AI Agents Paying Writers in Stablecoins

Michael, good to see you again. Welcome to StableDash live. Welcome

to the Thanks for hanging out. Thanks for being in the green room. We got two Thanks for having me. Here, so we were going down some rabbit holes. But, yeah, we're excited to to dive into drip today because if you open your time up timeline now, I think AgenTic payments, creator payments is a drum that is being beat

consist pretty consistently. And and, Michael, you're you're doing the podcast tour. You're going to all of our friends and all these great creators and podcasts that we love. I think I I first saw Drip hit the scene hard with Tokenized and love those guys over there. Give us maybe the intro of, you know, the the evolution of what you're building, how how it's now entered into what you're focused on today. For those who are watching live, we had a chance to catch up and kind of get to know the business.

So let's step we'll we'll play dumb and kind of step back a few steps. Yeah.

Sure. So I'll give you a big brief of my background then, I think, just to kind of paint the whole picture. So I spent three and a half years at a sixteen z crypto on the investment team there. And so I did I did that for a while. I focused a lot on, like, NFTs, creator products, and DeFi at my while I was there. I left about a year and a half ago, I guess, something like that, and I joined my brother who also was building a crypto startup in the space called Royal.

So Drift is actually a product of the larger Royal Royal company. And my passion as as someone who is a creator myself and my brother is also a creator is to you know, we love crypto, and our passion is to just explore programmable money for creators and come up with new and creative ways that people can actually use crypto and blockchain to have better economics on the Internet. And the latest experiment in that is this product called DRIP.

And the super high level concept of Drip is we're building a platform that allows writers, people that have paywalled newsletters or blogs to make all of their content available to AI agents on a pay per use basis.

And so instead of having AI agents subscribe to you for $10.20 bucks a month because I don't believe AI agents are going to do that, they can actually just pay for your articles and newsletters as they need it, when they need it, for whatever research and analysis they're doing for the the human user on the other side. When when you started to, like, think about this, you know, we see kind of this the two sides of this agentic conversation.

We see the side of, like, hey. Consumers are just not gonna use it. This is just gonna be they're using agents for payments. And then we see other people like, no. Everybody's gonna have their own personal agent that's going that's going around. And when you think about your your ideal customer right now, like, how are they using AI? How are they maybe using agents to do research and to go out and to, you know, kinda consume all of this all this content? Yeah. So I think the long term vision is

exactly what you said. Everyone who's using a ChatGPT or Claude, they're gonna have a wallet installed in some form. And as they ask questions for research, it's gonna go out and maybe spend money to to acquire the relevant data to answer a user's question. However, that I think is a longer term vision. I don't think that's gonna happen necessarily very very quickly.

However, what we're focusing on is very specific specifically, and you can see this from from the website that that you're sharing, is on financial analysis and financial writers. And so you can come and use DRIP to pay per use and get access to some of the top financial analysts on the Internet that are typically behind paywalls or or or gated on on a lot of different, like, newsletter platforms.

So, you know, we support Substack, Beehive, even just people that have their own newsletters that just do distribution via email, but it's like paywall. Yeah. So we allow allow you to access all of those things on a pay per use basis. And the reason why we're starting with that specific target audience is because one of the things that I learned in crypto was that a lot of the well, as you all know, crypto is full of a loss of friction to get started.

Like, setting up a wallet, onboarding is very hard, and it sort of found product market fit with, you know, more gambling, trading, investing type applications, which I have lots of thoughts about. And, but but the reality is that that is what has worked in crypto to date.

And so I'm sort of taking a page out of that book and saying that I think the type of person that's very willing to go through the friction of setting up Codex or Claude, giving them a wallet, whether that be a stablecoin wallet, like, with agent cash or Circle or more of a credit card wallet, like the link wallet from Stripe. These things are they're easy to, like, use on a, you know, individual basis, but they are still, you know, complicated, and they do have some friction.

And so the type of person that's gonna go through that friction and be willing to is the type of person that thinks that on the other side of that friction is something extremely valuable, which usually is someone who thinks that they might be able to get some investing advice or maybe make money on the other side.

And so for that reason, we're focusing primarily on the target user of a of a maybe retail trader, prosumer trader that is already reading all these newsletters and wants a really good tool to aggregate and synthesize across them. I'm gonna pop this back up. We were just showing the actual user

the user journey here. So we went onto the site. I'm assuming you guys are kind of you're looking at this as like, you're funding some of this maybe on the front end, these these queries. But when I when I go in here and I'm one of the ones I selected here with this chat is a are AI stocks in a bubble? Give me the bull and bear case from recent finance writers. Selected that, and then it went walk us through what actually happens here. Maybe I can start a new chat, we can go go through this.

But what's actually happening when I select one of these on chain and maybe walk us through the mechanics a bit?

Yeah. For sure. And actually, it's funny. If you do it again, you may run out of credits. So we give we give Okay. Yes. We we give $1 of free credits to anyone that comes to the site to use it, and we have a a default price of 10¢ per publication or 10¢ per article. So, actually, see how it says no articles are purchased for this response? And that's because I think you maybe ran out of credits in the last time did so. If you open an incognito window, it would it would sort Gotta be fiscally responsible.

That makes sense. But I like I think it's cool how you can just you know, you're not asking me don't you know, I have I don't have to create an account. I don't have to know what this is. I can just go in and, like, see if this thing works, and it's you know, I I like that approach to it looks like yeah. I have 40¢ left today sponsored. Yeah. This is cool. Alright. So Yeah. What's happening here? Sure.

So we're actually about to launch a total revamp of this interface, and it's gonna make it actually really, really much better. So I think this is still the demo gods that not in our favor. No. No. It's okay. You you you you could still show it. It it is great. It does work. And but what's happening behind the scenes is we basically have our own agent, which is sort of mocking what your agent would do if you went the skill wallet route.

So you could totally interact with our system by downloading a skill that that we've written and also downloading some sort of Stablecoin wallet, and we support a a bunch of them that are on our site. And you can interact with the system through through, like, Codex or or Claude. But on our website, what's happening is you're just kinda interacting with with our agent that's doing the same thing.

So when you ask the question, what's going on is it's searching our our index of ingested articles, podcasts, and all these different sources that that we believe are are are relevant for financial analysis. Once it comes up with a list of of relevant sources, it will pay for them on a pay per use basis. Like I said before, that's 10¢ default right now, so it'll pay pay for each of those, get them all together, and then basically give you a synthesis of all that information as a response.

And so what's interesting is if you were to actually go and subscribe to all of these newsletters yourself, besides the fact that it would cost, you know, maybe $10.20, sometimes even a $100 a month to subscribe to some of these premium newsletters. Yeah. I did I did it, by the way. My inbox exploded. I think I maybe got, like, three, four emails an hour. And these emails are long. These are financial analysts. These are, like, huge reports. There's no way I can read all of that.

And so regardless, I would I would still wanna put it into my my agent and synthesize that information into Interesting. More digestible form. And so yeah. Go for it. No. No. Go ahead. I'll I'll I'm just gonna say the the the point of the website is to kinda let you experience the magic instantly with zero setup. So we'll sponsor the money. You know, we'll host the agent for you. We'll choose the articles for you. All you have to do is ask the question and kinda feel the full the full experience.

Have you I know you've I think when we first talked, you mentioned you you found us through our stable minded episode with Sam Bronner because you guys crossed paths at a 16 z. Have did you get the chance to watch the episode, those two episodes before that, with Louis Amara from ATXP?

I did not get to watch that, but I I do I do know Louis.

And so he's awesome. I was gonna say I I'm I'm always curious. Like, I think even what we struggle with, obviously, there's a bunch of agents running around behind the scenes at at StableDash, and we always I think where we still struggle at times is to figure out, like, how how do they actually need to interact and where do they kind of fit in with our stack. Like, where is Claude? Where is Codex? Where is Accursor? Then where is the ATXP? And then, like, what else do we need to even build on top of the stack?

I think one thing that stuck out from our interview with him was he said, you know, a day that kinda was like an moment was when we got an e a a support email from an agent that was looking for a wallet, and they found ATXP, and they were emailing us to try to figure out how to, you know, how to install it. Have you had any other, you know, experiences like that where agents just kind of find you? It's not even, you know, their end user that is that is saying, hey. I want to install DripStack.

It's just an agent that was just trying to find this information and hit a paywall and said, hey. Trying to I'm trying to use it, but I can't figure it out. It did happen one time. So I actually got an email from some I it must have been an email that someone gave their agent because it was, like, some some random email. I think probably through agent mail, I I believe. And it was literally them saying, like, we found your source. There's a bug.

We're trying to get access to this article, but we're getting this error. Like, can you help fix it? Like, that that's actually what what, what it was. And Wow. There there was a bug, and we did fix it. But when that happened, I was like, that is freaking awesome. That is super cool. That that someone, like, instead of going to our site and went to our support box and emailed the support email, no. This was, like, directly from their agent reaching out to me, which I thought was really cool. Yeah.

I always wanted to. I'm like, what is the stack that they're using that kinda, like, led up to that point?

So, you know, I think it's kinda really easy to get, like, in these you kinda get in the habit of, like, doing the same thing, and then you see Hermes launch, and then you see the OpenCLaw launch. And then, you know, you have to decipher what you see on x with some of these people running just super complex, agentic workflows. I think Claude just came out with a loop feature now too. So it's like, what did they do to get their agent to what and, like, what was their end goal? Right?

What were they trying to do that brought their agent, you know, kinda to this fork in the road of trying to come to DRIP and and get this information? What Drew, cut me off after this. You can take the next question. I have one more. I'm thinking about this from the perspective of original content creation. How do you ensure, you know, quality control here that, you know, the financial analysts that you're kinda aggregating isn't just, you know, a bunch of other regurgitated AI aggregation? Right?

Because, obviously, I'm not I'm not I'm not the type that's like, hey. Don't use AI at all to write. But there's obviously a fine line of, you know, how deep down this this rabbit hole do we go of? This newsletter was written by AI, that was written by all these other sources that then, you know, you guys are indexing and then disseminating to maybe other people that are writing other analysis. And it's like, at some point, is it just this Like human attributes. Deep. Yeah.

Yeah. The the easy answer, and this might not be the answer that you're looking for, but today, it's completely curated by me. Yeah. So I you know, these are writers that I've either been referenced from from friends or that I have personally subscribed to and read myself and I trust and I and I think are great. That being said, there are a few newsletters that we've come across that are they are like aggregator newsletters. Yeah.

And so they just kinda, like, basically do what our chat would do, which is they they subscribe to a bunch. They kinda get the summaries, and they sort of give you the the TLDR. I think there's actually some use some usefulness in that for for for the agent to have access to, like, e e even that context. I think, like, right now, though, it's it's very highly curated.

The other thing I'd say is, at least for our initial target users, I think we're maybe leaning more toward, like, not, like, super retail investor, but, like, the slightly, you know, medium investor because one of the things that we're gonna be rolling out in the in the coming days is instead of just asking our chat, like, hey. Answer this question. Go pull all these relevant articles and give them to me.

That's actually not a good experience because you could totally have people that, like, put a lot of buzzwords in their newsletter and then get really high search rankings in our search engine. And just like Google, it's like SEO manipulation, and people just get to the top that might not not not actually be the best content. And so what we're what we're gonna actually eventually do is let you be the curator. Like, you actually Mhmm.

Personally might only care about five newsletters and only care about synthesis across those five. Right? So our our job is right now, we're opinionated. I don't wanna be opinionated. Right? I wanna let you Yeah. As the user, you know, if if you have three guys that you follow that real you really wanna synthesize across, you you should be able to do that. But right now, we're just kind of, you know, the the aggregator and the curator. But, eventually, you'll you'll be able to have that a curation yourself.

Yeah. So that's great. I mean, I think too you mentioned podcasts. I think that's where I see, like, the human conversations really starting to shine too.

Yeah. We so we actually just launched, not like publicly, but it's live, a few podcasts in the in the chat. Actually, I think the first query that you showed when you had shared the screen earlier was a a podcast called Thoughtful Money. So that actually pulled from from transcripts.

And what's interesting about that is, like I was saying before, with the newsletters, using this pay per use basis via an AI agent plus the synthesis that you get from using an AI, that helps with the the monetary cost of these subscriptions because they're very expensive, and they're hard to synthesize across. Right? The podcast angle is similar, but instead of it being a a monetary cost, it's a a time cost. And so, you know Yeah.

Are, you know, I follow, 10 podcasts that each release, you know, hour long episodes a week. Like, I don't have that much time in my day to listen to all of that. You don't have ten hours a week to watch our content, Michael? Come on, man. I hate you guys. You guys you of of all It is a huge problem.

Yeah. It is a huge problem. Especially I have ten hours a week to watch mine. Yeah. I mean, sometimes sometimes Zach is like, hey. This is a really great podcast episode. I'm like, dude, like, this has been out for a week. It's, like, actually really hard. We were talking about this last stream where we both of us can't watch everything. We can't even consume everything that we put out. So if that's if that's like, we obviously review it editorially, and we're looking over things.

But to actually sit down and each time we push publish, consume the end thing. It even for us as publishers getting difficult because of the the I mean, we're we're streaming now, like, six, seven, eight hours a week. It's it's a lot of content. So it's a huge problem, and I I feel I feel like for us too is we're focused in niche.

A huge trend also you're seeing is, like, the the rise of niche creators and people who are going after I think what TBPN just blew the top off of, which is going after a niche but high value audience extremely well with with thoughtfulness and care. That's a really important like, when you're exploring depth and quality as a creator like us, that's something that's you know, you you you get the right view or one to five views.

If that view then results in, like, an attributed business lead or something of value, it's a it's an investor or helps close your seed round, get a big deal done, like, that is such a bigger deal than, you know, the Joe Rogan alternative of just spray and pray awareness. So it is an interesting thing. I'm I'm wondering for you why you picked the financial use case. You said that you're going after, like, the retail investor or, like, prosumer, I think, was the word you used. Why why that user base?

What are you seeing there?

So a few a few reasons. First, I think that there tend to be people that are early adopters to new technology, and they're willing to push through a lot of friction to use something because their goal is, of course, to invest and make money. And so someone who has that as the end goal usually is willing to push through some, like, bad UX experience to to get to the other side. And so at least for early adopters, especially of a product that is pretty complicated today, I think that they're they're a great user base.

The second is I'm a I'm a trader and investor myself. And so I experienced this problem myself, and I wanted to build a product that I personally would love to use. And so this is not just me, you know, thinking in the clouds here. Like, like, I I actually use these I subscribe to these newsletters. I I I do use this stuff myself. I am an active investor and trader, and so I wanted this tool. And eventually, though, I think the goal is to expand.

I think right now, we are very focused on the financial use case, but I completely see a future where someone can come to us and say, what's the best, you know, chicken parm recipe? And it goes out and, like, pays pays for, like, some of the top, you know, data recipes, creators on the Internet and pulls in and synthesizes them and maybe edits it based on your dietary restrictions and gives you the answer, I think that would be really cool too. But What's it called?

Do invest AI workflow like, all the AI newsletters where it's like, oh, the new tool, the new workflow,

the new skill, the new you know, hey. Claude had this update. Chat should be that's another area where it's, information overload. What's your thoughts, though, especially going back to the financial use case on agents actually edge executing on the user's behalf, you know, with this knowledge. We talked with Berhan. We talked about the SpaceX investment memo that went out. You know, he was like, hey. This was amazing that they were able to put this together.

And then now imagine the world where the agent is going to not just trade SpaceX, but then they're gonna put it into a vault and lend against it and all, you know, all the agent's doing all of this. Are you on the side of, for now, let the let the agent get the information. Let me make the decision as a trader, or do you even want your agent going and trading on your behalf? I'm gonna give you a real concrete story that happened in the past forty eight hours. So I've been spending

an immense amount of time building my own personal, you know, I don't I hate to use the word vibe investing, but, like, genetic investing set up. Right? And I've been spending hours and hours and hours purely on the data aggregation and synthesis piece. Pulling in real time prices, pulling in things from trip, pulling in news sources, all that stuff, and coming up with a report, which basically is is just doing what I already did manually.

But now I can just go to, you know, my codex and say, tell me about x stock or whatever, and it will give me that report. And it's just doing what I already would have done. And this morning, I kinda got to a place where I was like, I was happy with it. It was like a good a good start. And then I just had this thought. So just let me let me back up for a second. The way that I'm building this is using Interactive Brokers as the back end. That's like the the the trading brokerage and API that that I'm using.

And I just had this thought, like, maybe literally two hours ago, you know what? Why can't I just I did all this analysis. I'm ready to make the trade. Why can't I just let it make the trade? And so I prompted my codex. I said, hook into the Interactive Brokers API to place a trade on the stock and, you know, take in the following inputs that I give you, like, the limit order price and and how many shares I want, etcetera. It wrote a quick example in, like, thirty seconds, and I'm like, do it. Okay.

Play place a trade, and it did. And I see, like, on one side of my screen, have, like, my Interactive Brokers terminal, and then I have my agentic set up with in in Codex over here. And I can just see, like, magically the little order pop up in my Interactive Brokers terminal. And I was like, this is freaking awesome. So I did that did I this actually today this morning before I start talking to you guys. So so to so to answer your question, I will definitely be be using that.

That being said, there is one slight nuance about this whole object trading thing, which as as someone who who is a trader, I care a lot about real time prices. And I know, and I think everyone here should know if you've been in the crypto space, that things can move drastically in a second. Mhmm. And so if I asked my agent to do some analysis and determine a price and then actually submit submit the trade, I will admit today, it is slow and it is stale.

And so it it's actually really not the best interface for placing a trade. I think right now for me at least, like, it's cool. Like, it's a cool gimmick, but I still personally prefer the manual only because I can adjust prices if I need to in the last second and click order, and it's instant. Right? Whereas when I decided to place an order today, it took, like, five, ten seconds for the LLM to read the context, submit the order, execute the like, that's too long. Yeah. That will get better.

That will get better over time. But right now, it's still a little bit stale and slow for it being How do you interface. Michael, how do you see the supply versus the demand side of what you're building here? I'd imagine the supply of creators

versus the also, maybe the supply of, like, Interactive Broker like pieces of that layer. Are you gonna start smart routing things? Like, both of those sides feel really uniquely challenging to create what you're what you're thinking through. Are you and also, I think we've talked about, like, owning the application side of this too, where you're actually the one interfacing with that flow on the demand side.

Like, how are where are you thinking about eventually the the business' focus or, like, where you really wanna live versus the things that maybe you're you're doing today because it's it's helping you test the growth and get the rails going? Like, where do you see that? Great question. Where I wanna live is I wanna be spending

every single hour of my day talking to writers and explaining to them what we're building and getting them to say, hell yeah. I wanna do this, and onboarding them to DRIP, and getting them to upload their content, and then making it available to the demand side. That's where I would love to live and how I started building Drip and what I it's just that's that's the dream. Unfortunately, like any new technology, especially when you're a creator, creators are slow adopters to new technology.

They they don't necessarily, like, love or are they're not, like, very willing to try new things all the time. Yeah. They're maybe more artsy. They're more opinionated. They're you know? So they're very, very hard to get on the phone. And so that's why what I'm doing now, is more of the short term side, is I'm I'm a 100% demand focused. Right? Right?

I wanna build an amazing product for the demand side of DRIP, get users that wanna come to us every day to synthesize across their favorite newsletters and news sources, and get a a great report on the financial markets and the and the economy and geopolitics and whatnot. And if that works and there is a lot of demand there, then I get to go to the writers and say to them, hey. You should try this out because you actually have made money here already.

Like, users have been paying you money for yourself on a pay per use basis. You should you should pay attention to us. And that is a, I think, a stronger go to market onboarding experience than trying to sell someone the vision of AI agents are gonna buy your thing. Honestly, I've spoken to a handful of writers and a lot of them, they're very skeptical, and they should be. Right? Because they don't understand the demand. They don't understand who these agents are.

They probably even haven't even used the more agentic forms of these AI systems yet. Like Yeah. And what I mean by that is Clog Code and Codex versus ChatGPT. And so they just it's very constrained right now to the x sphere of tech nerds.

So, like, let's just pretend I'm, like, a writer or well, I guess in some ways, I am. But let's pretend I'm actually doing a lot more. And you say, hey. I've been is the is the vision, like, that you can approach writers and hey. I've been paying for your subscription for for six months now. I wanna formally onboard you onto DRIP. And by the way, you have this much to claim because people have already been doing this. Like, is that what somebody can expect coming coming to you?

That is exactly what I what I would hope happens. Yeah. Exactly. That's so to just just today, of course, like, it's still early, so, like, I don't have, like, the super sexy numbers to to pitch people yet. There's 10,000

of Yeah. Asians that have been. But, hey. I would love to get there soon. Yeah.

Exactly. Let's come in. So much so much to cover here. Let me if you have anything well, as you kind of test the Interactive Brokers piece and things like that, like, really excited to see more of that. It sounds like you guys are about to ship an update of the site too, so excited to see this and really just, you know, personally, as as creators and people on the the supply side of what we're talking about, it it's just really great to see that there are people building and thinking through this.

And, yeah, really excited to watch the journey and continue to I mean, as we said last time, it's been really cool to see Royal. The the I mean, the same, like, the same thing that gets me excited about Drip is the same thing that gets that got me excited about the very early days of Royal. It's like the same core mission of helping creators I think you said say it again. Helping creators explore programmable money. Exactly. Yeah.

It's our our, like, sort of thesis and mission statement as a company is exploring programmable money for creators. Yeah. Same same mission. It's just a different approach. So thank you for all the work you're doing, and we'll

get you back on the show again soon. Definitely. Thanks for having me, guys. This was great. Thanks, Michael. See you.