PayPerQ is a bitcoin enabled ai interface that enables users to easily use all of the top ai tools without an account. Users pay per use with bitcoin and can switch the models they use on the fly without needing to provide an email address, phone number, or billing address.
Matt on X: https://x.com/MattAhlborg
PayPerQ on Nostr: https://primal.net/p/nprofile1qqsdy27dk8f9qk7qvrm94pkdtus9xtk970jpcp4w48k6cw0khfm06mss64u96
PayPerQ: https://ppq.ai/
EPISODE: 168
BLOCK: 905555
PRICE: 833 sats per dollar
(00:00:01) CNBC Intro
(00:03:14) Happy Bitcoin Monday
(00:05:00) PPQ AI
(00:14:05) Privacy and Payment Models in AI
(00:33:04) Cryptocurrency Payment Preferences
(00:57:28) Current State of Bitcoin and Future Perspectives
Video: https://primal.net/e/nevent1qqs8q9fgy40gxxkqpqjxmegqlnr0wt96qfqhgc2chjyeua7k8ptgtecnrukez
support dispatch: https://citadeldispatch.com/donate
nostr live chat: https://citadeldispatch.com/stream
odell nostr account: https://primal.net/odell
dispatch nostr account: https://primal.net/citadel
youtube: https://www.youtube.com/@CitadelDispatch
podcast: https://serve.podhome.fm/CitadelDispatch
stream sats to the show: https://www.fountain.fm/
rock the badge: https://citadeldispatch.com/shop
learn more about me: https://odell.xyz
[00:00:02]
CNBC Host 1:
Another all time high today topping a 118,000. The run up comes ahead of crypto week in Washington where congress is set to consider key legislation for the industry. Our next guest says the cryptocurrency could march to $1,000,000. Will Reeves is the CEO of Fold, the Bitcoin financial services platform, and he joins us now. I mean, Will, I'd love to sort of get your take on why we're seeing Bitcoin move higher. I know it doesn't always trade on fundamentals, but is anything that's happening in Washington potentially influencing the price of Bitcoin this week?
[00:00:32] Will Reeves:
You know, I think this is just inevitable. This is what happens when massive demand meets the finite supply of Bitcoin, and what we're seeing right now is the supply shock in motion. So I'd certainly say what we are seeing is the fundamentals of Bitcoin in play.
[00:00:48] CNBC Host 1:
And so you think that it will begin to start to move more on fundamentals and potentially even more so once it gets these sort of stamps of, hey. This is good stuff from Washington?
[00:00:58] Will Reeves:
I think the, stamps of approval have always been good. It enables new capital allocators to start to invest. And what we've seen recently, with the rise of Bitcoin treasury companies, the ETFs, is that entirely new classes of capital are able to invest in the asset. And what we're seeing is what happens when trillions of dollars start knocking on the door, the price tends to go up. And so the fundamentals have always been there. Investors have been waiting for their ability to get involved.
[00:01:30] CNBC Host 2:
What would it take for some of the other coins, altcoins, to kind of get the vote of approval or stamp of approval, if you will, in terms of Bitcoins? Because, I mean, if you think back to meme coins, the Doge's, the early stages of Bitcoin, it was very really the early adopters that really kind of formed that cohort that pushed it higher. Who will it take for us to see that transition into some of the other coins that are still branded as more speculative?
[00:01:56] Will Reeves:
You know, frankly, I think they're gonna remain speculative. Bitcoin dominance is at 60% in rising. I think what we're seeing is an inevitable trend towards the signal, the Apex asset. The issue is is that Bitcoin is the only credibly neutral digital asset that is in issuance. All the other tokens are still proving an extraordinarily speculative use case without the track record of Bitcoin. And so what we are seeing today is the continued march of Bitcoin dominating this entire market. And, I I am well, I'm waiting for the day when that changes, but ultimately, what we've seen is something that can't be reversed.
[00:03:15] ODELL:
Happy Bitcoin Monday, freaks. We got another still dispatch. The interactive live show focused on actual Bitcoin and Freedom Tech discussion. Looks like ZapStream is still under DDoS right now. I'm gonna keep an eye on that for us. But until then, I'm gonna remove that from our stream. Let's see what we got going here. Live producing going. We'll be streaming in the Twitch comments, the YouTube comments, and the x comments until ZapStream comes back online. Anyway, so dispatch the interactive live show focused on actionable Bitcoin and Freedom Tech discussion. We have another all time high Bitcoin rip going on Bitcoin maybe last night or yesterday, I guess, right before I fell asleep, past a hundred and twenty k.
Kieran from ZapStream said he just ordered a new router to deal with his DDoS. We have Matt Alborg here from PPQ, paper q dot a I. How's it going, Matt?
[00:04:20] Matt Ahlborg:
Hey. It's going well. Yeah. It's a pretty good day when Bitcoin's at an all time high to do a podcast. So
[00:04:27] ODELL:
Hopefully hopefully hopefully, this is the the rest of the summer. It's just constant all time high reps. I I'm here for it. So where should we get started? I mean, I think sorry. I'm reading troubleshooting. Your product's awesome. Let's do a quick to the freaks that don't know what ppq.ai is, what is it? Why does it exist? Why should people care?
[00:05:01] Matt Ahlborg:
Yeah. So PPQ, it's a website where you can get access to all of the latest and greatest AI models and tools in one place, and it's without subscription. It's on a pay per use basis, and it accepts all well, many cryptos, but we do have kind of a Bitcoin first, preference, 5% bonus for all lightning deposits in particular. But the basic the basic idea is that, you can come to our website. You can deposit as little as 10¢ worth of Bitcoin. You can try the product out. You're not committing to a monthly or annual subscription or anything like that. And, yeah, what we've what we've seen in the last year is that there is not one AI company that's releasing all of the best models.
Yeah. There are, half a dozen, awesome companies releasing really awesome stuff. And so it really pushes forward the idea that, like, the better the better, interfaces with AI, it's it's those that are plugged into all of the top company models.
[00:06:10] ODELL:
Yeah. I mean, I guess there's a bunch of things to unpack here, but trying to stay on top of, the AI curve or the AI acceleration phase where it just feels like new stuff is dropping every day. You kinda have you're you're you're kind of held back by as an individual that wants to use these tools, you're held back for two reasons. You're held back for one that you have a bunch of different companies that are iterating and innovating and shipping new stuff all the time. So in practice, you end up needing to have accounts everywhere. And then the second piece is it's very hard to use them with any semblance of privacy whatsoever.
They almost all require credit cards as a payment method, And then some of them even do other identity verification. Like, they try and make it even more difficult to use privately, let alone just the standard credit card billing stuff. So PPQ is really cool because you basically have all of the top models in one place, and then you can just easily pay with lightning. Probably. Awesome. And I I think it's I mean, it's pretty unique. Right? Is there anything else that kinda even competes with you?
[00:07:31] Matt Ahlborg:
Yeah. There are. There's one website in particular that I would consider to be our competitor, Nano GPT. They've done basically the same thing that we've done except they love the cryptocurrency Nano.
[00:07:45] ODELL:
I forgot that I forgot that shitcoin existed.
[00:07:48] Matt Ahlborg:
Yeah. But aside from that, aside from their, misguided shitcoining, they've done a great job executing. And so, we we are trying to answer in kind and and, we, we recently acquired a $500,000 investment from a couple of the, you know, popular Bitcoin VCs. And so we're scaling up now on the engineering side so that way we can really, really ship a lot of great stuff, in in the coming months. Yep.
[00:08:21] ODELL:
So let's so, I mean, if you look on if you look on your website, I mean, you can you can see what, different models you offer. But, you know, you have the OpenAI models there. You have the Anthropic models. You have XAI with Grok. You have the Google Gemini stuff, Mistral, Perplexity, Meta. How hard is it to actually integrate with their APIs? Like, do they make that difficult for you? Like, what is that process like to actually?
[00:08:54] Matt Ahlborg:
It it is, I would say, medium difficulty. We in many cases, with the top top models, we like to directly integrate with the provider. So we're getting our OpenAI models directly from OpenAI. We're getting our anthropic directly from anthropic. But with a lot of the, kind of second tier models and such, like, it would be a lot of overhead to set up direct integrations with all of these dozens of AI companies. So then we plug into, like, an aggregator that there are a bunch of companies coming up that are specializing in, like, providing all these models, in one place, which is kinda similar to what we do, as well. So, but I think our focus is that we want to make our platform, really user friendly where a lot of these kind of other aggregators, they're more at the they're more developer facing or even enterprise facing, whereas we're we're kind of trying to get to the, AI curious newbies or maybe crypto native person who's a little bit technologically minded, but maybe is is learning AI for the first time. That's the kind of users that we are capturing. Yeah.
[00:10:06] ODELL:
You're, like, more of the, like, the consumer focused front end while they're kinda offering, like, more of a b two b type of service.
[00:10:14] Matt Ahlborg:
Exactly. Yeah. Mhmm.
[00:10:17] ODELL:
And I does with Grok, is that, like, a direct API call? Like, they just open up an API for you? I assume you have to pay for it. But Yeah. So Grok is a tricky one. You can get a direct API access with XAI.
[00:10:31] Matt Ahlborg:
However, we're getting our Groc from one of the aggregators. When they released Groc four, like, a week ago, they kinda released it to a limited pool of of companies. And so for that reason, we are we're not getting it directly. But, we do have, you know, an XAI account. And, when they make it available, directly, we will probably plug in that way.
[00:10:54] ODELL:
And then what about, like, the open source models? Are you self hosting any like, are you hosting open source models, or are you, is it you're purely an API front end effectively? Yeah. We're we're basically
[00:11:05] Matt Ahlborg:
purely an API front end. Self hosting models is actually a lot of work, and, it's it's a lot of resources as well. Even, like Open Secret or or Maple AI, those guys, which I really like them, love what they're doing, and I've had several conversations with them. But the model that they're self hosting is it's still not the fully beefed up version
[00:11:32] ODELL:
of of, DeepSeq or or whatever open source model because those literally cost LLAMA they have LLAMA and DeepSeq Seek right now as their two supported models.
[00:11:42] Matt Ahlborg:
Yes. And but those particular instances of those models, they're not the beefed up version of the model. They're kind of like the middle of the road model. And and so because it costs a lot of money to host, one of those, the beefed up versions, like, I think on the order of $1,020,000 dollars a month, something like that. So you need a lot of demand to really justify hosting the fully beefed up open source version, And it's a lot of engineering to put into place and make sure that it's all, like, load balanced correctly and and all that stuff. So we've, avoided that stuff.
Another reason I would say is it's a lot of work, but another reason is the demand is not there, for a lot of the open source ones yet. They don't have the name brand recognition. There's still you know, when DeepSeq came out, a few months back, everybody everybody thought, like, oh my goodness. It's like, you know, it's just as good as as OpenAI and Anthropic. And then, like, as time passed, people figured out that, no. This model is not as good as the brand name models yet. And and there's still a lot of work to to to make up. And and as an end user of of AI, like, if I'm a developer every day, I would say a core user that we have is developers, using our API.
If I if I want to, do my job in the best way possible, I would rather use the absolute best model. There's no reason for me to use a model that's, like, 90% as good as something that that I you know, it's just not worth my time. So, yeah, so the open source models are not there yet. There's a really exciting one, that was released a few days ago. It's called Kimmy from Moonshot AI. It's it's like another Chinese open source model. Right? Yes. Yes. And it's getting a lot of buzz, and we're adding that, that one either tonight or tomorrow morning. And, so we'll see. It seems to have good reviews, but really, truly, the the test of time is really important for a lot of these models because, it's very easy to get your new model to pass all of the, like, the benchmarks, to kind of game it to to to do really well on benchmark testing. But then when it actually comes to, like, actual usage, it's it's a different story a lot of the time. So
[00:14:04] ODELL:
Right. I mean, the way I kinda think about it is, like, the open source models are usable right now for kind of, like, the average person chat interface type of scenario where you're asking for, like, populations of different countries and different stuff like that. But if you wanna do, like, really strong tasks, hard, aggressive tasks, whether that's, you know, generative art or vibe coding or something like that, there's a real big gap between the proprietary models and the Yes. And the open source ones. Would you agree on that? Totally. Yeah. And it it's but that that was not always the case. Like, even a year ago, when we've jumped from GPT 3.5
[00:14:48] Matt Ahlborg:
to GPT four turbo, like, GPT four turbo, I believe, was, like, the first model that really started doing well at answering those everyday questions, like population or whatever country. Like, 3.5 did the job, but Ford turbo really, like, was the first model that did well at all that, even the basic everyday questions. And so I think we are already at the point where, like, the basic everyday questions are free now. And and we've we've already moved past that, and now the expensive models are the ones that can do a lot of the fancy stuff. Yeah.
[00:15:26] ODELL:
Yeah. That makes sense to me. So, I mean, as an end user, like, obviously I mean, maybe I'm not the average end user, but I I personally have, you know, long term concerns about the viability of of open source models versus proprietary models. Like, I don't want us to go down a dystopian hole of, proprietary data surveillance heavy models. I mean, I think it puts a lot of disadvantages to the individual. It gives a lot of advantage to the corporations, to the governments that they are they collude with or get pressured by. But from a just an actual individual everyday point of view, the biggest issue is on the privacy side.
So for right now, like, the way you have PPQ AI set up, what is that trust model? Does that look like a is that like a VPN trust model? Like, are you basically sitting in between me and my, prompts versus with with the the AI model on the other side? Like, how does that look? How should a user think about the privacy model of your product?
[00:16:35] Matt Ahlborg:
Yeah. So, with PPQ, when a user comes and they enter in a question and then they hit the enter key, it that query, along with the content of that query, goes to our back end, and you have to trust that we are not logging, the content of your query on the back end. And we in our terms of service, we are telling our users that, like, we only log metadata, like, how many tokens the question was, how much it cost, all that stuff, and we're not logging the content, but we're not proving that. And so but what happens after it hits our back end is we send it on to the AI provider, and then the a AI provider returns the answer. The answer also goes through our back end including the content, And we're promising that we're not logging it, and we're sending it back to the front end.
That's the model right now, so you do have to trust us that we're not logging this stuff. I do think there is potential for, like, an architecture that could improve that where we could make better guarantees for our users. But, honestly, like, I did my best to get this product up and running as fast as possible. And I think that's, like, a very classic trade off between, like, a lot of what Bitcoiners want is, like, they always want you to to go absolutely perfectly in the direction of privacy. And if you're and if you're not all the way there, then you're you suck. And and, so I think that's actually one of my things that I'm good at is that I'm good at being practical and just shipping stuff that works.
And, I do very much appreciate privacy, and I wanna build that in a lot better over time. But that's where we're at right now. I wanna give people something that works.
[00:18:29] ODELL:
Yeah. I mean, I think it's, like, a perfect is the enemy of good classic example. And so it's basically the VPN trust model, more or less. So we we are you trust Mullvad or Proton not to log. Yeah. Yeah. Yeah. And so trust the relationship. But what about, presumably, if I'm doing a bunch of follow on prompts, the AI provider knows that's all from the same person, I assume. Right? I don't think so. So we to the AI provider, we are
[00:19:01] Matt Ahlborg:
one customer to them. So Okay. All of the queries from all of our users are being bundled into one, like so, now with certain data analysis and stuff, they theoretically could probably figure out, like, okay. These queries here are associated with They're very similar. Right? Yeah. Yeah. So they could do that, but I don't know. It I I But they're not getting, like, user x or user y or user ID code. User PPQ is what they're seeing. Yeah.
[00:19:33] ODELL:
And and so especially if it's if it's two different tasks that I'm doing. If I do if I, like, vibe code an app and then I go and just generate photos or something in a new chat window on PPQ, there's there's it's a pretty good privacy model there if if we trust you. I would say so. Yes. From the AI provider. Yep. Okay. That makes sense to me. The other cool aspect is is one of the few ways to actually use this stuff on, like, a pay per use basis, which, I mean, is hence the name pay per queue. How do you think about that? Like, are your costs like, it seems like the industry has more or less standardized, at least for consumer grade stuff, is standardized on, like, a monthly subscription model, where you just pay and you might you probably don't use it all or some power users are just destroying and destroying them and using way too much.
From an end user point of view, I love the pay per use model because, honestly, I'm I'm not a power user. I'm not using it all the time. And I like to know that I just when I use it, I pay for it, and when I don't, I don't. But how does that work for you? Are you pay are your bills,
[00:20:50] Matt Ahlborg:
usage based, or, like, how does that work? Yeah. My bills are usage based. PPQs are. We're we're when we plug into an API, we're not on any sort of, like, AI subscriptions with these companies. But, yeah, this is a really, really big question, actually. The the subscription model versus pay per use, how did we end up here? I think most people I have a lot of thoughts on this, but I'm not it's, like, not perfectly crystallized in my mind. So I'm probably just gonna ramble on here. But, with subscriptions, I think that, let's see. How do I start this?
So most people think that we arrived at this subscription models because, like, some MBA meeting at at all these companies agrees that it's like, oh, we should find a way to charge users for stuff that they don't use That's really good business model. Right? Like and then let's make them forget that they have a The gym memberships. Yeah. Exactly. Yeah. So I think that most people believe that that's how we ended up here because it's like a a nice way to steal money from your users, which I think there's some truth to that. But I think if you, I I worked at Bitrefill for five years, and one of my projects there was to kind of research the idea of getting the Bitrefill Visa card or Mastercard or whatever. And so I learned a lot about how the credit card companies work and how their fee structures work.
And, actually, I believe that we've arrived industry wide or, you know, global payments global payments, we've arrived at subscriptions not because of, it's it's a good way to steal money from users, but because the fees that Visa and Mastercard charge on microtransactions make microtransactions untenable in most cases. And, also, the fact that Visa and Mastercard have from their inception, they built in this chargeback mechanism, it makes it so that, anytime you do a payment on on these networks, there's a risk that it could be disputed and that it would be charged back. And for whatever reason, it means the what that amounts to is that small transaction numbers or or small volume numbers, are not, good for they they they become more expensive. So when you swipe a card, it's like 30¢ plus a percentage, typically. And so what that means is that a larger swipe is better, and that kind of rules out a lot of micro transactions.
But with crypto, it is there is no chargeback risk. It everything is instantly settled, and so we are now back at a place where pay per views actually starts to make sense. And and when you when you you know, there's all sorts of justifications for why we should have subscriptions, like, from a business point of view, like, you know, stealing money from users. And, like, we can also say, like, oh, now we have this recurring revenue thing, and that's good for you can you can, tell the the, the shareholders that we have our current Yeah. All this bullshit. But it's it's at the end of the day, what is a subscription? It's you're charging every user the same amount no matter their usage, and that just economically makes no sense at all.
You should be charging users based on their usage. And so I think that we are at the beginning of the end of subscriptions. I think that companies like PayPerQ, are going to start eating away at subscriptions. And the beautiful thing is is that you said you're a casual user. Right? So the the people who make the most money for these subscription providers are now gonna come to pay per queue because it costs them $2 a month instead of $20 a month. Right? And so who's left over in the subscriptions? It's the power. And I'm also I'm also, like, a casual user that wants
[00:24:52] ODELL:
the best model. Right? Like, I I don't I don't use it all the time, but I also I I would pay the best model. Use it. Yeah. Yeah.
[00:25:00] Matt Ahlborg:
Exactly. So that that actually helps us out against the freemium providers. Like, you can use ChatCBT for free, but they don't give you the document upload. They don't let you make images. They give you, like, a really shitty context window, all that stuff. So, yeah, PaperQ is a perfect place where you can actually get the high quality version of the model, the state of the art, but you're not paying a subscription. And so, yeah, it's it's great. Right. So I kinda cut you off, but what you were saying is
[00:25:30] ODELL:
you're kind of forcing the hand of the other providers to move to this model because the only ones who'd be he'll be left on the subscription model are the ones that they're not making money off of because they're just, oh, like, using the hell out of it. That's exactly where this is gonna end up, I I think. And I think AI is kind of the perfect,
[00:25:48] Matt Ahlborg:
way to to to really start forcing your hand because, like, Spotify subscription, I don't think that's going away anytime soon. Netflix subscription, that requires, like, massive licensing fees with all these big media conglomerates and stuff. Right? But this AI, I think, is it's the perfect way to really start eating away at the subscriptions.
[00:26:09] ODELL:
I mean, this is the first time that we've seen, like or maybe maybe I'm missing something, but it feels like one of the first times we've seen actual significant, like, on the Internet pay per use cost. Right? Like, real physical cost. Like, it has major computational cost every time you send them these prompts.
[00:26:30] Matt Ahlborg:
Yes. Yes. And that's kind of what makes it the perfect, like, the perfect, start to this end of subscriptions is that, yeah, there are, all sorts of views. Some are very casual. Some are power users. So the subscription models don't make sense anymore. And you can see that these subscription providers are already kind of, like they don't know what to do. Now they have, like, higher tier subscriptions. And even with, like, Cursor, if you've ever heard of Cursor, it's like this, it's like this, development tool that a lot of people are using. They have a 20 $20 subscription, then they have a $200 subscription.
The $20 subscription, they're already kind of, like, taking really big behind the scenes shortcuts. I I just recently hired a CTO, and he came in. And he and and I'm trying to he's he's, you know, he's got fifteen years of development experience. And this is a little bit of problem with with, like, some of the senior devs as you can't get them into the AI because they already know a lot of it. And, like, AI is really good for people like me who are not, seasoned developers. I'm kinda just I learned coding in my late twenties and, you know, I was never really a super full time person.
So it's great for me, but for these senior developers, like, they really have to be convinced on the value of AI. Well, I'm trying to convince him, like, you need to start using AI more because, like, this is our core user. You need to understand our product, blah blah blah. And the first week or two, he's like, hey. You know, I'm I'm doing my best, but, like but what what we figured out is that his $20 cursor subscription was, like, giving him really shit results. And that's he was forming his opinion of AI based on the shitty experience that he was having using $20 Cursor.
And then I was like, oh, that's crazy. Like, try using Cline. Cline is an open source version of Cursor that works on a pay per use model. And when it's pay per use, you're getting the model that you're paying for. It's giving you the full quality of the model. There's no secret, shortcuts behind the scenes. What we think is going on with Cursor is they're like, to save money, they're giving you lesser models at certain times, and and they're they're shrinking the context window. So it's not reading all of your files. It's like throttling. It's throttling you. Yes. And so when he switched over to Cline on paper use, he's like, holy shit. This thing is doing so much better than Cursor.
And so you you can already see there are really a lot of serious cracks forming in these subscription models.
[00:29:09] ODELL:
So with with your business model specifically, are you just charging, like, a flat rate margin on top of every usage?
[00:29:18] Matt Ahlborg:
That's what we're doing. I would like to transition to kind of a, kind of like a logarithmic pricing where, like, if you're a casual user, we're gonna charge you a higher margin. But if you use us a bunch over a thirty day window,
[00:29:33] ODELL:
your margin's gonna lower. Like a Bitcoin exchange.
[00:29:36] Matt Ahlborg:
Yeah. Exactly. There you go. Same thing. So I would like to implement that. That way, we win both user both types of users.
[00:29:44] ODELL:
But right now, it's just a fixed fee on top of that. That's pretty cool. I mean, you're just profitable and sustainable from the get go.
[00:29:50] Matt Ahlborg:
Yeah. Exactly. And it's and it's simple. That's the other thing with subscription. I told you, don't get me started on subscription. Every single AI website on the Internet right now, it has the same onboarding flow. You go to their website. They give you this watered down narrow version of what they actually offer, or maybe they'll give you a seven day trial or whatever. And then they try to get you on that subscription. Right? So there's so much, design thought that goes into converting a user to become a subscriber. Right? Because that's a big commitment. Right? To get somebody $20 subscription, that's $240 a year. So to get people to get a user to cross that, you have to design things so well, to get them to to to sign up for the subscription. On PPQ, like, we don't have to think about that stuff at all. You come. You make a deposit. We charge you whatever the price is, and and boom. You like it, good. If you don't like it, then you didn't lose anything. You're not committed. You know? Is is your margin is, like, the fee you charge on top public? Or it's not. But, I mean, we charge 40% on the UI.
Okay. And it still adds up to, like, you know, you're doing advanced queries and it's 2 pennies per query type thing. Yeah. It still feels pretty cheap. It's very cheap. Especially if you're not, if if you're not, like, an insane user. Yeah. Our median user is spending, like, a buck 50 a month. But on our API, we have a much lower margin, because the API users are using us, like, a lot more, and they're using us with the the IDEs and stuff. So, our I think our margin on the API is, like, 18, but we're gonna go even lower soon. Maybe 10%, something like that. So,
[00:31:47] ODELL:
what does an API user look like? Are they using something like Goose on the front end and they're connecting it? And, like, how does how are they using that product?
[00:31:56] Matt Ahlborg:
So they get so we offer an API key. Okay. What sort of PPQ, you get a little a p a PPQ API key. Our API key is what's considered OpenAI compatible, which means you can plug our API key in literally dozens, hundreds now of third party tools. And so, yes, Goose, Klein, Roo, the tons and tons and tons Zed.
[00:32:24] ODELL:
So they're just using a different front end with your API, basically. Yeah. They're they're using
[00:32:29] Matt Ahlborg:
I I would say, yes. They're they're using a different front end, but most of the time, it's an IDE, an integrated development environment. So it's like, yeah, they're they're looking at code. They're making calls. They're making many calls in succession as it reads their code files and stuff like that. Yeah. Got it.
[00:32:49] ODELL:
So, I mean, you said this in the beginning, and I wanna pull back to it. I mean, the biggest friction presumably with your product right now, and the death of subscription models is, like, people actually using Bitcoin, to make payments. Obviously, someone like me, I go to the website. I wanna test it out. It's like I just I have, like, 15 different Bitcoin wallets I could pull up to to pay it. But, like, the average person that doesn't have Bitcoin, maybe that's not the case. Maybe they get hit with some friction there. You said you mentioned that Lightning specifically you're basically giving a 5% discount to people. If they pay with Lightning, you'd rather than pay with Lightning. Obviously, I mean, our listeners, I think, would agree with me that Lightning is, like, the single best payment method for this type of thing. First of all, we get great privacy. But just from the UX point of view, it's just like point, click, go.
You have a bit refill background, but you also have a Useful Tulips background. I mean, that's what I knew you from. Freaks, if you didn't put two and two together, this is Matt Alberg from Useful Tulips, the OG p two p trading dashboard that was, like, tracking. I think it does it still exist?
[00:34:02] Matt Ahlborg:
Yep. No. It's not Not updated? Anymore. I think the website
[00:34:07] ODELL:
it was tracking all the p two p sales data from, like, all the different countries in the world, had great charts and whatnot. But you also have this bit refill background. And one of I've had Sergei on the show before, and I also, you know, have become relatively close with him. And one of the most fascinating things about him and his relationship with Bitrefill is the payment breakdown. So I'm kinda curious, like, what do you see payment wise in terms of Bitcoin on chain, Lightning, the other cryptos, currencies you support? Like, how does that look in practice?
[00:34:41] Matt Ahlborg:
Yeah. So Lightning is number one. It's just over majority. Number two, let's see. If you if you can make a guess on our number two. Is it Tether?
[00:34:54] ODELL:
Nope. Oh, is it Monero?
[00:34:57] Matt Ahlborg:
Monero. Yeah. So number two is Monero on chain Bitcoin. Yes. On chain Bitcoin is pretty small.
[00:35:04] ODELL:
And just because if they already have on chain Bitcoin, they'll probably just pay you in lightning, I guess.
[00:35:08] Matt Ahlborg:
Yeah. Also, our on chain Bitcoin has been kind of, like, finicky. So it's hard to give, like, statistical significance sometimes because, like, our payment methods were not always a 100% up. Okay. And it goes with our stable coins. Like, we had a really shitty stable coin payment processor with a lot of errors. So, like, the tether is probably a little lower than it should be. Got it. But coming from Bitrefill, I know exactly the payment popularity over there, and that's why I, chose to add the payment methods I did on PBQ because I'm already very familiar with what, like, the crypto payment landscape looks like. That's why I know that actually Litecoin is is actually pretty important payment coin to add.
[00:35:53] ODELL:
I think Do you see a lot of Litecoin payments?
[00:35:56] Matt Ahlborg:
It's lower than I thought it would be, but I think it's still around, like, 15%, something like that. Interesting. Way higher than I would expect. On Bitrefill, it's pretty high. Like, still not insanely high, but, yeah, 15% plus, something like that. Yeah. There there's there's a crazy thing with, like, the older the coin and and the more, you know, the older the coin, the more exchanges that it's listed on, the more wallets that support it. These coins gain massive network effects as well even if they're just the carbon copy of what Bitcoin is.
And it is I think it's very fascinating how Litecoin has kind of hung in there with without really any innovation or any nobody's ever been cheerleading Litecoin to a large degree. It's just kind of like this zombie project of sorts that just keeps kicking because it's been around so long. And they have actually improved Litecoin. They have, like, Mweb. It's like Mimble, Wimble, something or other. It's a privacy hack. They recently added the Litecoin. I don't know if it's good or not, but, yeah. I I I don't hate on the Litecoiners as much as I do some others, because I do think they have a lot of shared ideals with and they they a lot of them come from back in the day when the the early days of, you know, 2013, 2014 when when we were all learning about Bitcoin together and that type of thing. So I'm as the original shitcoin.
[00:37:32] ODELL:
I, so I'm looking here and, so I see you offer you offer lightning, Bitcoin, Monero, liquid Bitcoin
[00:37:43] Matt Ahlborg:
Yep.
[00:37:45] ODELL:
USDC, and USDT. And you also offer debit credit card. Do a lot of people do debit credit card?
[00:37:51] Matt Ahlborg:
It's it's pretty low, but but not nothing. And I think it's one of those that could really pop off if we get, like, the right tweet from the right person or the right I don't know. Because everybody has most people in the West have credit cards. And Yeah. As you said the standard. Are the people with a lightning wallet, around aside from us, Bitcoiners.
[00:38:15] ODELL:
So do you have a lot of fraud that you have to deal with? Because this is kind of a classic chargeback, part of the story. Right? Like, I'm I'm interesting thing.
[00:38:24] Matt Ahlborg:
As I said, the credit card payments are not that high, but we have had several instances where Stripe will send me an email and say, like, hey, this payment that happened a month ago, we might have to, like, take your money back. And it's like it it just proves to me, like, that this this isn't you know, it's it's it's Right. It's bullshit, like, how they can just steal your money back after an entire month, and it just kind of adds fuel to the fire that crypto and and, Bitcoin Lightning are the way to go in terms of, accepting payments because I've never had a single chargeback on Lightning. No. It's impossible. I mean,
[00:39:04] ODELL:
it's a it's a classic chargeback business risk for this particular use case because you're providing compute. Like, the it's gone. Like, you've already provided it. There's no you can't claw back the goods. You can't wait to see if there's a chargeback or not. It's weird, though, because the few times that it's happened, I'll go and I'll look at the user that that or the the credit ID that loaded.
[00:39:29] Matt Ahlborg:
And I don't think any of the times had they ever spent their entire balance. So it's I don't know what's going on there. But I mean, maybe you haven't actually gotten hit with a fraud, like, a fraud wave yet. You're just getting
[00:39:42] ODELL:
users that had regret for signing up for the service and just like issue. There was just some machine learning red flag that was the most positive or something. Yeah. I mean, that could become a major problem for you if you actually get scale on the credit card side. Yeah. It could. I understand. It's a trade off.
[00:40:01] Matt Ahlborg:
On the day that that happens, I'll be happy to tweet about it because this is a yeah. I think chargebacks are a huge problem. Like, I mean, I guess they have some some advantages, and there are reasons they added them. But I think if, like, if if if you went back to the nineteen seventies when they initiated this thing, like, I think Visa, Mastercard, they had no idea the cost that chargebacks were going to introduce into their systems, especially in the age of the Internet where, like, you know, fucking fraud is everywhere. I think they would really rethink, the fact that, you know, they would make payments final. And and that's why I think a lot of the the the crypto payments happening that are instantly settled, that's it's a huge advantage.
And I think that's gonna be in the same way that I'm clawing away at subscriptions, I think crypto is clawing away at Visa, Mastercard because chargebacks are like a really big thorn in their side.
[00:41:01] ODELL:
Yeah. I mean, it's actually the merchant. I also I mean, I just think it's a better payment experience from the user point of view,
[00:41:07] Matt Ahlborg:
just from, like, a UX point of view. Yeah. Even when you're a user and you make a credit card payment and it goes through and both you and the business are happy, there's still a fee involved that reflects the amount of fraud that's happening across the network. Like, when you that 3% that we all hear about, a big chunk of that 3% is because of all this fraud mitigation,
[00:41:31] ODELL:
that these companies have to put up with. So Yeah. They're social they're effectively socializing the I mean, it's a it's a they're socializing the cost. They're they're spreading it out. They're doing a risk reward adjustment to make profit. I see Gary Krause in the comments, rider or die freak, asking about eCash. What are your thoughts on that front? I assume he's talking about using eCash actually internally to try and add additional privacy to the system.
[00:42:01] Matt Ahlborg:
Yeah. So I'm not an eCash expert, but I've been learning a bit more. A lot of the ecash aficionados have come to me and asked for me to add this. There is routes that are out there, which I'm sure some some users are also familiar with, which uses the ecash tokens. So what you get when you add ecash, as far as I understand, is it basically makes it so that PPQ so right now, when a user makes deposits $10, they get a credit ID associated with them. So I can actually link If I wanted to, I could link all of the users' queries together, and I could see, like, are trusting you. Yeah. And I could I could say this person who ran five queries on Tuesday and then four queries on Wednesday and whatever. But with the eCash, it becomes very it it becomes anonymized, basically, where you can't even tell, you know, who who's even among your credit IDs, which credit ID is making that query. So there is a price to gain by adding that.
The problem with ECash is is the same problem as Lightning, only it's magnified by 10 x is that, like, not many people hold ECash to pay with it. Right? And so, like, I even tried to do Routester myself, and I was like, okay. I wanna go to Routester. I wanna, like, try this thing out. Oh, I don't have ecash tokens. Where the hell do I get ecash? No. But I think I think the point is
[00:43:28] ODELL:
for you to replace credits internally with ecash. So, like, I pay with lightning or credit card or Monero or whatever, and then you give me PPQ ecash tokens, right, that you manage.
[00:43:41] Matt Ahlborg:
I'm issuing them to the user. Yeah. That is another way to do it as well.
[00:43:46] ODELL:
And then that way and that way, then I would have the bearer tokens. I would have the ecash bearer tokens. Right? And then it disconnects me to a degree. It reduces trust in you. Yes. That is true. Trust model. Yeah.
[00:43:59] Matt Ahlborg:
Yeah. That now that I think about it, that is, there is a distinguished there's different ways you could use eCash, and the better way is, yeah, you don't have users bringing you eCash tokens Don't do that. Giving them eCash tokens when they make a lightning deposit.
[00:44:15] ODELL:
And and that improves, like so the trade off there is additional complexity on your side, particularly when you first set it up because you said you're not an eCash expert where there's about 10. So there's, like, not that many. I will say most of them are, you know, vibe coding aficionados. So, like, I'm sure a bunch of them would be willing to help out, like, specifically their leader, Cali.
[00:44:40] Matt Ahlborg:
It's it had Cali reach out to me a couple of times. And so, yes, it is something that that I I would, I would look into and and and and, you know, it it's just a matter of, like, is it gonna be next month, or, maybe two, three months from now? I I don't know. Fair enough.
[00:45:00] ODELL:
Yeah. Yeah. I mean, I I mean, I think the argument isn't really this isn't a, Bitcoiners pushing for perfect kind of thing. I mean, I think long term, you're in probably in a better position if you don't have the liability of having so much information on users. Right? Like, if PPQ is successful at scale and you can actually connect the dots on users, you're gonna get tons of requests and, you'll have tons of information liability, which we've seen with every VPN provider. Like, I can't imagine I would that's not a business I would ever wanna be in. It's like the hosted VPN provider business. And so if you can reduce some of that data liability, data exposure, it's probably a big win for you long term if you see scale.
[00:45:48] Matt Ahlborg:
That, yeah, that makes a lot of sense, actually. Yeah. I I I will definitely think about that. I don't know if the answer, though, if if if a regular comes to you and you say, like,
[00:46:01] ODELL:
sorry. I don't have the information that you're looking for. Like, I think most of the time, their knee jerk reaction is like, well, since you don't have it at all, then we're shutting you down. Well, no. They I mean, I don't think they could shut you down. I think they could try and pass additional regulations. I mean, it just adds it adds some level of friction. I mean, we see this with stuff like signal, right, where signal does respond to government requests. Just the data they hand over is, like, basically nothing. They they, like, hand over a phone number. They can hand over an I IP addresses that connect to their servers. But that's basically it. They don't have metadata, and they don't have the message contents.
Yeah. And so what happens what happens is is for the high value targets, governments just break into their phone or their computer and then get the signal messages that way, which they could do for this too. And so, presumably, it kinda plays out the same way where they just go that route. They don't give you too much trouble. And if they wanna add regulation on top, they can add regulation. The advantage of the ECash model, at least I mean, Cali's become a pretty good friend of mine. Is that, like, he even though it is adds a level of complexity, it is a use case that he's thought a lot about. And so they do basically have, like, an SDK, and it's relatively easy to implement.
Yeah. I mean, his dream is, like, Amazon gift cards being e cash, like, their own internal e cash, so you have some level of privacy
[00:47:31] Matt Ahlborg:
attached to it. Yeah. The one last, I do I do really, take point with what you're saying about, having too much information and and that being becoming a liability. The one last pushback I'll I'll make is that it's actually nice to have some level of metadata on users because to know what countries my users are coming from helps me in terms of marketing and business development, to know what you know, if if they if a user comes like, one question I have now is, like, how much value is adding all these media models giving me? Media models like image generation generation, like, if I can pull a new user in because they can do some cool image model, but then I can transition them to to using me on a daily basis for text queries or whatever, then it's worth it to add more image and video models. Right? So if I can actually see that about a particular credit ID where they come in, they make a they make a couple of image queries, and then I can see that they transition. That helps me from a business development standpoint.
I get it, though. It's not perfectly private. And so, you know, there's there's a gray area between, like
[00:48:51] ODELL:
but it's It's trade offs. It's trade offs all the way down. Yeah. Yeah. It's hard to build a product without having With knowing nothing about your users. Yeah. User I've seen that firsthand, you know, particularly on the ten thirty one side and some of the businesses we've invested in. It makes it really difficult. You're just throwing estimates around.
[00:49:11] Matt Ahlborg:
And and I don't even in most cases, I don't even get the email address of my users. So, like, I can't email them and say, like, hey. Why did you leave our platform? Like, what what did we do badly? Like, I have this little customer chat widget that is actually really awesome because that is my primary way that I'm talking to customers. But aside from that, like, yeah, I don't even have emails most of the time. And so it's really hard to to if I didn't have that customer chat thing, there's literally hundreds of bugs that users have reported to me, using that. And and it and it shows that, like, you have to have some level of connection with your users, in in, you know, I yeah. So it's really important.
[00:49:58] ODELL:
Yeah. Fair enough. Yeah. I mean, you could also you could just make it optional if you wanted to. You could give people a you could give people a discount too if they if they stay in the, you know, less private, non e cash model. No matter what, you still have the IP addresses of the you still get some data. Right? Like, it doesn't
[00:50:23] Matt Ahlborg:
Honestly, I haven't thought about exactly what data I would be losing that hard, but that was a question that I posed to Now the big thing you would lose is user behavior data, which is very helpful in product iteration.
[00:50:37] ODELL:
The other thing is you can iterate the product without it be and then you could turn on the private privacy preserving. And maybe maybe three months is too soon. Maybe it's maybe it's eighteen months, twenty four months. When you have a general idea of what the average user flow looks like, you're not losing as much. You've already iterated on the product more.
[00:50:57] Matt Ahlborg:
Yeah. There's Yeah. Always iteration to do, though, but, I I mean, I hear you. Yeah. Yeah.
[00:51:03] ODELL:
I just, is probably my single biggest concern with the AI acceleration is the disadvantage that privacy like, for all of digital humanity or for all of our lives, we're probably relatively similar in age. If you wanted to have privacy in the digital age and have freedom in the digital age, you were at a disadvantage. And, one of the cool things about Bitcoin, to be honest, is is that it actually flips the script on that. Someone who uses Bitcoin privately and in freedom oriented way actually has a lot of advantages over someone who just owns ETF or owns MSTR or something like that. So it's like a natural advantage that you have actually putting in the work and and doing that. But on the AI side, it really accelerates it. It's like, if if I'm competing with someone in the job market and they have access to chat GPT and they're using it in the least private way possible and feeding in all of their private information into it, that that model and that agent is gonna become so much more powerful to them. They're just gonna give them almost superpowers compared to the guy who's using the freedom focused one.
Yeah. And it becomes this massive data collection hellhole. Like, it's the first real model that incentivizes data collection on the Internet was the ad sales. Right? It was like, if you could track someone and you could sell their data to advertisers, you could monetize them better. It was like the your data cattle. Like, monetize the cattle as much as possible. This is the second major model. This is like the the second time in in digital human history where you where corporations have, like, a huge, huge incentive to to track people and collect as much data as possible and then train their models off of it, make them better, do all this other stuff.
So it's just something I think about a lot. I'm not pretending like there's a perfect way to do it. But
[00:52:57] Matt Ahlborg:
I mean, I think about it a lot too, and I haven't seen reasons yet to be super optimistic about it. Right. Which sucks. I think maybe maybe the maybe it is the fact that, like, you know, we did have DeepSeq, which came along, and it was a very Right. Good model. And so if we do continue to see, these open source models that are approaching the capability of of the proprietary models, then maybe there is, some hope in there. Okay. Yeah. I I think maybe maybe yeah. I'm just really theorizing here, but, like, there could be major incidents coming up in the future where, like, you know, we we have all sorts of data leaks nowadays. Right? Yeah. But, I mean, a lot of the time, I've never been impacted by a data leak myself. I'm sure were in the TransUnion,
[00:53:55] ODELL:
Social Security number thing where, like, every adult was.
[00:53:59] Matt Ahlborg:
I'm sure I was, but I guess my point is is I never felt the impact of it. You know? Like, I never my identity was never stolen. Like, I know it's common, but what I'm trying to say is, like, maybe there will be an incident where, like, you know, a large number of people are substantially impacted, and then that's the moment that governments actually step in and enforce privacy or start to enforce privacy. But maybe I'm just being really crazy and and, that's never gonna happen. I don't know. But I'm just trying to think about how it could come about
[00:54:31] ODELL:
when what are the possible inflection points? I mean, I'll say the bull case in my mind is that the data leaks are gonna get worse and worse. And but rather than government stepping in, even though they should, I think it's a national security issue. But rather than government stepping in, I think, you know, you get you get a handful of new users that seek out the better methods. Now the question is, and you don't get most of them. Right? Like, we just saw, what the main DNA provider like, they got, like, sold to, like, the Chinese. They got a lot of they got everyone's DNA that provided it to, 21 and me. And, like, most people didn't bat an eye. But some people did, and some people were like, holy shit. I just doxed my entire Yeah.
Genealogical branch. And but but then at that point, to have tools available to them so that they're not at a huge disadvantage, but they can still focus on the privacy freedom way, I think, is key. Because most of the time, they're at a disadvantage. Now I wanna be clear here. I'm not trying to, denigrate PPQ. I think actually PPQ right now is one of the best tools someone in a freedom privacy focused way can use. And even without eCash, like, I can log in with a VPN. And when I say log in, no login information whatsoever. Just pay with Lightning. I can do a set of prompts, and then I can just close that session and then do a new session from a different VPN later. And then I have similar privacy guarantees as if you were using eCash on your side.
Yeah. So, like, it is the best way for me to use top end models in a privacy focused way if the user knows what they're doing. Yeah. So thank you for that. I'm like, I'd I I wanna be clear here that I'm not like Yeah. Sure. Yeah. I appreciate that. And and, yeah, yeah.
[00:56:19] Matt Ahlborg:
Hot seat. We're trying to take the practical approach here, and I know that it's certainly not perfect. But,
[00:56:27] ODELL:
it's awesome. No. Quite frankly, it's awesome. So thank you. Yeah. Most people are appreciative, and and that's what keeps us going for sure. And I will say on the image generation side, you said, like, user behavior, does image generation bring you in? There's no other way for me to use, like, the top end image generation stuff. Exactly. There's no other way to use the software. Yeah. And it like, I don't use them for business stuff. I just use them because, like, I wanna create, like, shit posts or
[00:56:56] Matt Ahlborg:
memes or whatever. Or, like, yeah, meme troll troll family members and friends and stuff like that. Exactly. So,
[00:57:04] ODELL:
and I it it took me like, you were the first time I ever got to use a real image generation model, because I refuse to, you know, I never sign I'm probably one of I mean, there's plenty of people that haven't, but I've never signed up for chat g p t or
[00:57:20] Matt Ahlborg:
Yeah.
[00:57:21] ODELL:
Any of the major image generation proprietary models. Yeah. What else? I, I'm just kinda curious on your opinion on, like, the current state of Bitcoin. I mean, you were, like, the p two p guy. You were, like, the p two p freedom money guy, and you pivoted to AI. And then meanwhile, we have it's like paper Bitcoin somewhere. We got ETFs and MSTR and stuff. How are you thinking but we're also at all time highs. More people using Freedom Money than ever before. How do you think about where we're currently at in Bitcoin?
[00:58:05] Matt Ahlborg:
Yeah. It's I'm I have a big mess of thoughts. Yeah. I on the one hand, I think I underestimated the amount that the, you know, corporatization of Bitcoin would, like, help its price. I always thought that it would be the kind of, medium of exchange. Everybody kind of slowly grassroots adopting it, type thing, which hasn't popped off yet. And but I'm still I'm still in that corner, obviously, because I'm building PPQ. It's, hugely dependent on the success of Bitcoin as medium of exchange. Yeah. I don't know. Like I said, it's a big mess of thoughts about Bitcoin. I think that a lot of the p two p stuff went to stablecoins. And I think that, like, this gigantic plurality of all of these cryptos, they accomplish for a lot of users the goal of the user, and it doesn't necessarily require Bitcoin a lot of the time. So when they wanna do when a user wants to do they're they're living in a developing country and they wanna do a cross border transfer. Bitcoin is one option among literally thousands, and and stablecoin is probably the way that they're going to do it. Specifically Tether?
Yeah. Tether or even yeah. There's there's a there's a bunch. But, yeah. So in the way I pictured this journey on Bitcoin going is I always thought that, like like, Bitcoin is is kind of built for the financial oppression. Nuclear winner is kind of what I like to to I always thought the governments were gonna crack down hard, and we would need this very decentralized network, that that survived hard times. And we would need this lightning network, which is very difficult to build out and very hard for users to learn how to use, and there's so many technical, trade offs, that are happening or technical.
It's very conservative and because it wants to guarantee that sovereignty. But the the end result is, like, governments I guess they tried to crack down on in some degree, but, like, they were never really that successful. And so what we're left with is, like, there's literally thousands of shit coins out there. There's all these stable coins that were never cracked down upon, and now stable coins are actually becoming important for government, the US government in particular, because it's, you know, extending the timeline of the dollar.
So, yeah, I don't know. Bitcoin has gone in directions that I I guess in some ways I was right about, but in other ways, I was very wrong about. So, yeah, I don't know. A mess of thoughts, like I said. I mean, the way I think about it is
[01:01:02] ODELL:
is you really, I think it just really comes down to the extremes. Right? So, like, I think in, permissible times, right, and you can get around the, like, the Tron, Like, Tron can get away with just running, like, seven servers on Yeah. Corporate cloud servers and send Tether around the world and and be, like, the number one medium of exchange. Like, that works. Or you need to have, like, cockroach money, which is Bitcoin. Like, all the stuff in the middle is pretty much useless. Yeah. And and so it's like you owe Bitcoin is the safe haven. We have it no matter what. When there's no other option, you have Bitcoin. And then the majority of people are gonna use Tether, which is, you know, more or less, like, being indoctrinated as, like, almost a quasi CBDC, but, like, the best possible version of a CBDC because it's corporate run by Bitcoiners with that try and push back against seizing funds and whatnot, but still completely transparent and tied to the dollar. And everything in the middle is just, like, caught in no man's land.
Yeah. And that's kind of the reality we're in right now. And then 99% of new users that actually want Bitcoin price exposure are buying MSTR or
[01:02:19] Matt Ahlborg:
Do you think it's 99 or you think it's really that high? I think it's
[01:02:23] ODELL:
I think it's quite high. I mean, in terms of in terms of dollar amount, not in terms of users. Yeah. Yeah. Like, in terms of actual volumes, probably it's, you know, over 98%. Maybe I'm being provocative, but it's probably over 98%. But in terms of users, in terms of number of users, maybe we still have them beat because we have, like, the developing world and global permissionless markets. But in but but a lot of them are probably coming in with with smaller sums. Right? If someone's gonna allocate a $100,000,000 into Bitcoin, probably coming into iBit.
[01:03:01] Matt Ahlborg:
Yeah. Yeah. I don't know. It's yeah. I think you you said a lot of those things well, but,
[01:03:10] ODELL:
yeah, who knows who knows where we're at. And then my bull case my bull case is, I think, we have more people using Bitcoin as Freedom Money than ever before. Like, the number of people that are making self custody lightning payments higher than it's ever been. And the cool part is it's hard for us to calculate what that number is. The number of people self custodying Bitcoin, higher than it's ever been. Percentage wise, lower than it's been in the past. Yeah. But, in terms of absolute numbers, that number goes up.
[01:03:40] Matt Ahlborg:
It's true. And and I I think that's the fact that people say, like, well, we don't know how many lightning transactions are happening. That's a little bit of a I think, people are hoping that there's more than there are. There probably aren't. Yeah. But as I was the guy at Bitrefill, and Bitrefill is actually an excellent metric on, like because we accept all these cryptos and Yeah. And so we can actually see, like, how many users are actually bringing us lightning payments. And the answer is exactly what you said. It's in absolute numbers. It's never been higher. But in proportional numbers, it's actually, like, being beat out by a lot of other cryptos. So that's kind of a good news and bad news situation.
But, yeah, at the end of the day, I think I think Lightning is still, slowly but steadily improving, getting better, gaining users. I don't know. Is it enough to kind of keep up with with a lot of the other stuff that's happening? I don't know. And and, again, it it goes back to the question of, like, is this other stuff gonna be ever cracked down upon, like the Yeah. Tether on Tron type thing? I don't know. Probably not, though. I would I would say less likely than it was three years ago. Well, I mean, I think the cool thing about the vision of Bitcoin as, like, a Bitcoin on chain being, like, a foundational protocol layer that itself is cockroach
[01:05:07] ODELL:
type of trade off model, like, trying to survive a nuclear winter, is key because then you can build stuff on top of of it. And and Lightning itself took a certain set of trade offs to be censorship resistant and robust. And we just really haven't seen, besides Liquid, which kinda just never found product market fit, we've never seen, like, a more centralized trust model. And now there's some things being experimented on with, like, Arc and Spark that are coincidentally named the same, that rhyme with each other, or Cashew even. Right? Like, oh, how do we make more privacy preserving custodians? And so then all of a sudden, we could see we could see things go down to more centralized this like, maybe there's there's there's more usage on the centralized stuff.
But at the end of the day, if shit hits the fan, we can pull back to, like, the actual cockroach stuff. Maybe that's the bull case. I don't know. Is I also, like, I'm glad we're not getting cracked down on as much as I assumed. Like, I thought Bitcoin was gonna be straight up illegal in America in, like, 2017. So I still think it's the best timeline, and we'll kinda see how it all kinda see how it all plays out.
[01:06:18] Matt Ahlborg:
My bowl case is, like, I think it's, yeah, it's it's Bitcoin is its monetary properties. It's it's not the medium of exchange stuff that is really popping off. It's it's the fact that it is, competing as this beautiful, like, finite finance, finite supply, money on a on a global stage where everything else is just infinite, printing. I that's something I never that's the angle that I never really gave enough weight to, but I think it's totally starting to play out now where and I think, yeah, the the really great thing about one of the really great things about Bitcoin also is that it makes you start to ask questions about the fiat system and how it runs. Because until Bitcoin came along, nobody really ever asked questions. I'm like, okay. So you're just making money out of thin air. Like, what are the rules for that? Like, do we even know how much money you're printing? It says here you're printing this much, but then you have all these, like, classified budgets that are unaccounted for. So you're probably printing a lot more than you say you're printing and, like, the the holes just start getting poked into their their their presentation of what fiat is.
And and I think I think they're absolutely starting to lose the argument on, like, you know, before they could kind of just sit back and say, like, yeah. We we have this fiat thing, and we have all these, like, balancing mechanisms for it. And it's too complicated for you to understand, so just don't worry about it. But now because people are learning about, Finite money, and and the advantages of it, they're starting to ask questions about all these, like, bullshit, tricks that that the fiat system makes, and I think they're losing those arguments a lot. I love it when, like, the European Central Bank, like, post something on Twitter nowadays. It's, like, 98% angry Bitcoiners, like, just, telling them telling them, like, all of the lies that they're saying. And,
[01:08:28] ODELL:
yeah. I think they're losing the the the narrative war, the the story war. Yeah. I mean, even the term fiat, like, the term fiat was popularized by Bitcoiners. Like, that was not a term that was, like, in Yeah. Normal discourse and has become such recently, or, like, the last few years mostly because of the normalization from Bitcoiners. I mean, and just to bring it back, by the way, when I said, like, we don't know exactly how many people are making Bitcoin transactions or whatever, I actually like, I think it's on the lower end. Like, my question here's a question.
Just spitballing. No one's gonna hold you to it, and, also, no one can ever prove that you're wrong. How many Bitcoiners do you think exist in the world that own more than a half of Bitcoin in self custody. So more than I used to do this with a Bitcoin, but now we're at a 120. So I'll do a half of Bitcoin, $60,000 in self custody worth of Bitcoin. How many Bitcoins do you think exist?
[01:09:32] Matt Ahlborg:
Well, it can't be over 42,000,000. Right? So because that would be more than Right. So I would say, yeah, I would say maybe 5,000,000.
[01:09:44] ODELL:
I think it's less than that. So if you ask if you ask Probably is. Yeah. If you open PPQ and ask Grok three, how many own more than one Bitcoin, It comes up to a 100 to 250,000, and it gives you, like, all its reasoning or whatever.
[01:10:03] Matt Ahlborg:
Okay. Yeah. It's possible. Yeah.
[01:10:06] ODELL:
I think it's I think it's less than a million. What? What what is the point that you're driving at here? My point is is, first of all, it's cool that we have no idea. Just making a guess. Yeah. But my second point is is when Sailor says, like, there's a 100,000,000 Bitcoiners, it's like but it's just bullshit. Like, there's no we're still very small. Like, it's mostly a small vocal minority. Yeah. It's like, animals. You know, like, when you're, like, watching, like, National Geographic or something and they they show an animal, like, puff themselves up and look like a big bad beast. Like, we're a small town.
We're not even a city. Like, a city of a city of 200,000 people isn't a city. It's a large town Yep. Of actual Bitcoiners of of size.
[01:10:55] Matt Ahlborg:
This was a tweet that Marty put out a few years ago that really resonated with me. It said something about how, like, you know, 8% of colonials actually participated in the revolutionary war. So the vocal minorities really do change, history. So, you know,
[01:11:13] ODELL:
and I No. I agree. That's what we're doing. But it's I think it's important for people to ground themselves and realize, like, put it into perspective a little bit. Yeah. Anyway, Matt, this is a great chat. I'd love to have you on, like, I don't know, in, like, six to eight months and see where you're at and get some updates on AI. And, I mean, it's obviously a trend that I think everyone needs to keep an eye on. It'd be great to have you back. Do you have any final thoughts to the audience before we wrap?
[01:11:43] Matt Ahlborg:
Yeah. Don't have any final thoughts, but, thanks for listening. And and you can go to our, Noster end hub, and, it's probably posted somewhere around here. And, yeah, ask any questions you want. We are very much, building in the open as much as possible, and and, you know, user feedback and suggestions is is what drives the direction that we're going. So
[01:12:08] ODELL:
I love it. Yeah. Their website is ppq.ai. I'm gonna put links in the show notes to their master account and all the other relevant information. So just check the show notes if you wanna connect with Matt and the main account. Freaks, I see some comments in the YouTube chat. Yeah. The Nostra stream was down today. Kieran, who runs ZapStream, self hosts everything, and his router's getting DDoSed. I'm aware that it feels like Master isn't iterating quick enough. I still think it's amazing that that it's running and that it works to the extent that it does. We'll make those things better. But until then, we'll always be streaming no matter what. I'll make a way to get get the content to you guys even if we have to use YouTube, even if we have to use Twitch.
It is what it is. Next week, I got two shows lined up. One on Monday, we're gonna be talking about Bitcoin, with Ben Kaufman, great Bitcoin developer that works with Keeper Wallet right now. He used to work with Specter Wallet. And then on Friday with HuddleBot, who's mission focused on Nostr with Coracle and Flotilla, you can I'll I'll keep you guys posted on on Nostril with the actual schedule there. Primal.net/odell. Thank you all the freaks that continue to support the show. Thank you guys who joined us in the live chat, even though we weren't in our typical place, even though we weren't on the master live chat. Thank you for joining. You guys make the show unique. Huge shout out to Matt. Check out his awesome project. Give him feedback. It's very helpful at this point in their development.
Matt, thank you for joining.
[01:13:55] Matt Ahlborg:
Thank you so much for having me. Great conversation.
[01:13:58] ODELL:
Love it. Open Water Bitcoin is the best. Stay humble, StackSats.
Another all time high today topping a 118,000. The run up comes ahead of crypto week in Washington where congress is set to consider key legislation for the industry. Our next guest says the cryptocurrency could march to $1,000,000. Will Reeves is the CEO of Fold, the Bitcoin financial services platform, and he joins us now. I mean, Will, I'd love to sort of get your take on why we're seeing Bitcoin move higher. I know it doesn't always trade on fundamentals, but is anything that's happening in Washington potentially influencing the price of Bitcoin this week?
[00:00:32] Will Reeves:
You know, I think this is just inevitable. This is what happens when massive demand meets the finite supply of Bitcoin, and what we're seeing right now is the supply shock in motion. So I'd certainly say what we are seeing is the fundamentals of Bitcoin in play.
[00:00:48] CNBC Host 1:
And so you think that it will begin to start to move more on fundamentals and potentially even more so once it gets these sort of stamps of, hey. This is good stuff from Washington?
[00:00:58] Will Reeves:
I think the, stamps of approval have always been good. It enables new capital allocators to start to invest. And what we've seen recently, with the rise of Bitcoin treasury companies, the ETFs, is that entirely new classes of capital are able to invest in the asset. And what we're seeing is what happens when trillions of dollars start knocking on the door, the price tends to go up. And so the fundamentals have always been there. Investors have been waiting for their ability to get involved.
[00:01:30] CNBC Host 2:
What would it take for some of the other coins, altcoins, to kind of get the vote of approval or stamp of approval, if you will, in terms of Bitcoins? Because, I mean, if you think back to meme coins, the Doge's, the early stages of Bitcoin, it was very really the early adopters that really kind of formed that cohort that pushed it higher. Who will it take for us to see that transition into some of the other coins that are still branded as more speculative?
[00:01:56] Will Reeves:
You know, frankly, I think they're gonna remain speculative. Bitcoin dominance is at 60% in rising. I think what we're seeing is an inevitable trend towards the signal, the Apex asset. The issue is is that Bitcoin is the only credibly neutral digital asset that is in issuance. All the other tokens are still proving an extraordinarily speculative use case without the track record of Bitcoin. And so what we are seeing today is the continued march of Bitcoin dominating this entire market. And, I I am well, I'm waiting for the day when that changes, but ultimately, what we've seen is something that can't be reversed.
[00:03:15] ODELL:
Happy Bitcoin Monday, freaks. We got another still dispatch. The interactive live show focused on actual Bitcoin and Freedom Tech discussion. Looks like ZapStream is still under DDoS right now. I'm gonna keep an eye on that for us. But until then, I'm gonna remove that from our stream. Let's see what we got going here. Live producing going. We'll be streaming in the Twitch comments, the YouTube comments, and the x comments until ZapStream comes back online. Anyway, so dispatch the interactive live show focused on actionable Bitcoin and Freedom Tech discussion. We have another all time high Bitcoin rip going on Bitcoin maybe last night or yesterday, I guess, right before I fell asleep, past a hundred and twenty k.
Kieran from ZapStream said he just ordered a new router to deal with his DDoS. We have Matt Alborg here from PPQ, paper q dot a I. How's it going, Matt?
[00:04:20] Matt Ahlborg:
Hey. It's going well. Yeah. It's a pretty good day when Bitcoin's at an all time high to do a podcast. So
[00:04:27] ODELL:
Hopefully hopefully hopefully, this is the the rest of the summer. It's just constant all time high reps. I I'm here for it. So where should we get started? I mean, I think sorry. I'm reading troubleshooting. Your product's awesome. Let's do a quick to the freaks that don't know what ppq.ai is, what is it? Why does it exist? Why should people care?
[00:05:01] Matt Ahlborg:
Yeah. So PPQ, it's a website where you can get access to all of the latest and greatest AI models and tools in one place, and it's without subscription. It's on a pay per use basis, and it accepts all well, many cryptos, but we do have kind of a Bitcoin first, preference, 5% bonus for all lightning deposits in particular. But the basic the basic idea is that, you can come to our website. You can deposit as little as 10¢ worth of Bitcoin. You can try the product out. You're not committing to a monthly or annual subscription or anything like that. And, yeah, what we've what we've seen in the last year is that there is not one AI company that's releasing all of the best models.
Yeah. There are, half a dozen, awesome companies releasing really awesome stuff. And so it really pushes forward the idea that, like, the better the better, interfaces with AI, it's it's those that are plugged into all of the top company models.
[00:06:10] ODELL:
Yeah. I mean, I guess there's a bunch of things to unpack here, but trying to stay on top of, the AI curve or the AI acceleration phase where it just feels like new stuff is dropping every day. You kinda have you're you're you're kind of held back by as an individual that wants to use these tools, you're held back for two reasons. You're held back for one that you have a bunch of different companies that are iterating and innovating and shipping new stuff all the time. So in practice, you end up needing to have accounts everywhere. And then the second piece is it's very hard to use them with any semblance of privacy whatsoever.
They almost all require credit cards as a payment method, And then some of them even do other identity verification. Like, they try and make it even more difficult to use privately, let alone just the standard credit card billing stuff. So PPQ is really cool because you basically have all of the top models in one place, and then you can just easily pay with lightning. Probably. Awesome. And I I think it's I mean, it's pretty unique. Right? Is there anything else that kinda even competes with you?
[00:07:31] Matt Ahlborg:
Yeah. There are. There's one website in particular that I would consider to be our competitor, Nano GPT. They've done basically the same thing that we've done except they love the cryptocurrency Nano.
[00:07:45] ODELL:
I forgot that I forgot that shitcoin existed.
[00:07:48] Matt Ahlborg:
Yeah. But aside from that, aside from their, misguided shitcoining, they've done a great job executing. And so, we we are trying to answer in kind and and, we, we recently acquired a $500,000 investment from a couple of the, you know, popular Bitcoin VCs. And so we're scaling up now on the engineering side so that way we can really, really ship a lot of great stuff, in in the coming months. Yep.
[00:08:21] ODELL:
So let's so, I mean, if you look on if you look on your website, I mean, you can you can see what, different models you offer. But, you know, you have the OpenAI models there. You have the Anthropic models. You have XAI with Grok. You have the Google Gemini stuff, Mistral, Perplexity, Meta. How hard is it to actually integrate with their APIs? Like, do they make that difficult for you? Like, what is that process like to actually?
[00:08:54] Matt Ahlborg:
It it is, I would say, medium difficulty. We in many cases, with the top top models, we like to directly integrate with the provider. So we're getting our OpenAI models directly from OpenAI. We're getting our anthropic directly from anthropic. But with a lot of the, kind of second tier models and such, like, it would be a lot of overhead to set up direct integrations with all of these dozens of AI companies. So then we plug into, like, an aggregator that there are a bunch of companies coming up that are specializing in, like, providing all these models, in one place, which is kinda similar to what we do, as well. So, but I think our focus is that we want to make our platform, really user friendly where a lot of these kind of other aggregators, they're more at the they're more developer facing or even enterprise facing, whereas we're we're kind of trying to get to the, AI curious newbies or maybe crypto native person who's a little bit technologically minded, but maybe is is learning AI for the first time. That's the kind of users that we are capturing. Yeah.
[00:10:06] ODELL:
You're, like, more of the, like, the consumer focused front end while they're kinda offering, like, more of a b two b type of service.
[00:10:14] Matt Ahlborg:
Exactly. Yeah. Mhmm.
[00:10:17] ODELL:
And I does with Grok, is that, like, a direct API call? Like, they just open up an API for you? I assume you have to pay for it. But Yeah. So Grok is a tricky one. You can get a direct API access with XAI.
[00:10:31] Matt Ahlborg:
However, we're getting our Groc from one of the aggregators. When they released Groc four, like, a week ago, they kinda released it to a limited pool of of companies. And so for that reason, we are we're not getting it directly. But, we do have, you know, an XAI account. And, when they make it available, directly, we will probably plug in that way.
[00:10:54] ODELL:
And then what about, like, the open source models? Are you self hosting any like, are you hosting open source models, or are you, is it you're purely an API front end effectively? Yeah. We're we're basically
[00:11:05] Matt Ahlborg:
purely an API front end. Self hosting models is actually a lot of work, and, it's it's a lot of resources as well. Even, like Open Secret or or Maple AI, those guys, which I really like them, love what they're doing, and I've had several conversations with them. But the model that they're self hosting is it's still not the fully beefed up version
[00:11:32] ODELL:
of of, DeepSeq or or whatever open source model because those literally cost LLAMA they have LLAMA and DeepSeq Seek right now as their two supported models.
[00:11:42] Matt Ahlborg:
Yes. And but those particular instances of those models, they're not the beefed up version of the model. They're kind of like the middle of the road model. And and so because it costs a lot of money to host, one of those, the beefed up versions, like, I think on the order of $1,020,000 dollars a month, something like that. So you need a lot of demand to really justify hosting the fully beefed up open source version, And it's a lot of engineering to put into place and make sure that it's all, like, load balanced correctly and and all that stuff. So we've, avoided that stuff.
Another reason I would say is it's a lot of work, but another reason is the demand is not there, for a lot of the open source ones yet. They don't have the name brand recognition. There's still you know, when DeepSeq came out, a few months back, everybody everybody thought, like, oh my goodness. It's like, you know, it's just as good as as OpenAI and Anthropic. And then, like, as time passed, people figured out that, no. This model is not as good as the brand name models yet. And and there's still a lot of work to to to make up. And and as an end user of of AI, like, if I'm a developer every day, I would say a core user that we have is developers, using our API.
If I if I want to, do my job in the best way possible, I would rather use the absolute best model. There's no reason for me to use a model that's, like, 90% as good as something that that I you know, it's just not worth my time. So, yeah, so the open source models are not there yet. There's a really exciting one, that was released a few days ago. It's called Kimmy from Moonshot AI. It's it's like another Chinese open source model. Right? Yes. Yes. And it's getting a lot of buzz, and we're adding that, that one either tonight or tomorrow morning. And, so we'll see. It seems to have good reviews, but really, truly, the the test of time is really important for a lot of these models because, it's very easy to get your new model to pass all of the, like, the benchmarks, to kind of game it to to to do really well on benchmark testing. But then when it actually comes to, like, actual usage, it's it's a different story a lot of the time. So
[00:14:04] ODELL:
Right. I mean, the way I kinda think about it is, like, the open source models are usable right now for kind of, like, the average person chat interface type of scenario where you're asking for, like, populations of different countries and different stuff like that. But if you wanna do, like, really strong tasks, hard, aggressive tasks, whether that's, you know, generative art or vibe coding or something like that, there's a real big gap between the proprietary models and the Yes. And the open source ones. Would you agree on that? Totally. Yeah. And it it's but that that was not always the case. Like, even a year ago, when we've jumped from GPT 3.5
[00:14:48] Matt Ahlborg:
to GPT four turbo, like, GPT four turbo, I believe, was, like, the first model that really started doing well at answering those everyday questions, like population or whatever country. Like, 3.5 did the job, but Ford turbo really, like, was the first model that did well at all that, even the basic everyday questions. And so I think we are already at the point where, like, the basic everyday questions are free now. And and we've we've already moved past that, and now the expensive models are the ones that can do a lot of the fancy stuff. Yeah.
[00:15:26] ODELL:
Yeah. That makes sense to me. So, I mean, as an end user, like, obviously I mean, maybe I'm not the average end user, but I I personally have, you know, long term concerns about the viability of of open source models versus proprietary models. Like, I don't want us to go down a dystopian hole of, proprietary data surveillance heavy models. I mean, I think it puts a lot of disadvantages to the individual. It gives a lot of advantage to the corporations, to the governments that they are they collude with or get pressured by. But from a just an actual individual everyday point of view, the biggest issue is on the privacy side.
So for right now, like, the way you have PPQ AI set up, what is that trust model? Does that look like a is that like a VPN trust model? Like, are you basically sitting in between me and my, prompts versus with with the the AI model on the other side? Like, how does that look? How should a user think about the privacy model of your product?
[00:16:35] Matt Ahlborg:
Yeah. So, with PPQ, when a user comes and they enter in a question and then they hit the enter key, it that query, along with the content of that query, goes to our back end, and you have to trust that we are not logging, the content of your query on the back end. And we in our terms of service, we are telling our users that, like, we only log metadata, like, how many tokens the question was, how much it cost, all that stuff, and we're not logging the content, but we're not proving that. And so but what happens after it hits our back end is we send it on to the AI provider, and then the a AI provider returns the answer. The answer also goes through our back end including the content, And we're promising that we're not logging it, and we're sending it back to the front end.
That's the model right now, so you do have to trust us that we're not logging this stuff. I do think there is potential for, like, an architecture that could improve that where we could make better guarantees for our users. But, honestly, like, I did my best to get this product up and running as fast as possible. And I think that's, like, a very classic trade off between, like, a lot of what Bitcoiners want is, like, they always want you to to go absolutely perfectly in the direction of privacy. And if you're and if you're not all the way there, then you're you suck. And and, so I think that's actually one of my things that I'm good at is that I'm good at being practical and just shipping stuff that works.
And, I do very much appreciate privacy, and I wanna build that in a lot better over time. But that's where we're at right now. I wanna give people something that works.
[00:18:29] ODELL:
Yeah. I mean, I think it's, like, a perfect is the enemy of good classic example. And so it's basically the VPN trust model, more or less. So we we are you trust Mullvad or Proton not to log. Yeah. Yeah. Yeah. And so trust the relationship. But what about, presumably, if I'm doing a bunch of follow on prompts, the AI provider knows that's all from the same person, I assume. Right? I don't think so. So we to the AI provider, we are
[00:19:01] Matt Ahlborg:
one customer to them. So Okay. All of the queries from all of our users are being bundled into one, like so, now with certain data analysis and stuff, they theoretically could probably figure out, like, okay. These queries here are associated with They're very similar. Right? Yeah. Yeah. So they could do that, but I don't know. It I I But they're not getting, like, user x or user y or user ID code. User PPQ is what they're seeing. Yeah.
[00:19:33] ODELL:
And and so especially if it's if it's two different tasks that I'm doing. If I do if I, like, vibe code an app and then I go and just generate photos or something in a new chat window on PPQ, there's there's it's a pretty good privacy model there if if we trust you. I would say so. Yes. From the AI provider. Yep. Okay. That makes sense to me. The other cool aspect is is one of the few ways to actually use this stuff on, like, a pay per use basis, which, I mean, is hence the name pay per queue. How do you think about that? Like, are your costs like, it seems like the industry has more or less standardized, at least for consumer grade stuff, is standardized on, like, a monthly subscription model, where you just pay and you might you probably don't use it all or some power users are just destroying and destroying them and using way too much.
From an end user point of view, I love the pay per use model because, honestly, I'm I'm not a power user. I'm not using it all the time. And I like to know that I just when I use it, I pay for it, and when I don't, I don't. But how does that work for you? Are you pay are your bills,
[00:20:50] Matt Ahlborg:
usage based, or, like, how does that work? Yeah. My bills are usage based. PPQs are. We're we're when we plug into an API, we're not on any sort of, like, AI subscriptions with these companies. But, yeah, this is a really, really big question, actually. The the subscription model versus pay per use, how did we end up here? I think most people I have a lot of thoughts on this, but I'm not it's, like, not perfectly crystallized in my mind. So I'm probably just gonna ramble on here. But, with subscriptions, I think that, let's see. How do I start this?
So most people think that we arrived at this subscription models because, like, some MBA meeting at at all these companies agrees that it's like, oh, we should find a way to charge users for stuff that they don't use That's really good business model. Right? Like and then let's make them forget that they have a The gym memberships. Yeah. Exactly. Yeah. So I think that most people believe that that's how we ended up here because it's like a a nice way to steal money from your users, which I think there's some truth to that. But I think if you, I I worked at Bitrefill for five years, and one of my projects there was to kind of research the idea of getting the Bitrefill Visa card or Mastercard or whatever. And so I learned a lot about how the credit card companies work and how their fee structures work.
And, actually, I believe that we've arrived industry wide or, you know, global payments global payments, we've arrived at subscriptions not because of, it's it's a good way to steal money from users, but because the fees that Visa and Mastercard charge on microtransactions make microtransactions untenable in most cases. And, also, the fact that Visa and Mastercard have from their inception, they built in this chargeback mechanism, it makes it so that, anytime you do a payment on on these networks, there's a risk that it could be disputed and that it would be charged back. And for whatever reason, it means the what that amounts to is that small transaction numbers or or small volume numbers, are not, good for they they they become more expensive. So when you swipe a card, it's like 30¢ plus a percentage, typically. And so what that means is that a larger swipe is better, and that kind of rules out a lot of micro transactions.
But with crypto, it is there is no chargeback risk. It everything is instantly settled, and so we are now back at a place where pay per views actually starts to make sense. And and when you when you you know, there's all sorts of justifications for why we should have subscriptions, like, from a business point of view, like, you know, stealing money from users. And, like, we can also say, like, oh, now we have this recurring revenue thing, and that's good for you can you can, tell the the, the shareholders that we have our current Yeah. All this bullshit. But it's it's at the end of the day, what is a subscription? It's you're charging every user the same amount no matter their usage, and that just economically makes no sense at all.
You should be charging users based on their usage. And so I think that we are at the beginning of the end of subscriptions. I think that companies like PayPerQ, are going to start eating away at subscriptions. And the beautiful thing is is that you said you're a casual user. Right? So the the people who make the most money for these subscription providers are now gonna come to pay per queue because it costs them $2 a month instead of $20 a month. Right? And so who's left over in the subscriptions? It's the power. And I'm also I'm also, like, a casual user that wants
[00:24:52] ODELL:
the best model. Right? Like, I I don't I don't use it all the time, but I also I I would pay the best model. Use it. Yeah. Yeah.
[00:25:00] Matt Ahlborg:
Exactly. So that that actually helps us out against the freemium providers. Like, you can use ChatCBT for free, but they don't give you the document upload. They don't let you make images. They give you, like, a really shitty context window, all that stuff. So, yeah, PaperQ is a perfect place where you can actually get the high quality version of the model, the state of the art, but you're not paying a subscription. And so, yeah, it's it's great. Right. So I kinda cut you off, but what you were saying is
[00:25:30] ODELL:
you're kind of forcing the hand of the other providers to move to this model because the only ones who'd be he'll be left on the subscription model are the ones that they're not making money off of because they're just, oh, like, using the hell out of it. That's exactly where this is gonna end up, I I think. And I think AI is kind of the perfect,
[00:25:48] Matt Ahlborg:
way to to to really start forcing your hand because, like, Spotify subscription, I don't think that's going away anytime soon. Netflix subscription, that requires, like, massive licensing fees with all these big media conglomerates and stuff. Right? But this AI, I think, is it's the perfect way to really start eating away at the subscriptions.
[00:26:09] ODELL:
I mean, this is the first time that we've seen, like or maybe maybe I'm missing something, but it feels like one of the first times we've seen actual significant, like, on the Internet pay per use cost. Right? Like, real physical cost. Like, it has major computational cost every time you send them these prompts.
[00:26:30] Matt Ahlborg:
Yes. Yes. And that's kind of what makes it the perfect, like, the perfect, start to this end of subscriptions is that, yeah, there are, all sorts of views. Some are very casual. Some are power users. So the subscription models don't make sense anymore. And you can see that these subscription providers are already kind of, like they don't know what to do. Now they have, like, higher tier subscriptions. And even with, like, Cursor, if you've ever heard of Cursor, it's like this, it's like this, development tool that a lot of people are using. They have a 20 $20 subscription, then they have a $200 subscription.
The $20 subscription, they're already kind of, like, taking really big behind the scenes shortcuts. I I just recently hired a CTO, and he came in. And he and and I'm trying to he's he's, you know, he's got fifteen years of development experience. And this is a little bit of problem with with, like, some of the senior devs as you can't get them into the AI because they already know a lot of it. And, like, AI is really good for people like me who are not, seasoned developers. I'm kinda just I learned coding in my late twenties and, you know, I was never really a super full time person.
So it's great for me, but for these senior developers, like, they really have to be convinced on the value of AI. Well, I'm trying to convince him, like, you need to start using AI more because, like, this is our core user. You need to understand our product, blah blah blah. And the first week or two, he's like, hey. You know, I'm I'm doing my best, but, like but what what we figured out is that his $20 cursor subscription was, like, giving him really shit results. And that's he was forming his opinion of AI based on the shitty experience that he was having using $20 Cursor.
And then I was like, oh, that's crazy. Like, try using Cline. Cline is an open source version of Cursor that works on a pay per use model. And when it's pay per use, you're getting the model that you're paying for. It's giving you the full quality of the model. There's no secret, shortcuts behind the scenes. What we think is going on with Cursor is they're like, to save money, they're giving you lesser models at certain times, and and they're they're shrinking the context window. So it's not reading all of your files. It's like throttling. It's throttling you. Yes. And so when he switched over to Cline on paper use, he's like, holy shit. This thing is doing so much better than Cursor.
And so you you can already see there are really a lot of serious cracks forming in these subscription models.
[00:29:09] ODELL:
So with with your business model specifically, are you just charging, like, a flat rate margin on top of every usage?
[00:29:18] Matt Ahlborg:
That's what we're doing. I would like to transition to kind of a, kind of like a logarithmic pricing where, like, if you're a casual user, we're gonna charge you a higher margin. But if you use us a bunch over a thirty day window,
[00:29:33] ODELL:
your margin's gonna lower. Like a Bitcoin exchange.
[00:29:36] Matt Ahlborg:
Yeah. Exactly. There you go. Same thing. So I would like to implement that. That way, we win both user both types of users.
[00:29:44] ODELL:
But right now, it's just a fixed fee on top of that. That's pretty cool. I mean, you're just profitable and sustainable from the get go.
[00:29:50] Matt Ahlborg:
Yeah. Exactly. And it's and it's simple. That's the other thing with subscription. I told you, don't get me started on subscription. Every single AI website on the Internet right now, it has the same onboarding flow. You go to their website. They give you this watered down narrow version of what they actually offer, or maybe they'll give you a seven day trial or whatever. And then they try to get you on that subscription. Right? So there's so much, design thought that goes into converting a user to become a subscriber. Right? Because that's a big commitment. Right? To get somebody $20 subscription, that's $240 a year. So to get people to get a user to cross that, you have to design things so well, to get them to to to sign up for the subscription. On PPQ, like, we don't have to think about that stuff at all. You come. You make a deposit. We charge you whatever the price is, and and boom. You like it, good. If you don't like it, then you didn't lose anything. You're not committed. You know? Is is your margin is, like, the fee you charge on top public? Or it's not. But, I mean, we charge 40% on the UI.
Okay. And it still adds up to, like, you know, you're doing advanced queries and it's 2 pennies per query type thing. Yeah. It still feels pretty cheap. It's very cheap. Especially if you're not, if if you're not, like, an insane user. Yeah. Our median user is spending, like, a buck 50 a month. But on our API, we have a much lower margin, because the API users are using us, like, a lot more, and they're using us with the the IDEs and stuff. So, our I think our margin on the API is, like, 18, but we're gonna go even lower soon. Maybe 10%, something like that. So,
[00:31:47] ODELL:
what does an API user look like? Are they using something like Goose on the front end and they're connecting it? And, like, how does how are they using that product?
[00:31:56] Matt Ahlborg:
So they get so we offer an API key. Okay. What sort of PPQ, you get a little a p a PPQ API key. Our API key is what's considered OpenAI compatible, which means you can plug our API key in literally dozens, hundreds now of third party tools. And so, yes, Goose, Klein, Roo, the tons and tons and tons Zed.
[00:32:24] ODELL:
So they're just using a different front end with your API, basically. Yeah. They're they're using
[00:32:29] Matt Ahlborg:
I I would say, yes. They're they're using a different front end, but most of the time, it's an IDE, an integrated development environment. So it's like, yeah, they're they're looking at code. They're making calls. They're making many calls in succession as it reads their code files and stuff like that. Yeah. Got it.
[00:32:49] ODELL:
So, I mean, you said this in the beginning, and I wanna pull back to it. I mean, the biggest friction presumably with your product right now, and the death of subscription models is, like, people actually using Bitcoin, to make payments. Obviously, someone like me, I go to the website. I wanna test it out. It's like I just I have, like, 15 different Bitcoin wallets I could pull up to to pay it. But, like, the average person that doesn't have Bitcoin, maybe that's not the case. Maybe they get hit with some friction there. You said you mentioned that Lightning specifically you're basically giving a 5% discount to people. If they pay with Lightning, you'd rather than pay with Lightning. Obviously, I mean, our listeners, I think, would agree with me that Lightning is, like, the single best payment method for this type of thing. First of all, we get great privacy. But just from the UX point of view, it's just like point, click, go.
You have a bit refill background, but you also have a Useful Tulips background. I mean, that's what I knew you from. Freaks, if you didn't put two and two together, this is Matt Alberg from Useful Tulips, the OG p two p trading dashboard that was, like, tracking. I think it does it still exist?
[00:34:02] Matt Ahlborg:
Yep. No. It's not Not updated? Anymore. I think the website
[00:34:07] ODELL:
it was tracking all the p two p sales data from, like, all the different countries in the world, had great charts and whatnot. But you also have this bit refill background. And one of I've had Sergei on the show before, and I also, you know, have become relatively close with him. And one of the most fascinating things about him and his relationship with Bitrefill is the payment breakdown. So I'm kinda curious, like, what do you see payment wise in terms of Bitcoin on chain, Lightning, the other cryptos, currencies you support? Like, how does that look in practice?
[00:34:41] Matt Ahlborg:
Yeah. So Lightning is number one. It's just over majority. Number two, let's see. If you if you can make a guess on our number two. Is it Tether?
[00:34:54] ODELL:
Nope. Oh, is it Monero?
[00:34:57] Matt Ahlborg:
Monero. Yeah. So number two is Monero on chain Bitcoin. Yes. On chain Bitcoin is pretty small.
[00:35:04] ODELL:
And just because if they already have on chain Bitcoin, they'll probably just pay you in lightning, I guess.
[00:35:08] Matt Ahlborg:
Yeah. Also, our on chain Bitcoin has been kind of, like, finicky. So it's hard to give, like, statistical significance sometimes because, like, our payment methods were not always a 100% up. Okay. And it goes with our stable coins. Like, we had a really shitty stable coin payment processor with a lot of errors. So, like, the tether is probably a little lower than it should be. Got it. But coming from Bitrefill, I know exactly the payment popularity over there, and that's why I, chose to add the payment methods I did on PBQ because I'm already very familiar with what, like, the crypto payment landscape looks like. That's why I know that actually Litecoin is is actually pretty important payment coin to add.
[00:35:53] ODELL:
I think Do you see a lot of Litecoin payments?
[00:35:56] Matt Ahlborg:
It's lower than I thought it would be, but I think it's still around, like, 15%, something like that. Interesting. Way higher than I would expect. On Bitrefill, it's pretty high. Like, still not insanely high, but, yeah, 15% plus, something like that. Yeah. There there's there's a crazy thing with, like, the older the coin and and the more, you know, the older the coin, the more exchanges that it's listed on, the more wallets that support it. These coins gain massive network effects as well even if they're just the carbon copy of what Bitcoin is.
And it is I think it's very fascinating how Litecoin has kind of hung in there with without really any innovation or any nobody's ever been cheerleading Litecoin to a large degree. It's just kind of like this zombie project of sorts that just keeps kicking because it's been around so long. And they have actually improved Litecoin. They have, like, Mweb. It's like Mimble, Wimble, something or other. It's a privacy hack. They recently added the Litecoin. I don't know if it's good or not, but, yeah. I I I don't hate on the Litecoiners as much as I do some others, because I do think they have a lot of shared ideals with and they they a lot of them come from back in the day when the the early days of, you know, 2013, 2014 when when we were all learning about Bitcoin together and that type of thing. So I'm as the original shitcoin.
[00:37:32] ODELL:
I, so I'm looking here and, so I see you offer you offer lightning, Bitcoin, Monero, liquid Bitcoin
[00:37:43] Matt Ahlborg:
Yep.
[00:37:45] ODELL:
USDC, and USDT. And you also offer debit credit card. Do a lot of people do debit credit card?
[00:37:51] Matt Ahlborg:
It's it's pretty low, but but not nothing. And I think it's one of those that could really pop off if we get, like, the right tweet from the right person or the right I don't know. Because everybody has most people in the West have credit cards. And Yeah. As you said the standard. Are the people with a lightning wallet, around aside from us, Bitcoiners.
[00:38:15] ODELL:
So do you have a lot of fraud that you have to deal with? Because this is kind of a classic chargeback, part of the story. Right? Like, I'm I'm interesting thing.
[00:38:24] Matt Ahlborg:
As I said, the credit card payments are not that high, but we have had several instances where Stripe will send me an email and say, like, hey, this payment that happened a month ago, we might have to, like, take your money back. And it's like it it just proves to me, like, that this this isn't you know, it's it's it's Right. It's bullshit, like, how they can just steal your money back after an entire month, and it just kind of adds fuel to the fire that crypto and and, Bitcoin Lightning are the way to go in terms of, accepting payments because I've never had a single chargeback on Lightning. No. It's impossible. I mean,
[00:39:04] ODELL:
it's a it's a classic chargeback business risk for this particular use case because you're providing compute. Like, the it's gone. Like, you've already provided it. There's no you can't claw back the goods. You can't wait to see if there's a chargeback or not. It's weird, though, because the few times that it's happened, I'll go and I'll look at the user that that or the the credit ID that loaded.
[00:39:29] Matt Ahlborg:
And I don't think any of the times had they ever spent their entire balance. So it's I don't know what's going on there. But I mean, maybe you haven't actually gotten hit with a fraud, like, a fraud wave yet. You're just getting
[00:39:42] ODELL:
users that had regret for signing up for the service and just like issue. There was just some machine learning red flag that was the most positive or something. Yeah. I mean, that could become a major problem for you if you actually get scale on the credit card side. Yeah. It could. I understand. It's a trade off.
[00:40:01] Matt Ahlborg:
On the day that that happens, I'll be happy to tweet about it because this is a yeah. I think chargebacks are a huge problem. Like, I mean, I guess they have some some advantages, and there are reasons they added them. But I think if, like, if if if you went back to the nineteen seventies when they initiated this thing, like, I think Visa, Mastercard, they had no idea the cost that chargebacks were going to introduce into their systems, especially in the age of the Internet where, like, you know, fucking fraud is everywhere. I think they would really rethink, the fact that, you know, they would make payments final. And and that's why I think a lot of the the the crypto payments happening that are instantly settled, that's it's a huge advantage.
And I think that's gonna be in the same way that I'm clawing away at subscriptions, I think crypto is clawing away at Visa, Mastercard because chargebacks are like a really big thorn in their side.
[00:41:01] ODELL:
Yeah. I mean, it's actually the merchant. I also I mean, I just think it's a better payment experience from the user point of view,
[00:41:07] Matt Ahlborg:
just from, like, a UX point of view. Yeah. Even when you're a user and you make a credit card payment and it goes through and both you and the business are happy, there's still a fee involved that reflects the amount of fraud that's happening across the network. Like, when you that 3% that we all hear about, a big chunk of that 3% is because of all this fraud mitigation,
[00:41:31] ODELL:
that these companies have to put up with. So Yeah. They're social they're effectively socializing the I mean, it's a it's a they're socializing the cost. They're they're spreading it out. They're doing a risk reward adjustment to make profit. I see Gary Krause in the comments, rider or die freak, asking about eCash. What are your thoughts on that front? I assume he's talking about using eCash actually internally to try and add additional privacy to the system.
[00:42:01] Matt Ahlborg:
Yeah. So I'm not an eCash expert, but I've been learning a bit more. A lot of the ecash aficionados have come to me and asked for me to add this. There is routes that are out there, which I'm sure some some users are also familiar with, which uses the ecash tokens. So what you get when you add ecash, as far as I understand, is it basically makes it so that PPQ so right now, when a user makes deposits $10, they get a credit ID associated with them. So I can actually link If I wanted to, I could link all of the users' queries together, and I could see, like, are trusting you. Yeah. And I could I could say this person who ran five queries on Tuesday and then four queries on Wednesday and whatever. But with the eCash, it becomes very it it becomes anonymized, basically, where you can't even tell, you know, who who's even among your credit IDs, which credit ID is making that query. So there is a price to gain by adding that.
The problem with ECash is is the same problem as Lightning, only it's magnified by 10 x is that, like, not many people hold ECash to pay with it. Right? And so, like, I even tried to do Routester myself, and I was like, okay. I wanna go to Routester. I wanna, like, try this thing out. Oh, I don't have ecash tokens. Where the hell do I get ecash? No. But I think I think the point is
[00:43:28] ODELL:
for you to replace credits internally with ecash. So, like, I pay with lightning or credit card or Monero or whatever, and then you give me PPQ ecash tokens, right, that you manage.
[00:43:41] Matt Ahlborg:
I'm issuing them to the user. Yeah. That is another way to do it as well.
[00:43:46] ODELL:
And then that way and that way, then I would have the bearer tokens. I would have the ecash bearer tokens. Right? And then it disconnects me to a degree. It reduces trust in you. Yes. That is true. Trust model. Yeah.
[00:43:59] Matt Ahlborg:
Yeah. That now that I think about it, that is, there is a distinguished there's different ways you could use eCash, and the better way is, yeah, you don't have users bringing you eCash tokens Don't do that. Giving them eCash tokens when they make a lightning deposit.
[00:44:15] ODELL:
And and that improves, like so the trade off there is additional complexity on your side, particularly when you first set it up because you said you're not an eCash expert where there's about 10. So there's, like, not that many. I will say most of them are, you know, vibe coding aficionados. So, like, I'm sure a bunch of them would be willing to help out, like, specifically their leader, Cali.
[00:44:40] Matt Ahlborg:
It's it had Cali reach out to me a couple of times. And so, yes, it is something that that I I would, I would look into and and and and, you know, it it's just a matter of, like, is it gonna be next month, or, maybe two, three months from now? I I don't know. Fair enough.
[00:45:00] ODELL:
Yeah. Yeah. I mean, I I mean, I think the argument isn't really this isn't a, Bitcoiners pushing for perfect kind of thing. I mean, I think long term, you're in probably in a better position if you don't have the liability of having so much information on users. Right? Like, if PPQ is successful at scale and you can actually connect the dots on users, you're gonna get tons of requests and, you'll have tons of information liability, which we've seen with every VPN provider. Like, I can't imagine I would that's not a business I would ever wanna be in. It's like the hosted VPN provider business. And so if you can reduce some of that data liability, data exposure, it's probably a big win for you long term if you see scale.
[00:45:48] Matt Ahlborg:
That, yeah, that makes a lot of sense, actually. Yeah. I I I will definitely think about that. I don't know if the answer, though, if if if a regular comes to you and you say, like,
[00:46:01] ODELL:
sorry. I don't have the information that you're looking for. Like, I think most of the time, their knee jerk reaction is like, well, since you don't have it at all, then we're shutting you down. Well, no. They I mean, I don't think they could shut you down. I think they could try and pass additional regulations. I mean, it just adds it adds some level of friction. I mean, we see this with stuff like signal, right, where signal does respond to government requests. Just the data they hand over is, like, basically nothing. They they, like, hand over a phone number. They can hand over an I IP addresses that connect to their servers. But that's basically it. They don't have metadata, and they don't have the message contents.
Yeah. And so what happens what happens is is for the high value targets, governments just break into their phone or their computer and then get the signal messages that way, which they could do for this too. And so, presumably, it kinda plays out the same way where they just go that route. They don't give you too much trouble. And if they wanna add regulation on top, they can add regulation. The advantage of the ECash model, at least I mean, Cali's become a pretty good friend of mine. Is that, like, he even though it is adds a level of complexity, it is a use case that he's thought a lot about. And so they do basically have, like, an SDK, and it's relatively easy to implement.
Yeah. I mean, his dream is, like, Amazon gift cards being e cash, like, their own internal e cash, so you have some level of privacy
[00:47:31] Matt Ahlborg:
attached to it. Yeah. The one last, I do I do really, take point with what you're saying about, having too much information and and that being becoming a liability. The one last pushback I'll I'll make is that it's actually nice to have some level of metadata on users because to know what countries my users are coming from helps me in terms of marketing and business development, to know what you know, if if they if a user comes like, one question I have now is, like, how much value is adding all these media models giving me? Media models like image generation generation, like, if I can pull a new user in because they can do some cool image model, but then I can transition them to to using me on a daily basis for text queries or whatever, then it's worth it to add more image and video models. Right? So if I can actually see that about a particular credit ID where they come in, they make a they make a couple of image queries, and then I can see that they transition. That helps me from a business development standpoint.
I get it, though. It's not perfectly private. And so, you know, there's there's a gray area between, like
[00:48:51] ODELL:
but it's It's trade offs. It's trade offs all the way down. Yeah. Yeah. It's hard to build a product without having With knowing nothing about your users. Yeah. User I've seen that firsthand, you know, particularly on the ten thirty one side and some of the businesses we've invested in. It makes it really difficult. You're just throwing estimates around.
[00:49:11] Matt Ahlborg:
And and I don't even in most cases, I don't even get the email address of my users. So, like, I can't email them and say, like, hey. Why did you leave our platform? Like, what what did we do badly? Like, I have this little customer chat widget that is actually really awesome because that is my primary way that I'm talking to customers. But aside from that, like, yeah, I don't even have emails most of the time. And so it's really hard to to if I didn't have that customer chat thing, there's literally hundreds of bugs that users have reported to me, using that. And and it and it shows that, like, you have to have some level of connection with your users, in in, you know, I yeah. So it's really important.
[00:49:58] ODELL:
Yeah. Fair enough. Yeah. I mean, you could also you could just make it optional if you wanted to. You could give people a you could give people a discount too if they if they stay in the, you know, less private, non e cash model. No matter what, you still have the IP addresses of the you still get some data. Right? Like, it doesn't
[00:50:23] Matt Ahlborg:
Honestly, I haven't thought about exactly what data I would be losing that hard, but that was a question that I posed to Now the big thing you would lose is user behavior data, which is very helpful in product iteration.
[00:50:37] ODELL:
The other thing is you can iterate the product without it be and then you could turn on the private privacy preserving. And maybe maybe three months is too soon. Maybe it's maybe it's eighteen months, twenty four months. When you have a general idea of what the average user flow looks like, you're not losing as much. You've already iterated on the product more.
[00:50:57] Matt Ahlborg:
Yeah. There's Yeah. Always iteration to do, though, but, I I mean, I hear you. Yeah. Yeah.
[00:51:03] ODELL:
I just, is probably my single biggest concern with the AI acceleration is the disadvantage that privacy like, for all of digital humanity or for all of our lives, we're probably relatively similar in age. If you wanted to have privacy in the digital age and have freedom in the digital age, you were at a disadvantage. And, one of the cool things about Bitcoin, to be honest, is is that it actually flips the script on that. Someone who uses Bitcoin privately and in freedom oriented way actually has a lot of advantages over someone who just owns ETF or owns MSTR or something like that. So it's like a natural advantage that you have actually putting in the work and and doing that. But on the AI side, it really accelerates it. It's like, if if I'm competing with someone in the job market and they have access to chat GPT and they're using it in the least private way possible and feeding in all of their private information into it, that that model and that agent is gonna become so much more powerful to them. They're just gonna give them almost superpowers compared to the guy who's using the freedom focused one.
Yeah. And it becomes this massive data collection hellhole. Like, it's the first real model that incentivizes data collection on the Internet was the ad sales. Right? It was like, if you could track someone and you could sell their data to advertisers, you could monetize them better. It was like the your data cattle. Like, monetize the cattle as much as possible. This is the second major model. This is like the the second time in in digital human history where you where corporations have, like, a huge, huge incentive to to track people and collect as much data as possible and then train their models off of it, make them better, do all this other stuff.
So it's just something I think about a lot. I'm not pretending like there's a perfect way to do it. But
[00:52:57] Matt Ahlborg:
I mean, I think about it a lot too, and I haven't seen reasons yet to be super optimistic about it. Right. Which sucks. I think maybe maybe the maybe it is the fact that, like, you know, we did have DeepSeq, which came along, and it was a very Right. Good model. And so if we do continue to see, these open source models that are approaching the capability of of the proprietary models, then maybe there is, some hope in there. Okay. Yeah. I I think maybe maybe yeah. I'm just really theorizing here, but, like, there could be major incidents coming up in the future where, like, you know, we we have all sorts of data leaks nowadays. Right? Yeah. But, I mean, a lot of the time, I've never been impacted by a data leak myself. I'm sure were in the TransUnion,
[00:53:55] ODELL:
Social Security number thing where, like, every adult was.
[00:53:59] Matt Ahlborg:
I'm sure I was, but I guess my point is is I never felt the impact of it. You know? Like, I never my identity was never stolen. Like, I know it's common, but what I'm trying to say is, like, maybe there will be an incident where, like, you know, a large number of people are substantially impacted, and then that's the moment that governments actually step in and enforce privacy or start to enforce privacy. But maybe I'm just being really crazy and and, that's never gonna happen. I don't know. But I'm just trying to think about how it could come about
[00:54:31] ODELL:
when what are the possible inflection points? I mean, I'll say the bull case in my mind is that the data leaks are gonna get worse and worse. And but rather than government stepping in, even though they should, I think it's a national security issue. But rather than government stepping in, I think, you know, you get you get a handful of new users that seek out the better methods. Now the question is, and you don't get most of them. Right? Like, we just saw, what the main DNA provider like, they got, like, sold to, like, the Chinese. They got a lot of they got everyone's DNA that provided it to, 21 and me. And, like, most people didn't bat an eye. But some people did, and some people were like, holy shit. I just doxed my entire Yeah.
Genealogical branch. And but but then at that point, to have tools available to them so that they're not at a huge disadvantage, but they can still focus on the privacy freedom way, I think, is key. Because most of the time, they're at a disadvantage. Now I wanna be clear here. I'm not trying to, denigrate PPQ. I think actually PPQ right now is one of the best tools someone in a freedom privacy focused way can use. And even without eCash, like, I can log in with a VPN. And when I say log in, no login information whatsoever. Just pay with Lightning. I can do a set of prompts, and then I can just close that session and then do a new session from a different VPN later. And then I have similar privacy guarantees as if you were using eCash on your side.
Yeah. So, like, it is the best way for me to use top end models in a privacy focused way if the user knows what they're doing. Yeah. So thank you for that. I'm like, I'd I I wanna be clear here that I'm not like Yeah. Sure. Yeah. I appreciate that. And and, yeah, yeah.
[00:56:19] Matt Ahlborg:
Hot seat. We're trying to take the practical approach here, and I know that it's certainly not perfect. But,
[00:56:27] ODELL:
it's awesome. No. Quite frankly, it's awesome. So thank you. Yeah. Most people are appreciative, and and that's what keeps us going for sure. And I will say on the image generation side, you said, like, user behavior, does image generation bring you in? There's no other way for me to use, like, the top end image generation stuff. Exactly. There's no other way to use the software. Yeah. And it like, I don't use them for business stuff. I just use them because, like, I wanna create, like, shit posts or
[00:56:56] Matt Ahlborg:
memes or whatever. Or, like, yeah, meme troll troll family members and friends and stuff like that. Exactly. So,
[00:57:04] ODELL:
and I it it took me like, you were the first time I ever got to use a real image generation model, because I refuse to, you know, I never sign I'm probably one of I mean, there's plenty of people that haven't, but I've never signed up for chat g p t or
[00:57:20] Matt Ahlborg:
Yeah.
[00:57:21] ODELL:
Any of the major image generation proprietary models. Yeah. What else? I, I'm just kinda curious on your opinion on, like, the current state of Bitcoin. I mean, you were, like, the p two p guy. You were, like, the p two p freedom money guy, and you pivoted to AI. And then meanwhile, we have it's like paper Bitcoin somewhere. We got ETFs and MSTR and stuff. How are you thinking but we're also at all time highs. More people using Freedom Money than ever before. How do you think about where we're currently at in Bitcoin?
[00:58:05] Matt Ahlborg:
Yeah. It's I'm I have a big mess of thoughts. Yeah. I on the one hand, I think I underestimated the amount that the, you know, corporatization of Bitcoin would, like, help its price. I always thought that it would be the kind of, medium of exchange. Everybody kind of slowly grassroots adopting it, type thing, which hasn't popped off yet. And but I'm still I'm still in that corner, obviously, because I'm building PPQ. It's, hugely dependent on the success of Bitcoin as medium of exchange. Yeah. I don't know. Like I said, it's a big mess of thoughts about Bitcoin. I think that a lot of the p two p stuff went to stablecoins. And I think that, like, this gigantic plurality of all of these cryptos, they accomplish for a lot of users the goal of the user, and it doesn't necessarily require Bitcoin a lot of the time. So when they wanna do when a user wants to do they're they're living in a developing country and they wanna do a cross border transfer. Bitcoin is one option among literally thousands, and and stablecoin is probably the way that they're going to do it. Specifically Tether?
Yeah. Tether or even yeah. There's there's a there's a bunch. But, yeah. So in the way I pictured this journey on Bitcoin going is I always thought that, like like, Bitcoin is is kind of built for the financial oppression. Nuclear winner is kind of what I like to to I always thought the governments were gonna crack down hard, and we would need this very decentralized network, that that survived hard times. And we would need this lightning network, which is very difficult to build out and very hard for users to learn how to use, and there's so many technical, trade offs, that are happening or technical.
It's very conservative and because it wants to guarantee that sovereignty. But the the end result is, like, governments I guess they tried to crack down on in some degree, but, like, they were never really that successful. And so what we're left with is, like, there's literally thousands of shit coins out there. There's all these stable coins that were never cracked down upon, and now stable coins are actually becoming important for government, the US government in particular, because it's, you know, extending the timeline of the dollar.
So, yeah, I don't know. Bitcoin has gone in directions that I I guess in some ways I was right about, but in other ways, I was very wrong about. So, yeah, I don't know. A mess of thoughts, like I said. I mean, the way I think about it is
[01:01:02] ODELL:
is you really, I think it just really comes down to the extremes. Right? So, like, I think in, permissible times, right, and you can get around the, like, the Tron, Like, Tron can get away with just running, like, seven servers on Yeah. Corporate cloud servers and send Tether around the world and and be, like, the number one medium of exchange. Like, that works. Or you need to have, like, cockroach money, which is Bitcoin. Like, all the stuff in the middle is pretty much useless. Yeah. And and so it's like you owe Bitcoin is the safe haven. We have it no matter what. When there's no other option, you have Bitcoin. And then the majority of people are gonna use Tether, which is, you know, more or less, like, being indoctrinated as, like, almost a quasi CBDC, but, like, the best possible version of a CBDC because it's corporate run by Bitcoiners with that try and push back against seizing funds and whatnot, but still completely transparent and tied to the dollar. And everything in the middle is just, like, caught in no man's land.
Yeah. And that's kind of the reality we're in right now. And then 99% of new users that actually want Bitcoin price exposure are buying MSTR or
[01:02:19] Matt Ahlborg:
Do you think it's 99 or you think it's really that high? I think it's
[01:02:23] ODELL:
I think it's quite high. I mean, in terms of in terms of dollar amount, not in terms of users. Yeah. Yeah. Like, in terms of actual volumes, probably it's, you know, over 98%. Maybe I'm being provocative, but it's probably over 98%. But in terms of users, in terms of number of users, maybe we still have them beat because we have, like, the developing world and global permissionless markets. But in but but a lot of them are probably coming in with with smaller sums. Right? If someone's gonna allocate a $100,000,000 into Bitcoin, probably coming into iBit.
[01:03:01] Matt Ahlborg:
Yeah. Yeah. I don't know. It's yeah. I think you you said a lot of those things well, but,
[01:03:10] ODELL:
yeah, who knows who knows where we're at. And then my bull case my bull case is, I think, we have more people using Bitcoin as Freedom Money than ever before. Like, the number of people that are making self custody lightning payments higher than it's ever been. And the cool part is it's hard for us to calculate what that number is. The number of people self custodying Bitcoin, higher than it's ever been. Percentage wise, lower than it's been in the past. Yeah. But, in terms of absolute numbers, that number goes up.
[01:03:40] Matt Ahlborg:
It's true. And and I I think that's the fact that people say, like, well, we don't know how many lightning transactions are happening. That's a little bit of a I think, people are hoping that there's more than there are. There probably aren't. Yeah. But as I was the guy at Bitrefill, and Bitrefill is actually an excellent metric on, like because we accept all these cryptos and Yeah. And so we can actually see, like, how many users are actually bringing us lightning payments. And the answer is exactly what you said. It's in absolute numbers. It's never been higher. But in proportional numbers, it's actually, like, being beat out by a lot of other cryptos. So that's kind of a good news and bad news situation.
But, yeah, at the end of the day, I think I think Lightning is still, slowly but steadily improving, getting better, gaining users. I don't know. Is it enough to kind of keep up with with a lot of the other stuff that's happening? I don't know. And and, again, it it goes back to the question of, like, is this other stuff gonna be ever cracked down upon, like the Yeah. Tether on Tron type thing? I don't know. Probably not, though. I would I would say less likely than it was three years ago. Well, I mean, I think the cool thing about the vision of Bitcoin as, like, a Bitcoin on chain being, like, a foundational protocol layer that itself is cockroach
[01:05:07] ODELL:
type of trade off model, like, trying to survive a nuclear winter, is key because then you can build stuff on top of of it. And and Lightning itself took a certain set of trade offs to be censorship resistant and robust. And we just really haven't seen, besides Liquid, which kinda just never found product market fit, we've never seen, like, a more centralized trust model. And now there's some things being experimented on with, like, Arc and Spark that are coincidentally named the same, that rhyme with each other, or Cashew even. Right? Like, oh, how do we make more privacy preserving custodians? And so then all of a sudden, we could see we could see things go down to more centralized this like, maybe there's there's there's more usage on the centralized stuff.
But at the end of the day, if shit hits the fan, we can pull back to, like, the actual cockroach stuff. Maybe that's the bull case. I don't know. Is I also, like, I'm glad we're not getting cracked down on as much as I assumed. Like, I thought Bitcoin was gonna be straight up illegal in America in, like, 2017. So I still think it's the best timeline, and we'll kinda see how it all kinda see how it all plays out.
[01:06:18] Matt Ahlborg:
My bowl case is, like, I think it's, yeah, it's it's Bitcoin is its monetary properties. It's it's not the medium of exchange stuff that is really popping off. It's it's the fact that it is, competing as this beautiful, like, finite finance, finite supply, money on a on a global stage where everything else is just infinite, printing. I that's something I never that's the angle that I never really gave enough weight to, but I think it's totally starting to play out now where and I think, yeah, the the really great thing about one of the really great things about Bitcoin also is that it makes you start to ask questions about the fiat system and how it runs. Because until Bitcoin came along, nobody really ever asked questions. I'm like, okay. So you're just making money out of thin air. Like, what are the rules for that? Like, do we even know how much money you're printing? It says here you're printing this much, but then you have all these, like, classified budgets that are unaccounted for. So you're probably printing a lot more than you say you're printing and, like, the the holes just start getting poked into their their their presentation of what fiat is.
And and I think I think they're absolutely starting to lose the argument on, like, you know, before they could kind of just sit back and say, like, yeah. We we have this fiat thing, and we have all these, like, balancing mechanisms for it. And it's too complicated for you to understand, so just don't worry about it. But now because people are learning about, Finite money, and and the advantages of it, they're starting to ask questions about all these, like, bullshit, tricks that that the fiat system makes, and I think they're losing those arguments a lot. I love it when, like, the European Central Bank, like, post something on Twitter nowadays. It's, like, 98% angry Bitcoiners, like, just, telling them telling them, like, all of the lies that they're saying. And,
[01:08:28] ODELL:
yeah. I think they're losing the the the narrative war, the the story war. Yeah. I mean, even the term fiat, like, the term fiat was popularized by Bitcoiners. Like, that was not a term that was, like, in Yeah. Normal discourse and has become such recently, or, like, the last few years mostly because of the normalization from Bitcoiners. I mean, and just to bring it back, by the way, when I said, like, we don't know exactly how many people are making Bitcoin transactions or whatever, I actually like, I think it's on the lower end. Like, my question here's a question.
Just spitballing. No one's gonna hold you to it, and, also, no one can ever prove that you're wrong. How many Bitcoiners do you think exist in the world that own more than a half of Bitcoin in self custody. So more than I used to do this with a Bitcoin, but now we're at a 120. So I'll do a half of Bitcoin, $60,000 in self custody worth of Bitcoin. How many Bitcoins do you think exist?
[01:09:32] Matt Ahlborg:
Well, it can't be over 42,000,000. Right? So because that would be more than Right. So I would say, yeah, I would say maybe 5,000,000.
[01:09:44] ODELL:
I think it's less than that. So if you ask if you ask Probably is. Yeah. If you open PPQ and ask Grok three, how many own more than one Bitcoin, It comes up to a 100 to 250,000, and it gives you, like, all its reasoning or whatever.
[01:10:03] Matt Ahlborg:
Okay. Yeah. It's possible. Yeah.
[01:10:06] ODELL:
I think it's I think it's less than a million. What? What what is the point that you're driving at here? My point is is, first of all, it's cool that we have no idea. Just making a guess. Yeah. But my second point is is when Sailor says, like, there's a 100,000,000 Bitcoiners, it's like but it's just bullshit. Like, there's no we're still very small. Like, it's mostly a small vocal minority. Yeah. It's like, animals. You know, like, when you're, like, watching, like, National Geographic or something and they they show an animal, like, puff themselves up and look like a big bad beast. Like, we're a small town.
We're not even a city. Like, a city of a city of 200,000 people isn't a city. It's a large town Yep. Of actual Bitcoiners of of size.
[01:10:55] Matt Ahlborg:
This was a tweet that Marty put out a few years ago that really resonated with me. It said something about how, like, you know, 8% of colonials actually participated in the revolutionary war. So the vocal minorities really do change, history. So, you know,
[01:11:13] ODELL:
and I No. I agree. That's what we're doing. But it's I think it's important for people to ground themselves and realize, like, put it into perspective a little bit. Yeah. Anyway, Matt, this is a great chat. I'd love to have you on, like, I don't know, in, like, six to eight months and see where you're at and get some updates on AI. And, I mean, it's obviously a trend that I think everyone needs to keep an eye on. It'd be great to have you back. Do you have any final thoughts to the audience before we wrap?
[01:11:43] Matt Ahlborg:
Yeah. Don't have any final thoughts, but, thanks for listening. And and you can go to our, Noster end hub, and, it's probably posted somewhere around here. And, yeah, ask any questions you want. We are very much, building in the open as much as possible, and and, you know, user feedback and suggestions is is what drives the direction that we're going. So
[01:12:08] ODELL:
I love it. Yeah. Their website is ppq.ai. I'm gonna put links in the show notes to their master account and all the other relevant information. So just check the show notes if you wanna connect with Matt and the main account. Freaks, I see some comments in the YouTube chat. Yeah. The Nostra stream was down today. Kieran, who runs ZapStream, self hosts everything, and his router's getting DDoSed. I'm aware that it feels like Master isn't iterating quick enough. I still think it's amazing that that it's running and that it works to the extent that it does. We'll make those things better. But until then, we'll always be streaming no matter what. I'll make a way to get get the content to you guys even if we have to use YouTube, even if we have to use Twitch.
It is what it is. Next week, I got two shows lined up. One on Monday, we're gonna be talking about Bitcoin, with Ben Kaufman, great Bitcoin developer that works with Keeper Wallet right now. He used to work with Specter Wallet. And then on Friday with HuddleBot, who's mission focused on Nostr with Coracle and Flotilla, you can I'll I'll keep you guys posted on on Nostril with the actual schedule there. Primal.net/odell. Thank you all the freaks that continue to support the show. Thank you guys who joined us in the live chat, even though we weren't in our typical place, even though we weren't on the master live chat. Thank you for joining. You guys make the show unique. Huge shout out to Matt. Check out his awesome project. Give him feedback. It's very helpful at this point in their development.
Matt, thank you for joining.
[01:13:55] Matt Ahlborg:
Thank you so much for having me. Great conversation.
[01:13:58] ODELL:
Love it. Open Water Bitcoin is the best. Stay humble, StackSats.
