumbnails. love the one with the, I don't know what, there was something at the top and bottom like cool. Like flag thing, kind of haze, it was kind of cool. Not a flag, but it was like a color. It was kind of a purplish, kind of a haze-y color. It just looked pretty cool.

Pierson Marks (12:49)
I don't know.

flag thing. I'm not sure. Yeah, I'm not sure. I'll have to...

so yeah, was just going for like this retro pixel deep sea sort of vibe. ⁓ But yeah, like the one of the last videos that I created, I took a photo of me, like just literally, I was recording these tutorials, took a photo.

Bilal Tahir (13:04)
maybe that was a different one.

Pierson Marks (13:16)
snapshot of like me in my recording. And then I went to Photoshop and I removed my background, but now that has like an AI removal. And then I just pop that on top of a like a nano banana generated image background. And then I had like text that was placed on the bottom through that Satori MCP. And so it was just like, it was like such a

Bilal Tahir (13:23)
Hmm.

Pierson Marks (13:35)
easy workflow for me to envision this automating. Like, hey, you just have a snapshot. You can take a snapshot of me right here.

And then you're like, here's a YouTube thumbnail. And most of it's not AI. Most of it's just like, generate a background, remove my background, put me on top of this, and then put some text on top. And don't generate the text. Use actual React components to put the text in the right spot, the right size. And don't generate that. Just render it correctly.

Bilal Tahir (13:43)
Right.

Right.

Mm-hmm.

Right.

Right, yeah.

No, there's so much alpha. mean, I bet like the new Photoshop or Canva, like the UX is being built right now and we're gonna look back a few, I'm like, of course that totally makes sense. was, it'd be like one of those slam dunk obvious things. Like, yeah, that's obviously the best UX, but right now, you know, we're still figuring out what it is, you know, where, cause I feel like we shouldn't be, it's still so much of a pain, like making these decisions, like putting in an endpoint, like me actually, okay, I now have to configure a file or, I have to go to Photoshop just for this one removal thing.

Pierson Marks (14:21)
Mm-hmm.

Bilal Tahir (14:33)
It should just be there for us, like, you know, in one UX. a lot of alpha there, for sure.

Pierson Marks (14:34)
Totally. Totally. Totally.

100%. And so one thing I wanted to touch on just because

we didn't get to last week and we mentioned it earlier was there was two pretty viral videos that happened last week. One before Halloween, the Frankenstein video. I wanted to talk about that real quick for anybody that hasn't listened. And then also this Nike video that came out too. I don't know if you saw this one, the runner. Yeah, yeah, yeah, yeah, yeah.

Bilal Tahir (14:55)
Right. Yeah.

yeah. I think

it was, remind me who was the character in that one.

Pierson Marks (15:07)
It was the guy with the chainsaw mask and... ⁓

Bilal Tahir (15:11)
Oh, yes. Oh, shit. See. So kind of the Frankenstein video, which is basically Frankenstein kind of like talking about how people are scared of him is he's just talking like, you know, he's doing therapy and he's lonely. The chainsaw video was awesome, too, because it's like basically the Texas Chainsaw Massacre movie on touch. But it's like the first half is just him trying out, you know, working out and just wearing Nike sneakers. And you think is just a regular pump, you know, sports motivational video. And then he starts running after this girl with the chainsaw. Is this the

Pierson Marks (15:34)
Right.

Bilal Tahir (15:39)
the most amazing twist but for me what really stood out for both these videos was and it's kind of a kind of a slap in the face from someone who I think I'm creative and stuff you know we always had that excuse you know I can be Steven Spielberg if you give me a billion dollars I can make Jaws now you have no excuse I mean the storytelling is what separates like these people you know people like who make these videos and the ones who don't and I look at this I'm like

Pierson Marks (15:56)
It's true.

Mm-hmm.

Bilal Tahir (16:04)
I wanna come up with this. It just, it gets me little jealous. I'm like, shit, I wanna create something like this. This would be awesome, right?

Pierson Marks (16:05)
Mm-hmm.

Right.

100%. No,

totally. And it's really cool because also what you're going to see is that you can iterate more quickly. So you can iterate more quickly with the tools and you can also collaborate with people because I don't, I'm not in Hollywood, never been in Hollywood. So don't, I can't speak, but I'm assuming that I think oftentimes tech is much further ahead than any other industry in terms of both the technology and also in the workflows. like the way that we manage code repositories and we like, and we collaborate how I can code.

you can code on the same thing, and then we can both merge our code together, right? In a lot of other industries, that still doesn't happen. You share a document back and forth, and you have to, it's crazy. And I think that you're gonna start to see tech, those processes, get built into movie making, filmmaking, script writing.

Bilal Tahir (16:45)
Right.

Right.

Oh, 100%. I

think I've always said, it's what happens when content becomes as cheap as software. I feel like these paradigms like get up merging, forking, they become, I think, much more easier and accessible. And like you said, right now, the majority of this tech is accessible to tech nerds like me and you. But people who are actually artists and stuff, who are maybe not technical, but they understand story making and stuff, I feel like they're the ones who are actually going to be able to take, once they get their hands

on this and it's easy enough for them to use that's where we're going to see this amazing explosion of wow you know that's like like pj and stuff right so

Pierson Marks (17:28)
Mm-hmm.

Totally, totally,

totally. It'll just be, yeah, it'll be interesting. Like the collaboration side on things is going to be very cool. I can imagine like very much where, like what we just saying, yeah, there is opportunity.

Bilal Tahir (17:40)
Yeah

Right. And there's opportunity. Yeah. Actually, on that note,

I don't know if you saw, someone just made a Billboard Top 100 song on Suno and she got signed by a label. So the first instance of a human being being signed by a label based on AI music. I forget her name, but yeah, I think she was a musician she was making, but this is like, she made basically AI-generated music using Suno and she got signed. So.

Pierson Marks (17:55)
Really?

Let's this up.

Bilal Tahir (18:08)
This is like, feel like, you know, there's this right now, there's this amazing opportunity, you know, for you guys to like you, if you want to be signed by Nike, you make the Nike commercial, whatever, and you know, get the deal. You know, if you want to be the next Taylor Swift, you know, start putting out the songs, you know, start building your image.

Pierson Marks (18:24)
Totally.

Totally, so this is a Suno created AI artist named Zanya Monet and signed a multi-million dollar record deal with Hallwood Media. So Zanya Monet is the artist name and the deal is estimated to be at three million dollars and the creator behind the AI artist is Talisha Jones. So she uses Suno for composition and vocals but she writes lyrics herself because that's fair, I think AI is still not that great.

Bilal Tahir (18:28)
cough

Yeah.

Pierson Marks (18:52)
at lyrics. Even

Bilal Tahir (18:52)
Right. Yeah.

Pierson Marks (18:53)
though if you listen to any like hop or pop song today it's like the lyrics suck.

Bilal Tahir (18:57)
yeah, it's so literal. I

like going to them, like make me a song about, you know, like walking alone in a park. And it's the lyrics would be, I'm walking alone in a park. I'm like, be a little more metaphorical, man. I mean, it's like, I don't literally want, you know, I'm going for the vibes. And then the lyrics will be, my vibe is walking alone in a park. I'm like, that's not what I meant.

Pierson Marks (19:14)
Yeah.

And then it's

gonna give you like Green Day's like Boulevard of Broken Dreams like, I walk a lonely road. It's just gonna copy it. So there's my two seconds of singing that's never gonna be heard again. ⁓

Bilal Tahir (19:24)
There you go. The only one I love.

Yeah.

Yeah, well, the other

one on that, it's funny, you you were saying that the other one that's been popping off on social media is like, they'll take like Justin, like a Jay-Z song and make like a jazz version of it. And it actually sounds amazing. I don't know if you've seen these clips. And I mean, I feel like there could a lot of these old artists could have a resurgence where you can remix their songs, you know, maybe a Tupac, you know, singing like, you know, like Frank Sinatra and stuff or, or, you know, what if, you know, Tupac and ⁓ Elvis were in the same era? Like, what if they had done a duet?

Pierson Marks (19:43)
yeah.

totally.

Bilal Tahir (20:02)
You can recreate these reality, like what if scenarios. And I think there will be a huge market for this, for sure.

Pierson Marks (20:10)
I

think it's super interesting. I think that there'll be like two dimensions to this. One, very similar to how Suno, you can use your likeness and you can show your likeness off to the world. Like Mark Cuban, you can just use Mark Cuban in your Suno videos and he's given that explicit approval that you could do. I think that's definitely gonna happen. There's gonna be a marketplace where you can, if you're an artist, mainly a vocalist, I would think, because your likeness is...

Bilal Tahir (20:13)
Yeah

Pierson Marks (20:36)
very much your voice and your face and your personality. like you can like if you're Chris Martin from Coldplay, you know, and you want to give your voice out to the world to make Coldplay as songs with your voice. Maybe you allow that to be like happen and you get a royalty every time it's been used or has been played. Maybe this is what Spotify does. I know we've talked about Spotify. I'm very curious to see how they enter the A.I. I think, because they're entered. They're going to enter into the generation space.

Bilal Tahir (20:39)
Yeah.

They have to,

mean they have to margin.

But I think it'll be very interesting

Pierson Marks (21:06)
So.

Bilal Tahir (21:07)
because I think you're gonna see two camps. You're gonna see people, you know, like the Cubans of the world or even like the Grimes, you know, who are more tech forward and they say, hey, here's my voice, use it. As long as I get a look, why do I care? And then there'll be people, you know, artists, actually, I think a huge number of artists, least legacy artists who are like, no, it's my voice, it's art. Nobody else gets to sing like me, you know, I'm Madonna. And I get that point too, because if you're Madonna, you know, if there's a million Madonna songs, you know, it kind of takes away.

Pierson Marks (21:15)
Mm-hmm. Right.

Sure.

Totally.

Bilal Tahir (21:33)
the value of what is a Madonna song, right? So I get, cause art in a way is scarcity is always been part of art. And so it'll be a very interesting to see, cause I imagine if you're a new artist, you have nothing to lose. You want to, you know, you're trying to pay a rent, whatever, right? But then if you're an established artist, I mean, you already have your millions. And so I feel like the calculus will be, if you're an old star, you probably will be a little resistant. And if you're new, you're like, no, I, you know, I want to jump in. So it'll be interesting.

Pierson Marks (21:40)
Mm-hmm.

work.

Totally.

It's interesting.

You ever know this company called Splice? Have you heard of them? So Splice is, they've been around for a while. They are kind of, they're the largest royalty free sound company out there. They work with like...

Bilal Tahir (22:05)
Uno.

Hmm.

Pierson Marks (22:16)
every big musician you just splice. If you're an artist or you're producer, you're on splice day in, day out, you're finding songs. And it's a marketplace where producers, like ghost producers essentially, will create samples, like vocal samples, 15 second tracks, vocal or instrumental things, and they'll put on splice. you essentially, as a user, you pay a splice subscription, like 20 bucks a month, and you get access to these credits. And then

Bilal Tahir (22:30)
Mm-hmm.

Hmm.

Pierson Marks (22:44)
you can download those those songs and use them commercially in your in your music. And so if you listen to house music ever and you get into the weeds of house music, you'll hear the same splice sample in multiple songs. And it's very funny because like they sound like the songs are completely different, but it's the same vocal track and it's either pitched up or down or like sometimes the exact same. the song is completely different. And I could see that also being a case where if you're just a good

Bilal Tahir (22:49)
Huh.

in.

Right.

Pierson Marks (23:13)
vocalist or something, you can kind of just let people, like you just clone your voice, clone your singing, and let people generate the lyrics with your voice, and it'd be another source of income. So I wonder if they'll give me a do it.

Bilal Tahir (23:24)
Yeah. No, it is very interesting. It's funny because I

remember there's actually, by the way, a really good if you've seen there's a YouTuber called Cleo Abrams. You know, she does amazing interviews about technology. So she did this video about how the copyright on music works. And she actually took a in the video she talks about. There's a Doja Cat song called, you know, let the town.

Pierson Marks (23:32)
clear.

Bilal Tahir (23:44)
that the town paint the town red, paint the town red. I think it's a pretty famous song. And she actually does, she's like, there's a sample in there, which is walk on by and walk on by is like a fame, like an old beat from like 80s or whatever. And so she's like, okay, so they have to calculate, okay, it's like a three second vocal, this is a song. And then they do some math. There's a math mathematical formula, how much the state of that artist will get because uses that sample. So it's funny, you know,

Pierson Marks (23:54)
Right, right, right, right, right.

Mmm.

Bilal Tahir (24:10)
And I guess we will just make this even more of a thing, right? It get super granular, I guess. So very cool.

Pierson Marks (24:17)
Great. That

is interesting. Well, so yeah, that's music industry, except for shaking. ⁓ So well, we talked about video, talked about music today. There was a few things. we want? There's just two ways we could take this. We can either talk about stable coins and Stripe, or we could talk about these new image models that are coming out. Maybe image models? Or what do you want to?

Bilal Tahir (24:25)
Yeah, but

Cool.

well, we've been on general media. Let's let's do stable coins for a bit. So you and I, went to was a Stripe Builders event, which was this past Tuesday. It's pretty cool. Went to the Stripe office, super sick office, like one of the nicer offices I've seen, you know. And they had so Stripe has. started to accept stable coins as a payment. It's very interesting because

Pierson Marks (24:44)
stay there. Okay.

Bilal Tahir (25:04)
I feel like it's still the Wild West. You should figure out stuff like, okay, if you have a wallet address and you instantly settle, how do you collect sales tax and stuff like that? And so I was asking the guy, like, okay, how do you do sales tax? And he's like, it's my shipping address. I'm like, but what if it's online? Right? So I feel like there's still as an industry, we have to figure out what it is. ⁓

you know, will mean when you can just have an anonymous wallet address and you can buy something. stuff like blocking, so appear an IP address or, you know, your credit card, like that goes out the window. It's like a whole new paradigm and which will probably lead to.

Pierson Marks (25:34)
Mm.

Bilal Tahir (25:35)
issues, you know, probably some issues, we'll figure those issues out. But one good thing about USVC settlement or crypto settlement is fraud isn't a thing because you basically pay, it's like paying cash, you know, I gave you 20 bucks, so there's no concept of fraud. Now I can't dispute that. And it's interesting because in Stripe,

when you can kind of dispute that because you get a payout, it takes a couple of days to get the payout. So Stripe technically does hold the funds, even though you've settled it. So they may kind of act like an escrow there, but for the most part, I stable coin payments are supposed to be irreversible. I mean, that's like the whole paradigm. And so we might have a bifurcation where anything less than a certain amount, people don't care, but then over that, they do care and they want old school escrow.

Pierson Marks (26:19)
So let's take a step back real quick. So there's two questions I want to get into. One, like why stable coins? Why is Stripe caring about stable coins if they're doing the money that they're doing today and the infrastructure? What's their incentive? Two, it's also stable coins the hype? like,

Bilal Tahir (26:31)
Yeah.

Pierson Marks (26:37)
Yeah, so like what are stable coins? So think the first thing that people, if you're listening to this, I mean, think most people probably know what stable coins are. It's a dollar or physical currency pegged coin that is pegged to like actual currency. So the USDC is a stable coin. There's like Stripe has their own stable coins, like Bridge. Is Bridge a stable coin or is that the network?

Bilal Tahir (26:41)
yeah.

Yep.

They

do have their own strip of coin. I think you're right. I know they acquired the company for that. So, but I don't know who uses it right now. The biggest one is USDC, which is Coinbase kind of manages that. And the biggest is actually Tether, which is USDT. And that's a very interesting company with a very interesting history that's gone over, but most people kind of trust it. It's too big to fail kind of at this point, even though they were kind of sketchy at the beginning.

Pierson Marks (27:08)
Right. Yeah. And so. Right.

Mm-hmm. Totally. All right.

I mean, you know, when people say

that about cryptocurrency, I was like...

It's not too big to fail. It can absolutely fail. Like look at the banking system. Now we talk about banking systems, we have trillions of dollars and movement and like that, like maybe that's too big to fail. But when you talk about cryptocurrency and the cryptocurrency industry is like still super, super small. Like you could wipe out 600 million or $600 billion. Like I think the whole crypto industry is like around, it's less than a trillion today. And so like you could like wipe that out. Like yes, well that's suck. It's a lot of money. It could, but like for just tether to go under.

Bilal Tahir (27:33)
Right.

Mm-hmm.

Pierson Marks (27:57)
Hopefully it never happens. But the reason why stable coins are all the hype now is because of the Genius Act that passed in Congress this year. And the Genius Act essentially were regulatory clarity around stable coins, not cryptocurrencies, not all cryptocurrencies, but stable coins specifically requiring issuers of stable coins to have and prove that they actually have dollar backed.

Bilal Tahir (28:03)
Hmm.

Pierson Marks (28:19)
funds so that if you issue one dollar of USDC you have to prove that you have a dollar of know treasuries like bonds ⁓

Bilal Tahir (28:28)
Right.

Pierson Marks (28:29)
or cash on hand and you had to prove that and that was kind of the way it's like, are you just making up all this money that's just pegged to the dollar and that's why Tether I think originally was like really, like hey, how can you guarantee if there's a run in the bank, a bank run kind of like, where's this money coming from if everybody wanted to withdraw it and how's it pegged?

Bilal Tahir (28:41)
You ⁓

Yay.

Yeah,

mean USTC is much more, I mean, I think transparent. Tether's thing was, I think they did fracture reserve. So it wasn't one-to-one backed for a while. And then it got lucky because there was no crash or no one had a run on them. And by the time people are asking questions, they had made so much money that they were one-to-one backed. So I think they're fine now. They're pretty solid now, but still, the transparency is not there, which is why I prefer USTC.

But what I think for me, the biggest thing about stable coins is it basically lets the world dollarize. Like it's so hard, everyone wants dollars. From Pakistan, everyone wants dollars there. For all the, we talk about Jerome Power printing my trillions of dollars, the Fed printing trillions and it's being unsustainable. The US dollar is still the most coveted asset currency in the world.

Pierson Marks (29:22)
Right.

Bilal Tahir (29:38)
and most people would love to have access to it but they can't because there's only a limited number of channels you go through and you have to pay fees and stuff what stable coins allow is like for anyone to have usdci i mean anyone from zimbabwe to you know cuba or whatever they you know nicaragua they can all have argentina they can all get usdc and basically have access to dollars and i think that is

what is the most powerful thing about stable coins? That access aspect. The other aspect, obviously, is the payment one, which is Salmon, which is cool. I think it's great. I can pay you a million dollars and the fee will be a cent. And that's the other thing. So the middleman is cut out. It's fast, it's instant. It doesn't take two days as it does with any, forget the name of the bank, iBank system, whatever.

Pierson Marks (30:01)
Right.

Right. Well, think about this. mean, the interesting thing is like this is advantageous for most people, except for if you're like a competing currency. But for the United States, it was like it's very important for them to be aligned here. And I think that's why the genius act passed, because one, you need. So if you're a U.S., if you're Coinbase, if you're Stripe, if you're an issuer of these stable coins pegged to the U.S. most of these issuers are going to be American companies. And so you have to follow American laws. So the American law says you have to

Bilal Tahir (30:25)
Right.

Yes. ⁓

Pierson Marks (30:53)
have one dollar for every dollar of cryptocurrency. so if Coinbase is like, hey, we're going to print or mint a billion USDC and allow you to buy these from us, they're going to have to go buy a million million dollars of treasuries. You're to be supporting, you know, the treasury markets and then you're going to be offloading those to other countries. And so it's like it's the easy way just to get our currencies into the hands of other countries.

And it's pretty bad. Like honestly, like though if you're in another country, like the power of the purse and like that's super, like when you lose the ability to print your money as a country, it's very like, that's one of the most powerful things you have is currency control as a government. so.

I see that if this becomes very powerful, non allies or like neutral parties that like are wary, like if the US becomes the international currency for consumers, because it's the reserve currency, but it's not like the transactional currency in other countries, oftentimes, because it's hard to get. And not everyone's gonna accept it. Like, oh, what do do with these dollars kind of, because you can't really get them. It takes away the power of a lot of smaller country governments.

Bilal Tahir (31:52)
Right.

Mm-hmm.

It does. mean, sometimes they prefer that because they've been so mismanaged, like, you know, like Argentina or whatever. They would rather have that and not worry about some other government coming and just driving inflation or trillion percent or whatever, and just wiping out the wealth of all their citizens. So for them, a lot of small nations, they would rather have that. That's why I mean, part of the EU, right? The other thing I think is super powerful about stablecoins versus just having access to dollars is, you know, I don't want to get too much into politics, but the administrations of the past couple of administrations have definitely weaponized

Pierson Marks (32:07)
Totally.

Totally. Totally.

Bilal Tahir (32:32)
the dollar and this is one of the reasons why you see a lot of nations not buying a lot of Treasuries because they don't want to be in position where suddenly they're strangled because they own a lot of US assets, denominated assets. With stable coins you don't get to weaponize it because you can't really freeze those assets or anything like that and so it kind of gives I think a second win to the dollar.

Pierson Marks (32:33)
Mm-hmm.

Bilal Tahir (32:56)
And I hope it does. I actually think there's a huge story that is still playing out about how stable coins can become the thing that kind of props up our basic whole economy in some way, because we can just keep printing. I mean, we're printing right now and in a way, the proxy has been stocks. Everyone just buys, US companies really the top 20 tech companies and stuff and we...

kind of keep printing and there's this whole story about it. With stablecoins, you can just have more one-to-one with that where we just print money and everyone just owns USDC.

Pierson Marks (33:28)
And some

more efficient economy. mean, like we said this last week, until there's no efficiency in the world, there's always going to be room for opportunity and growth. And so if you have the rails that are cheaper to do commerce and that commerce is transacting in US denominated cryptocurrency, you're going to have smaller transactions. You're going to be able to generate your image for one cent, like one stable coin cent. ⁓ So that's good for everybody, good for the entire economy.

Bilal Tahir (33:52)
Yes.

Pierson Marks (33:55)
good for Visa and MasterCard but you know who cares. ⁓ Yeah totally.

Bilal Tahir (33:55)
Yeah. Yeah.

I mean, they're probably going to jump in. think they are doing some profits and stuff.

same like to your question about why Stripe interested, because they understand this is going to be a huge part of the future and they want in on that action. so I think I forgot the stats, the guy shared on the talk, but it basically is like doubling or tripling like crazy, like in the volume and stuff. it's very much early days and getting towards a very explosion part of the curve. They want to be well placed for that.

Pierson Marks (34:21)
Right. And

so the one question I had before we wrap about Stripe and everything is that I get that Stripe sees this as the future, it makes sense, but where does Stripe fit in? If you're transacting me to you and I have a stable coin and I want to give you my stable coin.

Bilal Tahir (34:39)
Mm-hmm.

Pierson Marks (34:41)
Where does Stripe fit in that picture? Like, are they just saying, are they providing you the wallet? If you're the merchant, when you set up your SaaS platform to accept stable coins, like, you can just use Stripe and then you say, hey, here's a wallet address, like, allow people to pay. Like, how does Stripe fit into the picture of payments? Right.

Bilal Tahir (34:55)
Yeah.

I think as someone who's been in crypto for so many years, I still

find the idea of setting up a wallet a nightmare. There's different test nets, there's base nets, there's main nets, there's all these decisions which most users, let's be real, they're not gonna make. And so I think you're probably gonna see the stripes of the world, or other newer companies, they basically abstract this away. And so I can see...

me rather than me using a wallet right now the decision it's you use a wallet and then you go to Shopify merchant which is the partner they had the biggest partner they had for this crypto integration you go there you buy something and then the stable coins go to their stripe wallet

Pierson Marks (35:26)
Mm-hmm.

Bilal Tahir (35:34)
I can see that the wallet going away too. I just have a Stripe account, like a card or something. go and I just like, you know, I see that you support whatever coins and like, you know, like credit cards, master card, whatever. And I just say, go and go from a Stripe to Stripe account. I think there's, you're probably going to see more abstractions because most people are not going to just download Metamask wallets and you know, put coins in there and say, I don't know. I'm skeptical.

Pierson Marks (35:56)
I wonder if it just,

I wonder really if it just becomes like, don't even, consumers don't even know. Like do we move to a future where like when you pay on a Stripe supported account, it's just like automatically being paid through stable coins and like your card, your Bank of America or your Wells Fargo or whatever, like your balance is like probably just sitting in stable coins instead.

Bilal Tahir (36:03)
Yeah.

I mean, we figure, this has already

been figured, this is a solved problem within a closed ecosystem. It's called Cash App, it's called Venmo. I Venmo you money, whatever. But the thing is, I can't take those Venmo. I mean, I can deposit it to bank because they're dollars. And I think it should just be USDC or whatever. And you can take those, you can move them off, you can move them off platform, et cetera. And so, you know, just have.

basically a nice app for it. People understand the concept of having a cash amount, like a Venmo type account or whatever, and they can send you money, et cetera.

Pierson Marks (36:42)
Totally. That's super interesting. Well, like always, we start off in general media, we get to crypto, and it's just more of a tech talk. But like, it's always fun. It's always fun. well.

Bilal Tahir (36:48)
Yeah. Why not? Yeah. Yeah, there's

so many aspects to this. You know, gotta explore all of them.

Pierson Marks (36:58)
I know. I love it. It's super fun. talked about a lot. I think so next week we'll do episode 21. We got to episode 20. And yeah. So.

Bilal Tahir (37:08)
Yeah, this is great. 20

episodes in, top 1%, 1%, at least in episodes. Hopefully other metrics too soon. We'll see.

Pierson Marks (37:16)
Top

1 % of haters, just kidding. Not yet. Okay, well, sweet. I'll see you next week.

Bilal Tahir (37:19)
There you go.

Yep, take care.