1 00:00:00,040 --> 00:00:03,320 What we believe is that we're actually witnessing the rise of 2 00:00:03,320 --> 00:00:06,880 a new consumer that's going to manifest as trillions of AI 3 00:00:06,880 --> 00:00:10,240 agents. And in order to scale these 4 00:00:10,240 --> 00:00:15,120 systems, we're going to need to rethink and rebuild a big chunk 5 00:00:15,120 --> 00:00:18,720 of the payments infrastructure. Being able to understand what's 6 00:00:18,720 --> 00:00:22,440 actually being used in that analytical pipeline, like how 7 00:00:22,440 --> 00:00:25,640 much of A given data set, how many times is an algorithm 8 00:00:25,640 --> 00:00:28,080 called? What's like the computational 9 00:00:28,080 --> 00:00:32,960 cost of executing that algorithm in conjunction with that data 10 00:00:32,960 --> 00:00:35,320 set to say, train a model or what have you? 11 00:00:35,480 --> 00:00:40,040 That output then gets commercialized and everybody in 12 00:00:40,040 --> 00:00:43,400 that value chain is somehow rewarded because without that, 13 00:00:43,400 --> 00:00:47,080 there's actually nothing to pass back upstream. 14 00:00:47,360 --> 00:00:50,160 Something that needs to be addressed in the web free space 15 00:00:50,160 --> 00:00:54,760 is there's often times a trivial outlook on payments. 16 00:00:54,920 --> 00:00:58,320 And that's because I I personally believe people 17 00:00:58,320 --> 00:01:00,680 conflate settlement for payments. 18 00:01:01,160 --> 00:01:03,640 Welcome to Epicenter, the show, which talks about the 19 00:01:03,640 --> 00:01:06,360 technologies, projects and people driving decentralization 20 00:01:06,360 --> 00:01:09,200 in the blockchain revolution. I'm Rodrique ants and today I'm 21 00:01:09,200 --> 00:01:13,000 speaking with with Don Gossan, who is the CEO and Co founder of 22 00:01:13,000 --> 00:01:16,760 Nevermind. So Nevermind as in Nevermind as 23 00:01:16,760 --> 00:01:21,920 in has never been mined because there's quite a few projects 24 00:01:21,920 --> 00:01:23,880 kind of like with very similar names, right? 25 00:01:24,280 --> 00:01:29,240 And Nevermind is positioning itself as the PayPal for AI to 26 00:01:29,240 --> 00:01:33,240 AI payments. Before I talk with Don, these 27 00:01:33,240 --> 00:01:37,360 are our sponsors this week. If you're looking to stake your 28 00:01:37,360 --> 00:01:40,440 crypto with confidence, look no further than course one. 29 00:01:40,880 --> 00:01:44,480 More than 150,000 delegators, including institutions like Bid 30 00:01:44,480 --> 00:01:47,160 Gold, Pantera Capital and Ledger Trust Course One with their 31 00:01:47,160 --> 00:01:50,440 assets, they support over 50 block chains and are leaders in 32 00:01:50,440 --> 00:01:53,640 governance on networks like Cosmos, ensuring your stake is 33 00:01:53,640 --> 00:01:56,600 responsibly managed. Thanks to the advanced MEV 34 00:01:56,680 --> 00:01:59,440 research, you can also enjoy the highest staking rewards. 35 00:01:59,880 --> 00:02:02,920 You can stake directly from your preferred wallet, set up a white 36 00:02:02,920 --> 00:02:06,680 label node, restake your assets on Eigenia or Symbiotic, or use 37 00:02:06,680 --> 00:02:08,960 the SDK for multi chain staking in your app. 38 00:02:09,440 --> 00:02:12,560 Learn more at Chorus .1 and start staking today. 39 00:02:14,240 --> 00:02:17,480 This episode is proudly brought to you by Gnosis, a collective 40 00:02:17,480 --> 00:02:19,960 dedicated to advancing a decentralized future. 41 00:02:20,440 --> 00:02:24,640 Nosys leads innovation with Circles, Nosys Pay and Metri, 42 00:02:24,720 --> 00:02:26,920 reshaping open banking and money. 43 00:02:27,280 --> 00:02:30,360 With Hashi and Nosys VPN, they're building a more 44 00:02:30,360 --> 00:02:32,840 resilient, privacy focused Internet. 45 00:02:33,280 --> 00:02:37,040 If you're looking for an L1 to launch your project, Nosys Chain 46 00:02:37,040 --> 00:02:40,400 offers the same development environment as Ethereum with 47 00:02:40,400 --> 00:02:44,480 lower transaction fees. It's supported by over 200,000 48 00:02:44,480 --> 00:02:47,920 ballot errors, making Nosys Chain a reliable and credibly 49 00:02:47,920 --> 00:02:49,840 neutral foundation for your applications. 50 00:02:50,360 --> 00:02:54,160 Gnosis Dow drives Gnosis governance, where every voice 51 00:02:54,160 --> 00:02:56,880 matters. Join the Gnosis community in the 52 00:02:56,880 --> 00:03:01,320 Gnosis Dow forum today. Deploy on the EVM compatible 53 00:03:01,320 --> 00:03:06,400 Gnosis Chain or secure the network with just one GNO and 54 00:03:06,400 --> 00:03:09,440 affordable hardware. Start your decentralization 55 00:03:09,440 --> 00:03:12,400 journey today at gnosis dot IO. Done. 56 00:03:12,800 --> 00:03:16,440 Thank you so much for coming on. Thanks for having me. 57 00:03:17,880 --> 00:03:23,680 Maybe we can get off to a start by talking about you yourself. 58 00:03:23,680 --> 00:03:26,040 What what's your background and kind of like what brought you 59 00:03:26,040 --> 00:03:31,800 here? Canadian transplant in Europe, 60 00:03:33,640 --> 00:03:39,920 so I live in Lisbon now, by way of Berlin, by way of London, by 61 00:03:39,920 --> 00:03:42,720 way of Tokyo, by way of Los Angeles. 62 00:03:43,000 --> 00:03:46,880 So I've lived all over the world, studied engineering 63 00:03:48,080 --> 00:03:53,120 university, went into commodities trading after 64 00:03:53,120 --> 00:04:01,320 university I was on the risk side and so doing back then what 65 00:04:01,320 --> 00:04:07,960 was called statistical modeling, which became machine learning 66 00:04:08,200 --> 00:04:12,680 and now has been Co opted by the AI branding. 67 00:04:14,000 --> 00:04:19,560 But yeah, just basically augmenting internal credit 68 00:04:19,560 --> 00:04:24,760 histories with external credit scoring and figuring out which 69 00:04:24,760 --> 00:04:28,120 of our clients were deadbeats and which ones weren't, so which 70 00:04:28,120 --> 00:04:32,560 ones we could like loan money to in order for them to hedge and 71 00:04:32,560 --> 00:04:36,920 stuff like that. So it was pretty boring, to be 72 00:04:36,920 --> 00:04:42,520 honest. And then I went into IT 73 00:04:42,520 --> 00:04:47,400 consulting for the better part of a decade and 1/2 as a subject 74 00:04:47,400 --> 00:04:49,040 matter expert in data and analytics. 75 00:04:49,040 --> 00:04:50,880 And that's what took me all over the world. 76 00:04:50,880 --> 00:04:59,360 So I built large scale data states for ML purposes for some 77 00:04:59,360 --> 00:05:02,360 of the biggest companies on the planet. 78 00:05:02,400 --> 00:05:07,280 So I was at HSBC and L'Oreal and AXA and Mizuho and stuff like 79 00:05:07,280 --> 00:05:12,400 that. And so yeah, I've, I've been in 80 00:05:12,400 --> 00:05:21,040 the machine learning space my entire career, so 20 years and 81 00:05:21,040 --> 00:05:25,240 then added the crypto flavor in 2016. 82 00:05:27,040 --> 00:05:37,520 So I got introduced to blockchain not as a, you know, a 83 00:05:37,560 --> 00:05:41,680 system or platform or ecosystem for like speculation and 84 00:05:41,680 --> 00:05:46,320 payments and settlement and stuff, but more on the side of 85 00:05:46,440 --> 00:05:51,800 providential integrity. So block chains are very elegant 86 00:05:54,000 --> 00:06:01,000 provenance machines and asset provenance is a really hard 87 00:06:01,000 --> 00:06:08,040 problem to contend with within the confines of large analytical 88 00:06:08,040 --> 00:06:10,920 estates, right? You've got to, you've got to 89 00:06:10,920 --> 00:06:14,600 answer 4 questions with high fidelity, where is the asset 90 00:06:14,600 --> 00:06:16,040 coming from? Where is it going? 91 00:06:16,600 --> 00:06:19,040 Who's using it and what are they doing with it, right. 92 00:06:19,280 --> 00:06:22,640 And the assets can be like data sets, they can be algorithms, 93 00:06:24,080 --> 00:06:27,640 all kinds of stuff. And if you can't answer one of 94 00:06:27,640 --> 00:06:32,560 those questions, it undermines the integrity of the, the 95 00:06:32,560 --> 00:06:35,640 output. So they're quite critical 96 00:06:35,640 --> 00:06:38,640 questions to answer, but they it just so happens that with 97 00:06:38,640 --> 00:06:43,600 contemporary software, it's really hard to actually answer 98 00:06:43,600 --> 00:06:47,760 that stuff. So what ultimately happens or 99 00:06:47,760 --> 00:06:52,400 usually happens is that you create these very like bespoke 100 00:06:53,280 --> 00:07:00,120 one off patchwork solutions that like cobble together information 101 00:07:00,120 --> 00:07:04,240 from all of your different sources and destinations to try 102 00:07:04,240 --> 00:07:06,920 and figure out the topology of what's going on. 103 00:07:07,560 --> 00:07:11,320 Anyway. Block chains help plug into that 104 00:07:11,320 --> 00:07:16,120 and make it much more seamless in terms of understanding the 105 00:07:16,120 --> 00:07:20,800 answers to those four questions. And so that's what kind of got 106 00:07:20,800 --> 00:07:25,520 me hooked initially, sort of within the grander scope of my 107 00:07:26,040 --> 00:07:30,160 my career. And so have been at this 108 00:07:30,160 --> 00:07:36,400 crossroads of AI and, and web 3 for going on a decade now 'cause 109 00:07:36,400 --> 00:07:39,800 that shortly after, you know, I, I kind of went down the crypto 110 00:07:39,800 --> 00:07:41,960 rabbit hole. I Co founded a project in Berlin 111 00:07:41,960 --> 00:07:45,600 called ocean protocol, which was one of the first projects at 112 00:07:45,600 --> 00:07:47,320 this intersection of AI and Web 3. 113 00:07:48,160 --> 00:07:52,960 And I've just, you know, kept beating this drum ever since. 114 00:07:52,960 --> 00:07:59,640 And now it's we've we've sort of distilled the learnings and 115 00:07:59,640 --> 00:08:04,240 experiences and understanding of merging these two technologies 116 00:08:04,240 --> 00:08:14,160 into a very hyper focus on AI payments can get into why we're 117 00:08:14,680 --> 00:08:17,440 hyper focused on this stuff. But yeah, that's the background, 118 00:08:17,440 --> 00:08:18,840 yeah. Absolutely. 119 00:08:19,360 --> 00:08:25,640 So you talked about kind of the, the provenance of data and kind 120 00:08:25,640 --> 00:08:30,520 of this entire data economy that kind of come comes that kind of 121 00:08:30,520 --> 00:08:34,880 crucially hinges on that How, how, how does payments actually 122 00:08:34,880 --> 00:08:39,320 fit into that picture? It's a this is a good question 123 00:08:39,320 --> 00:08:45,120 and and it took us quite a while to realize sort of the gravity 124 00:08:45,120 --> 00:08:50,480 of the payments piece. So where what we were focused on 125 00:08:50,480 --> 00:08:55,720 for a long time was in establishing the provenance 126 00:08:55,960 --> 00:09:00,840 component so that we could do by, you know, extension the 127 00:09:00,840 --> 00:09:05,360 attribution piece, right? So taking this holistic view 128 00:09:05,960 --> 00:09:11,760 that we want to build these analytical systems, and then now 129 00:09:11,760 --> 00:09:15,040 take, let's say, the form factor of an AI agent, right? 130 00:09:16,160 --> 00:09:20,880 We want to build these things in a decentralized landscape. 131 00:09:21,520 --> 00:09:26,080 So what does that mean kind of holistically? 132 00:09:27,080 --> 00:09:31,600 And it means being able to understand what's actually being 133 00:09:31,600 --> 00:09:33,920 used in that analytical pipeline. 134 00:09:34,960 --> 00:09:38,040 How much of anything's being used, right? 135 00:09:38,080 --> 00:09:41,240 Like how much of A given data set, how many times is an 136 00:09:41,240 --> 00:09:44,920 algorithm called? What's like the computational 137 00:09:44,920 --> 00:09:51,120 cost of executing that algorithm in conjunction with that data 138 00:09:51,120 --> 00:09:56,760 set to say, train a model or what have you, and, and 139 00:09:57,080 --> 00:10:02,800 basically accounting for all of that and then extrapolating or 140 00:10:02,800 --> 00:10:04,880 extending that into the attribution piece. 141 00:10:04,880 --> 00:10:11,360 So Frederica, you provide some contextual data, I provide some 142 00:10:11,520 --> 00:10:16,680 training data, somebody else provides the algorithm, another 143 00:10:16,680 --> 00:10:19,240 third party provides the infrastructure for bringing all 144 00:10:19,240 --> 00:10:24,200 of this together. And when combined, we create 145 00:10:24,200 --> 00:10:27,520 this inference or actionable insight or whatever we want to 146 00:10:27,520 --> 00:10:31,040 call the output. That output then gets 147 00:10:31,680 --> 00:10:36,760 commercialized and everybody in that value chain is somehow 148 00:10:36,760 --> 00:10:40,760 rewarded. Right now, what we realized a 149 00:10:40,760 --> 00:10:46,320 few years ago is actually the most critical piece in that 150 00:10:46,320 --> 00:10:51,600 whole workflow is the end state. It's the last mile. 151 00:10:51,640 --> 00:10:56,760 It's commercializing that output, enabling that thing to, 152 00:10:57,560 --> 00:11:02,160 you know, enabling payment for that inference or that 153 00:11:02,160 --> 00:11:05,840 actionable insight. Because without that, there's 154 00:11:05,840 --> 00:11:10,120 actually nothing to pass back up stream. 155 00:11:10,360 --> 00:11:14,080 There's really other than recognition like a nice pat on 156 00:11:14,080 --> 00:11:16,160 the back. Hey Federica, thanks for 157 00:11:16,160 --> 00:11:20,720 providing that context data, thumbs up, reputational reward, 158 00:11:20,720 --> 00:11:23,120 whatever. There's actually nothing if you 159 00:11:23,160 --> 00:11:29,800 if you can't capture the end state utility, there's nothing 160 00:11:29,800 --> 00:11:34,240 actually to translate back upstream to the different 161 00:11:34,240 --> 00:11:37,080 participants. And so in and amongst everything 162 00:11:37,080 --> 00:11:42,040 that we were building, there was that piece that that payment 163 00:11:42,240 --> 00:11:44,480 system in the attribution component. 164 00:11:45,280 --> 00:11:47,520 And well, it's part of the attribution component. 165 00:11:48,040 --> 00:11:50,440 And what we realized a couple of years ago was like that's 166 00:11:50,440 --> 00:11:51,960 actually the most important piece. 167 00:11:52,400 --> 00:11:56,200 If we don't get that piece right, then we can't do all of 168 00:11:56,200 --> 00:12:00,200 the other stuff that we want to do from a providential integrity 169 00:12:00,200 --> 00:12:05,160 and attribution point of view. So we became hyper focused on 170 00:12:05,160 --> 00:12:09,040 that one component within this whole workflow. 171 00:12:09,920 --> 00:12:13,520 So Ocean is very much sit around and kind of like Ocean, I mean, 172 00:12:13,960 --> 00:12:15,880 trends been on this podcast multiple times. 173 00:12:15,880 --> 00:12:18,680 So kind of like listeners will probably know that it it's kind 174 00:12:18,680 --> 00:12:22,560 of it's focused on the data economy and having a marketplace 175 00:12:22,560 --> 00:12:27,200 for, for, for data sets. It seems to me that kind of like 176 00:12:27,200 --> 00:12:34,720 payments, it is just a very natural part of any marketplace. 177 00:12:34,720 --> 00:12:36,000 Why? Why do you think it kind of 178 00:12:36,000 --> 00:12:38,640 makes part to kind of ply these two things apart? 179 00:12:41,080 --> 00:12:48,200 Because payment systems are complex and and they're non 180 00:12:48,200 --> 00:12:54,760 trivial, let's put it that way right as our marketplaces. 181 00:12:56,320 --> 00:12:59,960 The other realization purely from a marketplace point of view 182 00:12:59,960 --> 00:13:10,840 is marketplaces are, are difficult propositions from a 183 00:13:10,840 --> 00:13:13,320 commercial and operational point of view. 184 00:13:14,600 --> 00:13:21,000 They, they tend to not make a lot of money and therefore 185 00:13:21,000 --> 00:13:25,240 they're they're hard to persist. They're hard to like accrue 186 00:13:25,240 --> 00:13:27,120 enough revenue to keep the thing going. 187 00:13:27,360 --> 00:13:32,000 The way that so that it basically marketplaces are one 188 00:13:32,000 --> 00:13:35,520 in the margins and the way that you win marketplaces is through 189 00:13:35,520 --> 00:13:39,000 monopolization. And that monopolization usually 190 00:13:39,000 --> 00:13:43,800 comes from a very discreet focus on a specific domain or a niche 191 00:13:43,800 --> 00:13:49,680 within a specific domain, right? So the common marketplaces that 192 00:13:49,680 --> 00:13:54,160 are sort of presented are like Bloomberg, right? 193 00:13:54,760 --> 00:13:59,080 Or maybe I'll Sevier for a more abstract 1 and like the research 194 00:13:59,480 --> 00:14:04,400 domain, but they're very difficult to win and they 195 00:14:04,400 --> 00:14:10,640 require a significant amount of focus in order to actually make 196 00:14:10,640 --> 00:14:15,160 a successful marketplace. You know, personally, I would 197 00:14:15,160 --> 00:14:18,920 say one of the learnings out of the last decade for me is that 198 00:14:20,880 --> 00:14:25,160 I'm pretty skeptical about general purpose marketplaces. 199 00:14:25,160 --> 00:14:30,480 I think it's, you know, I think there's a lot more examples of 200 00:14:30,480 --> 00:14:33,080 those failing than there are of successes. 201 00:14:35,520 --> 00:14:40,760 So all that aside, that's one part of the argument and then 202 00:14:40,760 --> 00:14:45,200 the other is a realization that that payments are are quite 203 00:14:45,960 --> 00:14:49,840 complex in their in their manifestation, right. 204 00:14:50,720 --> 00:14:55,840 So it's not, I think there's a, there something that needs to be 205 00:14:55,840 --> 00:15:00,280 addressed in the web free space is there's often times that a 206 00:15:00,280 --> 00:15:04,960 trivial outlook on payments. And that's because I, I 207 00:15:05,080 --> 00:15:09,920 personally believe people conflate settlement for payments 208 00:15:10,640 --> 00:15:14,080 and, and payment processing and, and that sort of thing. 209 00:15:16,400 --> 00:15:25,480 And so just anecdotally, right, there's, you know, like if you 210 00:15:25,480 --> 00:15:28,160 don't want to take my word for, for the complexity of these 211 00:15:28,160 --> 00:15:32,800 things, there's a company called Metronome that does like 212 00:15:33,400 --> 00:15:38,000 payments and billing services, predominantly for SAS providers, 213 00:15:38,240 --> 00:15:44,800 right? And the, the CEO of that company 214 00:15:44,800 --> 00:15:48,000 and, and a few of the founders, they come out of, of Box of, out 215 00:15:48,000 --> 00:15:53,480 of Dropbox, right? What they recognized was there 216 00:15:53,480 --> 00:15:56,400 was a discrete issue with the payments and billing 217 00:15:56,400 --> 00:15:58,680 functionality that Drop was creating. 218 00:15:58,680 --> 00:16:05,800 So they had if, if, if they had something to the effect of 60 219 00:16:05,800 --> 00:16:13,440 engineers at Dropbox just supporting the payments, the 220 00:16:13,440 --> 00:16:17,440 price setting the infrastructure and the billing at that company. 221 00:16:17,880 --> 00:16:23,720 That's like a huge undertaking from an engineering standpoint, 222 00:16:23,720 --> 00:16:25,760 right? The reason for this is like 223 00:16:26,360 --> 00:16:29,960 operating across a bunch of different divergent, you know, 224 00:16:30,000 --> 00:16:31,520 jurisdictions and stuff like that. 225 00:16:31,520 --> 00:16:34,960 So they have like different price points for different 226 00:16:34,960 --> 00:16:38,320 localities, all of these different, you know, depending 227 00:16:38,320 --> 00:16:42,000 on the customer and the volume and all of this like prices 228 00:16:42,000 --> 00:16:45,920 would would need to fluctuate. And so they had a team of 60 229 00:16:45,920 --> 00:16:49,080 engineers supporting this. Their realization was, well, 230 00:16:49,080 --> 00:16:51,120 we're not going to be much different from our competitors. 231 00:16:51,120 --> 00:16:54,160 We're not going to be much different from Databricks or you 232 00:16:54,160 --> 00:16:57,520 know, these these other like SAS providers for these services. 233 00:16:58,240 --> 00:17:04,200 So why don't we take what would build here, extract it and offer 234 00:17:04,200 --> 00:17:08,720 it as a service to multiple companies. 235 00:17:09,040 --> 00:17:14,160 And so it's that kind of application that we're looking 236 00:17:14,160 --> 00:17:18,640 at providing, but in this case discreetly for agentic 237 00:17:18,640 --> 00:17:21,760 transactions, AI to AI. Transactions. 238 00:17:21,760 --> 00:17:26,200 So I understand that kind of like pricing and kind of price 239 00:17:26,720 --> 00:17:30,320 pricing strategy can be arbitrarily complex. 240 00:17:31,000 --> 00:17:34,720 But given that kind of I would assume most of these agents kind 241 00:17:34,720 --> 00:17:38,080 of set their own prices or kind of they kind of that this is 242 00:17:38,080 --> 00:17:42,160 something that's pre negotiated. Walk us through the complexity, 243 00:17:42,280 --> 00:17:44,760 the teas of payments because kind of like, I mean, I have, I 244 00:17:44,760 --> 00:17:48,080 have as someone who also works in payments, I have said this 245 00:17:48,080 --> 00:17:51,280 time and time again that kind of I think it makes sense to kind 246 00:17:51,280 --> 00:17:55,400 of like productionize things in web three first that have simple 247 00:17:55,400 --> 00:17:57,200 use cases. And I always say kind of like I 248 00:17:57,200 --> 00:18:00,120 always say payments and principles of fairly simple use 249 00:18:00,120 --> 00:18:01,640 case. If you compare it to other 250 00:18:01,640 --> 00:18:04,480 things that also warrant disruption like social media and 251 00:18:04,480 --> 00:18:06,120 so on. Because kind of like it's, it's 252 00:18:06,120 --> 00:18:09,600 kind of like it's balances that go from one place to another. 253 00:18:09,600 --> 00:18:12,160 Kind of like it ideally kind of they should be consulted. 254 00:18:12,440 --> 00:18:13,520 You shouldn't. Nuance. 255 00:18:14,360 --> 00:18:15,440 To it, right? Exactly. 256 00:18:15,440 --> 00:18:17,360 So walk us through kind of like, what makes it difficult. 257 00:18:18,320 --> 00:18:22,560 I I, I I'm, I'm clearly speaking with with a kindred spirit here, 258 00:18:22,680 --> 00:18:25,600 right? So this is the other kind of 259 00:18:25,600 --> 00:18:35,440 revelation, like it's simple within the context of like it's 260 00:18:35,520 --> 00:18:38,760 easy to understand, it's not esoteric. 261 00:18:38,800 --> 00:18:42,920 And it like the infrastructure, the technology that we're using 262 00:18:42,960 --> 00:18:48,240 actually makes sense to focus on payments, right? 263 00:18:48,720 --> 00:18:54,920 So in that sort of very general sense, it's easy, you know what 264 00:18:54,920 --> 00:18:58,360 I mean? So why is this? 265 00:18:58,840 --> 00:19:00,640 Why is this important? OK. 266 00:19:00,640 --> 00:19:05,360 So our view of the world is like a particular one. 267 00:19:07,440 --> 00:19:12,800 I think up until about 6 months ago, it was relatively unique 268 00:19:14,360 --> 00:19:18,680 because most people back then didn't know what an AI agent is 269 00:19:18,680 --> 00:19:21,280 like. It's AI agents are now starting 270 00:19:21,280 --> 00:19:26,240 to emerge conceptually for some people, I'm not going to say 271 00:19:26,240 --> 00:19:31,000 most people, but those that are kind of either in the AI space 272 00:19:31,000 --> 00:19:38,880 or adjacent to it, you know, are are this is entered the lexicon, 273 00:19:38,880 --> 00:19:40,680 right? Like people are getting familiar 274 00:19:40,680 --> 00:19:44,360 with what an AI agent is. Even those in the AI space like 275 00:19:44,680 --> 00:19:51,560 agentic AI mixtures of experts like this concept was even an AI 276 00:19:51,560 --> 00:19:57,000 was quite fringe up until relatively recently. 277 00:19:57,720 --> 00:20:03,520 So our view of the world, what we believe is that we're 278 00:20:03,520 --> 00:20:07,480 actually witnessing the rise of a new consumer that's going to 279 00:20:07,480 --> 00:20:10,160 manifest as trillions of AI agents. 280 00:20:11,200 --> 00:20:16,560 And in order to scale these systems, we're going to need to 281 00:20:16,560 --> 00:20:21,240 rethink and rebuild a big chunk of the the payments 282 00:20:21,240 --> 00:20:25,360 infrastructure. You know if if. 283 00:20:26,240 --> 00:20:29,600 You hold this vision of the world like we don't we don't 284 00:20:29,600 --> 00:20:33,760 believe in the monolithic one AI to rule them all right, like the 285 00:20:33,760 --> 00:20:36,760 God AI. So we believe in this concept of 286 00:20:36,760 --> 00:20:41,280 mixture of experts that there's going to be sort of finely 287 00:20:41,280 --> 00:20:46,000 trained agents that will provide very discreet expertise that 288 00:20:46,000 --> 00:20:49,400 will be called upon when and where needed, right. 289 00:20:50,360 --> 00:20:56,800 And so holding this view, it kind of becomes relatively 290 00:20:56,800 --> 00:21:01,200 obvious that like, actually we need to work on at a minimum 291 00:21:01,200 --> 00:21:04,840 standardization or protocolization of this payments 292 00:21:04,840 --> 00:21:07,600 process, right? Because if you just think about 293 00:21:07,600 --> 00:21:10,760 this scenario where you have ecosystems with effectively 294 00:21:10,760 --> 00:21:16,560 trillions of agents, all either, you know, collaborating or in 295 00:21:16,560 --> 00:21:20,920 competition, but ultimately consuming, buying and selling 296 00:21:20,920 --> 00:21:23,480 from one another. If each one has their own 297 00:21:23,480 --> 00:21:28,800 payments mechanism, not only do you have to negotiate the price 298 00:21:28,800 --> 00:21:31,240 for that good or service that the counterpart is providing, 299 00:21:31,240 --> 00:21:34,440 you also have to negotiate which payment system you're going to 300 00:21:34,440 --> 00:21:36,400 use, right. But how is? 301 00:21:36,440 --> 00:21:39,920 It different from from non AI. So kind of like if I kind of 302 00:21:39,920 --> 00:21:43,640 make a deal with someone who's human, kind of like why, why do 303 00:21:43,640 --> 00:21:46,120 you need one? That's that's it's kind of kind 304 00:21:46,120 --> 00:21:48,880 of specifically to a is. Sure. 305 00:21:48,880 --> 00:21:52,320 So it's intrinsically like it's not that dissimilar. 306 00:21:52,320 --> 00:21:57,280 It's the way that effectively the information gets packaged 307 00:21:57,280 --> 00:22:00,040 around the service that's actually being provisioned. 308 00:22:00,560 --> 00:22:03,400 So let me unpack that a little bit. 309 00:22:05,760 --> 00:22:09,600 If you're familiar with AI and, and, and how this stuff works 310 00:22:11,560 --> 00:22:16,800 and in particular an AI agent. So we're like our tech can be 311 00:22:16,800 --> 00:22:19,520 used for like pricing models and stuff like that. 312 00:22:19,520 --> 00:22:23,800 But really we're looking at these analytical pipelines in 313 00:22:23,800 --> 00:22:27,320 the form factor of an AI agent, which is a compilation of 314 00:22:27,320 --> 00:22:33,320 different AI tools, right? So at a minimum, it's like the 315 00:22:33,320 --> 00:22:38,320 ability to likely source an inference from more than one 316 00:22:38,320 --> 00:22:41,880 model, right, depending on the complexity. 317 00:22:41,880 --> 00:22:47,320 So like I mean operator does this if you've used say 4 O or 318 00:22:47,320 --> 00:22:56,280 O1 from Open AI, you put in a prompt and you and I can be like 319 00:22:56,280 --> 00:23:00,760 completely divergent users. There is sort of there's under 320 00:23:00,760 --> 00:23:03,320 the hood, there's logic that takes place that routes the 321 00:23:03,320 --> 00:23:09,480 request to the the model that will adequately handle the 322 00:23:09,480 --> 00:23:11,560 complexity of that request, right. 323 00:23:11,920 --> 00:23:17,160 So the very general broad kind of use case that I provide, it's 324 00:23:17,160 --> 00:23:22,520 not or or example that I provided, it's not use case to 325 00:23:22,520 --> 00:23:25,840 lie. You and I are users of some 326 00:23:25,840 --> 00:23:29,560 third party agent, right? And let's just say that that 327 00:23:29,560 --> 00:23:35,880 agent within its architecture, it is composed of the GPT series 328 00:23:35,880 --> 00:23:40,200 of models from Open AI. So GBD 3, GBD 3.5, GBD 4, soon 329 00:23:40,200 --> 00:23:46,320 to be GBD 4.5 and maybe five. I'm a simple user. 330 00:23:46,320 --> 00:23:51,120 In this case, you're a complex user, OK, You and I can both 331 00:23:51,200 --> 00:23:58,080 interface with the same agent and it can sufficiently respond 332 00:23:58,080 --> 00:24:01,240 to our level of request. It does that by, like I said, 333 00:24:01,240 --> 00:24:06,560 calling on a different model or set of assets, AI services 334 00:24:06,920 --> 00:24:10,560 that's going to allow it to provide a sufficient level of of 335 00:24:11,000 --> 00:24:15,000 of inference or response. So I submit my simple request to 336 00:24:15,000 --> 00:24:19,120 this agent that gets decomposed by the agents back in and 337 00:24:19,120 --> 00:24:21,760 optimally rooted to the model that will sufficiently handle 338 00:24:21,760 --> 00:24:24,560 that simple request, GPT 3, right? 339 00:24:25,320 --> 00:24:28,160 You on the other hand, you submit a multimodal request to 340 00:24:28,160 --> 00:24:31,400 the agent that has to go to a model that can handle the multi 341 00:24:31,400 --> 00:24:34,760 modality of your request. So in this case GPT 4. 342 00:24:35,840 --> 00:24:41,560 The cost difference between invoking GPT 3 and GPT 4 is like 343 00:24:41,560 --> 00:24:44,520 an order of magnitude if not more indifference. 344 00:24:45,520 --> 00:24:48,600 And accounting for that, especially with most 345 00:24:48,600 --> 00:24:52,960 contemporary pricing solutions is non trivial. 346 00:24:53,120 --> 00:24:58,200 It's relatively complicated. So Stripe, which is what a lot 347 00:24:58,200 --> 00:25:01,760 of people turn to when they go to commercialize their AIS and 348 00:25:01,760 --> 00:25:06,040 their AI agents. It's a it's a skew based 349 00:25:06,040 --> 00:25:10,680 architecture, right? You price per skew SKU. 350 00:25:11,160 --> 00:25:13,400 It's set up to sell T-shirts on the Internet. 351 00:25:13,400 --> 00:25:16,000 So I set the price of my small T-shirt. 352 00:25:16,240 --> 00:25:21,080 That price doesn't change from one day to the next versus an AI 353 00:25:21,080 --> 00:25:28,440 agent. It's cost is variable depending 354 00:25:28,440 --> 00:25:33,320 on the complexity of the request that's served to it as well as 355 00:25:33,320 --> 00:25:38,360 the the tools that it has at its disposal to respond to said 356 00:25:38,360 --> 00:25:43,440 request. So an agent can take in dynamic 357 00:25:43,440 --> 00:25:47,680 requests and respond to those dynamic requests in a variable 358 00:25:47,680 --> 00:25:52,160 fashion by invoking a variable set of services which 359 00:25:52,160 --> 00:25:56,120 correspondingly have a variable set of costs to them. 360 00:25:56,440 --> 00:26:01,320 So what we've built is a system of unit accounting, effectively 361 00:26:01,320 --> 00:26:05,840 an accounting module that straps onto an agent's observability 362 00:26:05,840 --> 00:26:11,400 function and translates the metered cost of that service to 363 00:26:11,400 --> 00:26:14,920 provide that inference into a settlement cost for the 364 00:26:14,920 --> 00:26:19,720 requester for you and I depending on the variable 365 00:26:20,360 --> 00:26:24,640 response and and invocation of of the corresponding services on 366 00:26:24,640 --> 00:26:28,640 the back end. That sounds very prescribed. 367 00:26:29,080 --> 00:26:33,440 So as someone, for instance, if I were to offer different AI 368 00:26:33,440 --> 00:26:36,640 models, I mean, I, I could come up with different pricing 369 00:26:36,640 --> 00:26:39,080 strategies, right? Kind of I, I could kind of sell 370 00:26:39,680 --> 00:26:43,320 a flat rate to all my models. I could say, OK, I'll give you a 371 00:26:43,320 --> 00:26:50,800 flat rate until kind of you hit a certain number of requests and 372 00:26:50,800 --> 00:26:54,720 then kind of you have to pay per request or kind of I, I, I mean, 373 00:26:54,840 --> 00:26:58,560 there's different strategies. So how much of that do you 374 00:26:58,560 --> 00:27:02,840 impose on people and how much can you actually tailor the the 375 00:27:03,000 --> 00:27:07,880 these this pricing strategy? So this is where again, like 376 00:27:08,440 --> 00:27:13,880 you're clearly like well versed in, in the nuance of this. 377 00:27:15,520 --> 00:27:18,760 What we're what we are trying to accommodate for is as much 378 00:27:18,760 --> 00:27:23,040 variability in that price control setting mechanism as 379 00:27:23,040 --> 00:27:26,280 possible. So if you want to have sort of 380 00:27:27,800 --> 00:27:34,640 a, a fixed price subscription that maybe or may not rate limit 381 00:27:34,800 --> 00:27:38,240 or time limit a service, you can do that. 382 00:27:38,240 --> 00:27:44,240 If you want to go as granular as like pure pay to play and you 383 00:27:44,240 --> 00:27:49,840 know each, each access is cost this or or each GPU cycle for 384 00:27:49,840 --> 00:27:54,080 that matter costs X, you can do that with this system. 385 00:27:54,720 --> 00:28:01,200 What I mean that part of the rationale is we want to provide 386 00:28:01,200 --> 00:28:06,040 that flexibility. The other is we are in the 387 00:28:06,040 --> 00:28:11,080 process of discovering what is like sort of the dominant set of 388 00:28:11,080 --> 00:28:15,280 attributes and characteristics for these pricing mechanisms 389 00:28:15,280 --> 00:28:20,720 within the agentic landscape. Because the reality is this 390 00:28:20,720 --> 00:28:25,840 stuff is all quite new and we don't know what the dominant 391 00:28:26,240 --> 00:28:34,200 system for, for costing and billing wrapped up in some 392 00:28:34,200 --> 00:28:37,200 pricing component is actually going to be yet. 393 00:28:37,800 --> 00:28:42,560 So we are trying to provide as much variability. 394 00:28:42,560 --> 00:28:47,640 So basically in an intrude decentralized fashion, we 395 00:28:47,640 --> 00:28:51,120 attempt to give that control to the user as opposed or the 396 00:28:51,120 --> 00:28:53,640 builder as opposed to setting it for them upfront. 397 00:28:55,160 --> 00:28:57,280 OK. And then kind of I, I understand 398 00:28:57,280 --> 00:29:00,920 that kind of like you give me variability and kind of like how 399 00:29:00,920 --> 00:29:05,200 I set my pricing strategy, but do you, you, I, you also take 400 00:29:05,200 --> 00:29:07,480 care of settlement, right? So kind of you also make sure 401 00:29:07,480 --> 00:29:09,800 that I actually get paid. Right. 402 00:29:10,680 --> 00:29:12,440 Yeah. And how, how do you do that? 403 00:29:12,440 --> 00:29:15,160 So kind of like how how does how does the payment work? 404 00:29:15,160 --> 00:29:16,960 Yeah. So this is a good question too. 405 00:29:17,520 --> 00:29:25,800 So we, we leverage a concept called license tokenization. 406 00:29:26,760 --> 00:29:33,480 So we believe quite strongly. And, and so again, like going 407 00:29:33,480 --> 00:29:36,600 back to the, this is a function of a marketplace, right? 408 00:29:37,320 --> 00:29:40,920 Marketplaces are hard to win, especially as you get as, 409 00:29:41,240 --> 00:29:46,960 especially as you start to sell assets that become more and more 410 00:29:46,960 --> 00:29:50,560 commoditized, right? Like trying to markets, trying 411 00:29:50,560 --> 00:29:54,400 to push the price to 0. So and then you're, your, your 412 00:29:54,400 --> 00:29:57,480 revenues generated at the margin, right? 413 00:29:57,720 --> 00:30:02,480 So being is having as much fidelity on the actual 414 00:30:02,480 --> 00:30:06,200 operational cost means that you can like eke out as much margin 415 00:30:06,200 --> 00:30:08,920 as possible. So like recognizing that that's 416 00:30:08,920 --> 00:30:16,400 sort of the this the set of, I don't know, maximal constraints 417 00:30:16,640 --> 00:30:20,640 that we have to deal with here. We want to enable these agents 418 00:30:20,880 --> 00:30:24,960 to eke out as slim a margin as possible and and also allow them 419 00:30:24,960 --> 00:30:27,680 to continue to be functional from a business and operational 420 00:30:27,680 --> 00:30:34,360 point of view. So that means setting a single 421 00:30:34,360 --> 00:30:40,000 price and just giving free rein access doesn't really make a lot 422 00:30:40,000 --> 00:30:42,760 of sense. It works right now for 423 00:30:42,760 --> 00:30:47,120 propositions that are broadly speaking toys. 424 00:30:47,840 --> 00:30:51,800 And then more importantly there were there's not a a lot of 425 00:30:52,120 --> 00:30:55,760 competition that's yet pushing that price point down. 426 00:30:55,760 --> 00:30:59,320 But then we are already seeing this manifest, right? 427 00:30:59,320 --> 00:31:07,680 Like Open AI very clearly has, you know, a price per usage 428 00:31:07,680 --> 00:31:09,640 function. And then a deep sea comes along 429 00:31:09,640 --> 00:31:10,840 and has like an open source model. 430 00:31:10,840 --> 00:31:13,160 And it's like, here you go have it for free, right? 431 00:31:14,280 --> 00:31:16,880 Trying to find the equilibrium between the two that that's 432 00:31:16,880 --> 00:31:19,920 somewhere in between, right? It's not free. 433 00:31:19,920 --> 00:31:24,320 It's also probably not maybe as price gougey as some of the 434 00:31:24,320 --> 00:31:29,520 things that Open AI is doing. So anyway, in all of this work 435 00:31:29,520 --> 00:31:32,520 that we've done, there's this recognition that actually 436 00:31:32,520 --> 00:31:36,880 understanding sort of the ML OPS that that the observability 437 00:31:36,880 --> 00:31:39,640 piece of translating that metered cost into a settlement 438 00:31:39,640 --> 00:31:42,760 cost, it's likely going to be important, especially as the 439 00:31:43,080 --> 00:31:46,840 these AI agents and their services get commoditized. 440 00:31:47,640 --> 00:31:50,960 And so we kind of looked around and said, OK, like if we're 441 00:31:50,960 --> 00:31:55,080 taking the position that like a traditional subscription style 442 00:31:55,080 --> 00:31:57,640 model isn't the right one, what is? 443 00:31:59,680 --> 00:32:03,120 What does look and feel right based on experience? 444 00:32:03,520 --> 00:32:05,560 And it's this concept of license tokenization. 445 00:32:05,560 --> 00:32:07,080 So it has nothing to do with crypto. 446 00:32:08,520 --> 00:32:12,560 It's a traditional licensing scheme, but it differs in 447 00:32:12,560 --> 00:32:16,160 comparison to like named user licensing and concurrent access 448 00:32:16,160 --> 00:32:21,920 licensing, where in a named user license, Federica, you negotiate 449 00:32:21,920 --> 00:32:26,560 your usage for a platform and the underlying sets of tools 450 00:32:26,560 --> 00:32:30,000 within that platform, right? Concurrent access license would 451 00:32:30,000 --> 00:32:32,640 be you and I are on a team together. 452 00:32:32,640 --> 00:32:37,160 We negotiate our usage of said platform and the corresponding 453 00:32:37,160 --> 00:32:40,360 tools of that platform. It's a pretty laborious process 454 00:32:40,360 --> 00:32:43,480 in the grand scheme of things or very rigid, right? 455 00:32:43,680 --> 00:32:48,640 One of the two either like everybody gets the same thing or 456 00:32:48,680 --> 00:32:51,960 it takes a long time to negotiate what you get, right. 457 00:32:52,720 --> 00:32:55,480 The response to this is this thing is this concept called 458 00:32:55,480 --> 00:33:02,560 license tokenization, where the the platform is tokenized and 459 00:33:02,560 --> 00:33:07,120 the tools that make up that platform have redemption 460 00:33:07,120 --> 00:33:14,400 criteria in those tokens. So you buy 1000 tokens, I buy 461 00:33:14,480 --> 00:33:22,880 10,000 tokens and that platform is made-up of tool A B&C and a 462 00:33:22,880 --> 00:33:29,080 has redemption criteria of 100 tokens BA 1000 tokens C5000. 463 00:33:29,080 --> 00:33:30,840 Token usage credits or something? 464 00:33:30,920 --> 00:33:34,320 It's use, it's exact. That's exactly, it's this, yeah. 465 00:33:35,960 --> 00:33:42,880 So it's this emergent licensing model we've taken that looked at 466 00:33:42,880 --> 00:33:46,000 agents. They are again, this form factor 467 00:33:46,000 --> 00:33:49,720 of a platform compilation of a bunch of different tools where 468 00:33:50,040 --> 00:33:54,960 you can issue tokens, or in our case, we call them credits for 469 00:33:54,960 --> 00:34:00,160 each of these agents or set swarm of agents and the tools 470 00:34:00,160 --> 00:34:06,160 and, or agents within that that agent is composed of or those 471 00:34:06,160 --> 00:34:09,400 sets of agents of swarms of agents are composed of. 472 00:34:09,639 --> 00:34:13,480 They can have their own redemption criteria in those 473 00:34:13,480 --> 00:34:18,239 credits or those tokens. So that's how we facilitate sort 474 00:34:18,239 --> 00:34:28,159 of that the, the legibility and and the the, the fine grain 475 00:34:29,320 --> 00:34:32,679 component of the, the payment aspect. 476 00:34:32,679 --> 00:34:36,120 And how do you set it so because kind of like you have, you then 477 00:34:36,120 --> 00:34:38,320 have to transfer them, right? Right. 478 00:34:38,320 --> 00:34:43,520 So there's so like in this scenario where you and I are the 479 00:34:43,520 --> 00:34:46,960 user of this third party set of credits, we, let's say we pay a 480 00:34:46,960 --> 00:34:50,520 dollar and we each get 1000 credits. 481 00:34:50,639 --> 00:34:56,760 And so under the hood of this agent, GPT 3 has a redemption 482 00:34:56,760 --> 00:35:02,640 criteria of 10 credits. GPT 4 A3 3.5 of 100 GPT 4400 483 00:35:02,640 --> 00:35:06,320 credits, right? We both pay a dollar. 484 00:35:06,600 --> 00:35:10,160 I get 1000 credits in my wallet. You get 1000 credits in your 485 00:35:10,160 --> 00:35:11,960 wallet. We make these requests to this 486 00:35:11,960 --> 00:35:15,440 agent. My simple request goes to GPD 3 487 00:35:15,720 --> 00:35:19,720 out of my pool of 1000 credits, I get charged 10. 488 00:35:20,360 --> 00:35:25,560 You with your multimodal request that gets routed to GPD 4, you 489 00:35:25,560 --> 00:35:28,720 get charged 400 out of your pool of 1000. 490 00:35:29,320 --> 00:35:32,600 So that's how. And then that function, that 491 00:35:32,600 --> 00:35:37,120 accounting, that redemption is a burn function on chain. 492 00:35:38,440 --> 00:35:43,760 OK, so it it's a burn function and that, but then how does the 493 00:35:43,960 --> 00:35:48,320 the AI that actually did the work receive the payment? 494 00:35:49,280 --> 00:35:52,400 Good question. So we have basically 2 forms of 495 00:35:52,400 --> 00:35:57,400 settlement. One is the the piece that that 496 00:35:57,400 --> 00:36:02,280 the settlement that authorizes you to use the system. 497 00:36:02,840 --> 00:36:10,080 So the payment of a dollar for that 1000 credits, that's the 498 00:36:10,080 --> 00:36:12,520 first settlement. And in our case, you know, 499 00:36:12,520 --> 00:36:18,640 recognizing that there's a large swath of like we, we view this 500 00:36:18,640 --> 00:36:23,200 as an AI solution or adjacent solution. 501 00:36:23,480 --> 00:36:32,520 So we don't distinguish between like Web 3 AI versus Web 2 AI is 502 00:36:32,520 --> 00:36:34,480 just AI. What we're trying to build is 503 00:36:34,480 --> 00:36:36,480 something that's like general purpose for AI. 504 00:36:37,320 --> 00:36:43,920 Recognizing that there is a large swath maybe if not a 505 00:36:43,920 --> 00:36:49,840 majority, like a, there's, yeah, probably a majority of AI, that 506 00:36:50,360 --> 00:36:54,160 of the AI community that is not very conversant in Web 3. 507 00:36:55,200 --> 00:36:57,800 So one of the things that we've done is like full account 508 00:36:57,800 --> 00:37:00,840 abstraction. You still get a wallet, right? 509 00:37:00,840 --> 00:37:03,760 And the agent still gets a wallet, but you can use socials 510 00:37:03,760 --> 00:37:06,000 to set it up. It's an MPC solution. 511 00:37:06,400 --> 00:37:12,360 It's fully gas less. So there's no like extraneous 512 00:37:12,880 --> 00:37:16,680 signature signing and stuff like that that that's required for 513 00:37:16,680 --> 00:37:19,080 anybody that's listening. We're paying for the gas right 514 00:37:19,080 --> 00:37:21,440 now if there's questions around that. 515 00:37:21,800 --> 00:37:24,920 But anyway, so in this case, in this scenario that that I was 516 00:37:24,920 --> 00:37:27,520 describing where I'm the simple user and you're the power user, 517 00:37:28,280 --> 00:37:32,160 I don't have any affinity towards web three. 518 00:37:32,240 --> 00:37:35,160 I don't have a wallet. I just want to use to say I OK, 519 00:37:35,160 --> 00:37:37,040 I use nevermind. I go through this checkout. 520 00:37:37,040 --> 00:37:44,240 Part of the checkout process is I register with the system that 521 00:37:44,240 --> 00:37:48,000 creates my wallet. We've gone so far as to 522 00:37:48,000 --> 00:37:51,240 integrate Stripe. So I'm now in the ecosystem of 523 00:37:51,240 --> 00:37:53,200 registered. I've created this MPC based 524 00:37:53,200 --> 00:37:56,280 wallet that's attached to me. If I don't know where to look, I 525 00:37:56,280 --> 00:37:58,320 don't even know that it's, you know, a crypto wallet. 526 00:37:58,880 --> 00:38:03,560 And then I can just take out my debit card or credit card and 527 00:38:03,560 --> 00:38:09,000 pay a dollar for these thousand for these 1000 credits. 528 00:38:09,000 --> 00:38:12,440 Now in the background, the builder that's registered this 529 00:38:12,440 --> 00:38:15,040 agent, this third party agent that you and I are using, 530 00:38:15,560 --> 00:38:18,880 they've linked that to their bank account through a Stripe 531 00:38:18,880 --> 00:38:21,920 integration. What they've also done is linked 532 00:38:21,920 --> 00:38:26,040 that to a wallet because in this case, in this scenario, the 533 00:38:26,040 --> 00:38:28,880 builder is going to take both Fiat and crypto as payment. 534 00:38:29,320 --> 00:38:34,760 So I pay my $1.00 with my debit card back goes to the the 535 00:38:34,760 --> 00:38:37,600 builder's bank account. You on the other hand, you're 536 00:38:39,080 --> 00:38:44,160 well versed in crypto, you have a wallet, you've got USDC, so 537 00:38:44,160 --> 00:38:49,200 you pay 1 USDC and that goes into the, the agent or, or the 538 00:38:49,200 --> 00:38:54,080 builder's wallet in that case. And so in this case we're, we're 539 00:38:54,080 --> 00:38:55,960 handling both. But as you can see, there's two 540 00:38:55,960 --> 00:39:00,520 forms of settlement. 1 is this sort of overarching 541 00:39:00,960 --> 00:39:05,040 authentication gatekeeping function for access to the agent 542 00:39:05,720 --> 00:39:10,520 that gives you the set of credits or tokens the the usage 543 00:39:11,200 --> 00:39:15,400 asset to start utilizing and authorize. 544 00:39:17,280 --> 00:39:19,760 I mean, I you know and authenticate the usage of of 545 00:39:20,080 --> 00:39:24,880 that agent. How do I as a user know that 546 00:39:24,880 --> 00:39:31,840 kind of algorithms that I solicit, How how do I know that 547 00:39:31,840 --> 00:39:35,160 they are metered in a fairway? So kind of like if there's 548 00:39:35,160 --> 00:39:39,080 different ways that kind of like different, different underlying 549 00:39:39,080 --> 00:39:42,120 functions that I could call in kind of they're all metered in 550 00:39:42,240 --> 00:39:45,720 some way. Is, is there like some rubber 551 00:39:45,720 --> 00:39:49,160 stamp of approval somewhere that says, OK, this is actually this 552 00:39:49,160 --> 00:39:50,680 is, this is an OK pricing scheme? 553 00:39:50,680 --> 00:39:53,520 Because kind of like I could, I could kind of like make a really 554 00:39:53,520 --> 00:39:58,240 obscure pricing scheme where kind of I overcharge massively 555 00:39:58,240 --> 00:40:01,600 for certain parts because it's somewhat intransparent to to the 556 00:40:01,600 --> 00:40:04,240 user, right? Yeah. 557 00:40:04,520 --> 00:40:12,920 So right now it's it's the very finger in the air and cottage 558 00:40:12,920 --> 00:40:22,400 industry they're they're there's a lot of price discovery going 559 00:40:22,400 --> 00:40:25,640 on at the moment, a lot of like guesstimation. 560 00:40:26,120 --> 00:40:29,160 We're actually working on something that we hope is going 561 00:40:29,160 --> 00:40:34,920 to help both with the price setting piece as well As for an 562 00:40:34,920 --> 00:40:38,920 understanding from a user base point of view what maybe the 563 00:40:38,920 --> 00:40:42,720 pricing should be for a given agent. 564 00:40:44,400 --> 00:40:50,440 So that's, that will be, it's, it's like a, a pricing engine. 565 00:40:50,440 --> 00:40:55,160 So that's going to come down the, the pipe relatively soon, 566 00:40:56,080 --> 00:40:59,160 but it's something conceptually that we've been working on for 567 00:40:59,160 --> 00:41:05,240 about 6 months. And over the last two months 568 00:41:05,240 --> 00:41:10,440 we've like put pen to paper and actually started to, we, we POC 569 00:41:10,440 --> 00:41:12,720 did. And now it looks like we can 570 00:41:12,720 --> 00:41:15,680 actually do what we want to accomplish to kind of address 571 00:41:15,960 --> 00:41:19,440 not the buy side, but more the sell side to start. 572 00:41:20,160 --> 00:41:22,920 Because the other, the, the, the flip side of your question is 573 00:41:22,920 --> 00:41:29,240 what do I price this at, right? So helping answer that question 574 00:41:29,400 --> 00:41:31,240 is where we're trying to get to 1st. 575 00:41:31,760 --> 00:41:36,160 And I think the knock on effect of that will be disclosing that 576 00:41:36,160 --> 00:41:44,840 sort of price setting mechanism will help those on the buy side 577 00:41:45,600 --> 00:41:49,600 also understand maybe what their cost structure should be. 578 00:41:50,760 --> 00:41:52,800 Maybe that's kind of switch gears a little bit. 579 00:41:53,000 --> 00:41:57,120 So all of this kind of is built on blockchain infrastructure 580 00:41:57,120 --> 00:42:02,920 kind of like walk us through kind of like what kind of stack 581 00:42:02,920 --> 00:42:08,680 this is built on and why you chose that stack so. 582 00:42:10,080 --> 00:42:15,800 We're at adapt in the classical sense, so we're chain agnostic 583 00:42:15,800 --> 00:42:19,360 though we are. We're an EVM based solution. 584 00:42:21,280 --> 00:42:26,320 So you know, from a deployment point of view, we're on mainnet 585 00:42:26,320 --> 00:42:32,920 and Polygon and Arbitrum and bass and Nosis and cello and 586 00:42:33,080 --> 00:42:36,440 bunch of EVM based chainsaw ones and L twos. 587 00:42:37,840 --> 00:42:42,680 Yeah, the the code base is Python And TypeScript. 588 00:42:44,520 --> 00:42:53,600 You know where needed. We've got backends that are what 589 00:42:54,000 --> 00:42:59,440 do we have? I don't know, it's post Christ 590 00:42:59,440 --> 00:43:05,040 database. Yeah, I mean it's relatively 591 00:43:05,040 --> 00:43:07,640 run-of-the-mill from the architecture. 592 00:43:07,920 --> 00:43:11,280 Point of view, OK, let's talk about kind of like the 593 00:43:11,280 --> 00:43:13,000 interoperability aspects here, right? 594 00:43:13,000 --> 00:43:16,760 So kind of like say I as a user kind of come to your dab, how, 595 00:43:16,760 --> 00:43:22,320 how do you determine kind of like which of which of these 596 00:43:23,880 --> 00:43:28,440 chains I kind of buy my credits on and settle on and so on? 597 00:43:28,440 --> 00:43:30,480 How, how is that? How is that determined? 598 00:43:30,760 --> 00:43:33,480 Because in principle, that's something that the user probably 599 00:43:33,480 --> 00:43:39,040 doesn't care about, right? No, I, I, I, I disagree with 600 00:43:39,040 --> 00:43:42,200 that statement when you're talking about AI Web 3 builders 601 00:43:42,200 --> 00:43:45,520 'cause they usually have a network that they want to 602 00:43:45,520 --> 00:43:48,200 default to. OK, but then let's let's talk 603 00:43:48,200 --> 00:43:50,640 about kind of like the the people who kind of consume, 604 00:43:50,760 --> 00:43:54,440 right, Who kind of consume your. AI, well, they don't care. 605 00:43:55,160 --> 00:43:56,600 Yeah, they don't care at all, right? 606 00:43:57,000 --> 00:44:02,800 So as someone who wants to consume, how do I decide which 607 00:44:03,040 --> 00:44:07,520 network to kind of consume my to which network to pay for? 608 00:44:07,520 --> 00:44:10,760 My in this case, you don't the the builder would. 609 00:44:11,280 --> 00:44:16,000 So OK, so here's where like from a crypto point of view, you run 610 00:44:16,000 --> 00:44:18,680 into friction. But again, the choice is up to 611 00:44:18,880 --> 00:44:24,400 the agent or the, you know, the builder to decide which chain or 612 00:44:24,400 --> 00:44:29,920 chains this thing the agent is anchored to right is connected 613 00:44:29,920 --> 00:44:33,680 to. Usually it's one the dominant 614 00:44:33,680 --> 00:44:40,640 chain at the moment is base, so that's the default. 615 00:44:43,240 --> 00:44:47,840 From a consumption point of view, you don't really care 616 00:44:47,920 --> 00:44:55,240 other than if you are in your case, this power user that is 617 00:44:55,440 --> 00:45:01,360 crypto native or savvy, you know, you're in a condition 618 00:45:01,360 --> 00:45:08,160 where shit, I don't have any USDC in a wallet on base. 619 00:45:09,120 --> 00:45:13,120 So now you're in and you want to use an agent that's anchored 620 00:45:13,120 --> 00:45:16,960 there and it needs, you know, it's, you got to pay 1 USDC for 621 00:45:16,960 --> 00:45:19,840 those 1000 credits. Well, now you've got to bridge 622 00:45:19,840 --> 00:45:24,920 that. That's outside of the scope of 623 00:45:25,080 --> 00:45:30,520 of our operation. You know, there's, I think 624 00:45:30,520 --> 00:45:33,640 there's enough bridging tools out there that you could 625 00:45:33,640 --> 00:45:36,080 probably figure it out if you need to bridge. 626 00:45:36,080 --> 00:45:38,600 OK. So basically as a consumer, kind 627 00:45:38,600 --> 00:45:43,160 of like I decide what model I want to use, what kind of agent 628 00:45:43,160 --> 00:45:48,960 I want to frequent and then kind of I I just have to pay on the 629 00:45:48,960 --> 00:45:51,360 commensurate chain. Is that fair? 630 00:45:51,400 --> 00:45:54,880 Yeah, exactly. Yeah, Yeah, that's that's the 631 00:45:54,880 --> 00:45:57,160 way it would work, yes. OK. 632 00:45:57,160 --> 00:46:03,800 So how, how does Nevermind currently integrate with other 633 00:46:03,800 --> 00:46:06,800 kind of decentralized platforms or protocols, right? 634 00:46:06,800 --> 00:46:11,360 Because kind of like you primarily enable the, the 635 00:46:11,360 --> 00:46:15,360 payments here, there's a lot of functionality that kind of that 636 00:46:15,400 --> 00:46:18,720 that has to come together to kind of make this into a good 637 00:46:18,720 --> 00:46:21,640 user experience that kind of goes beyond payment, right? 638 00:46:21,880 --> 00:46:24,560 So how, how, how do you interoperate here? 639 00:46:25,560 --> 00:46:28,680 Yeah. So we kind of have, we have 3 640 00:46:29,520 --> 00:46:41,320 levels of engagement. So there's the SDK, which is the 641 00:46:41,320 --> 00:46:43,440 most robust. It provides the most set of 642 00:46:43,440 --> 00:46:51,800 features for integration. You know, that's if you're like 643 00:46:51,840 --> 00:47:01,280 a pretty serious Web 3 builder. Moonlighting is an AI developer. 644 00:47:01,320 --> 00:47:07,360 You're probably like, you might gravitate towards that on the 645 00:47:07,800 --> 00:47:10,440 those that are gravitate more towards the AI side that are 646 00:47:10,440 --> 00:47:16,000 less familiar with like the the, the full suite of capabilities 647 00:47:16,000 --> 00:47:18,560 from a blockchain point of view, they're going to use the 648 00:47:18,560 --> 00:47:23,400 libraries that we have on offer. So we've got the SDK then on top 649 00:47:23,600 --> 00:47:30,040 a more refined set of libraries and Python based ones, 'cause 650 00:47:30,040 --> 00:47:33,720 it's the dominant language for building a is. 651 00:47:34,440 --> 00:47:37,640 And then on top of that, for anybody that's kind of like 652 00:47:37,880 --> 00:47:40,800 there's this new subset of builder, right? 653 00:47:41,960 --> 00:47:47,240 This like non-technical or let's call it pseudo technical. 654 00:47:48,760 --> 00:47:51,840 They can build, you know, there's, there's emergent tools, 655 00:47:51,840 --> 00:47:56,200 especially on the model side where you can like prompt 656 00:47:56,200 --> 00:48:01,400 engineer a relatively sophisticated agent. 657 00:48:02,400 --> 00:48:05,240 And now there's tooling that's coming out that makes building 658 00:48:05,320 --> 00:48:13,960 agents even easier for that, for that demographic. 659 00:48:13,960 --> 00:48:18,440 We have an app which is an even more refined set of, of 660 00:48:18,880 --> 00:48:22,760 functionality. And so that's, yeah, they're, 661 00:48:22,760 --> 00:48:27,360 those are the three kind of mechanisms or means of engaging 662 00:48:27,360 --> 00:48:32,000 with what we built. What's the value proposition 663 00:48:32,280 --> 00:48:36,320 that kind of you put forward to each of these groups? 664 00:48:36,320 --> 00:48:40,080 So kind of what, why shouldn't they kind of just buy kind of 665 00:48:40,080 --> 00:48:44,880 credits with open AI or cloud or kind of use deep seek for free? 666 00:48:47,760 --> 00:48:51,160 OK. I would say this, well, even in 667 00:48:51,160 --> 00:48:56,560 the case where all of these services are cost nothing and 668 00:48:56,560 --> 00:49:00,080 that's just long term probably untenable unless they become, 669 00:49:00,440 --> 00:49:08,240 you know public goods, which I think most of these companies 670 00:49:08,240 --> 00:49:10,080 will fight tooth and nail against. 671 00:49:10,080 --> 00:49:16,480 But well, let's see how it plays out, you know, barring that from 672 00:49:16,480 --> 00:49:19,600 happening. But even if that does occur, 673 00:49:21,480 --> 00:49:27,640 there's still sort of the aggregate that these agents 674 00:49:27,640 --> 00:49:40,000 represent any tuned expertise that can be captured and then 675 00:49:40,000 --> 00:49:48,080 subsequently deployed in these packaged agentic services that 676 00:49:48,080 --> 00:49:55,360 any of themselves can be can be priced and paid for. 677 00:49:57,160 --> 00:50:02,480 So going back to your question, like what, which, which solution 678 00:50:02,480 --> 00:50:05,640 would you gravitate towards? Like it splice around, like 679 00:50:05,640 --> 00:50:09,920 what's the pricing and payment mechanism and more, what's the 680 00:50:09,920 --> 00:50:15,040 level of functionality that you want to have within your, your 681 00:50:15,040 --> 00:50:17,840 agent and or your and, or your swarm? 682 00:50:19,240 --> 00:50:26,920 So for example, we have in our SDK the like the attribution 683 00:50:26,920 --> 00:50:29,240 function. So once payment has occurred, 684 00:50:29,240 --> 00:50:34,120 like if you have a swarm of agents and that swarm is 685 00:50:34,120 --> 00:50:37,360 ultimately what's priced, that you can actually like 686 00:50:37,360 --> 00:50:44,360 redistribute funds within the, the commercial, the, the, the 687 00:50:44,360 --> 00:50:48,360 value capture piece, right? And redistribute those amongst 688 00:50:48,360 --> 00:50:51,280 the agents proportionately to their contribution within that 689 00:50:51,280 --> 00:50:55,280 swarm. But that's like super low level 690 00:50:55,320 --> 00:50:58,960 and really only going to be interested interesting to 691 00:50:58,960 --> 00:51:02,960 somebody that's been working on swarms for a long time, right? 692 00:51:03,560 --> 00:51:11,040 So like that functionality in the SDK is not really being used 693 00:51:11,040 --> 00:51:12,880 at present. What most people are just trying 694 00:51:12,880 --> 00:51:16,480 to do is wrap their Http://endpoint in some sort of 695 00:51:16,480 --> 00:51:23,120 gatekeeping functionality with a payment mechanism attached as 696 00:51:23,120 --> 00:51:27,280 part of that gatekeeping functionality. 697 00:51:28,320 --> 00:51:34,400 So there's a broad spectrum of requirements and demands. 698 00:51:37,360 --> 00:51:41,280 We're trying to cater to the simplest set of those demands. 699 00:51:41,280 --> 00:51:45,720 Initially, though, we have built in some relatively complex 700 00:51:45,720 --> 00:51:48,920 functionality. Just, you know, I don't know, 701 00:51:48,920 --> 00:51:52,400 because it's interesting. OK. 702 00:51:52,400 --> 00:51:55,400 So maybe let me reframe the question a little bit. 703 00:51:56,000 --> 00:52:06,280 So with never mind how, how is the AI data payments landscape 704 00:52:06,720 --> 00:52:12,120 becoming better for actual users, consumers of data or 705 00:52:12,120 --> 00:52:14,720 providers of data and algorithms? 706 00:52:15,480 --> 00:52:19,480 How is it becoming better to what we currently have in kind 707 00:52:19,480 --> 00:52:21,640 of like this, this centralized model? 708 00:52:23,280 --> 00:52:25,920 I'm going to answer this in a relatively flippant way. 709 00:52:26,840 --> 00:52:39,600 We don't actually care what we're trying to offer and enable 710 00:52:41,120 --> 00:52:48,840 is a higher degree of fidelity on that transactional piece. 711 00:52:49,920 --> 00:52:55,520 So weather and and what I mean like there's two aspects of 712 00:52:55,520 --> 00:52:58,960 this. One is on the accounting piece, 713 00:52:59,080 --> 00:53:01,760 right. So enabling that and doing it in 714 00:53:01,760 --> 00:53:05,920 a way that's more dynamic than existing systems today. 715 00:53:07,200 --> 00:53:10,920 And then the other, and this is whether or not over the long 716 00:53:10,920 --> 00:53:14,160 term this actually matters in time will tell. 717 00:53:14,920 --> 00:53:21,040 But like these services that are being rendered discreetly don't 718 00:53:21,040 --> 00:53:27,360 cost that much. They cost like fractions of a 719 00:53:27,360 --> 00:53:34,440 fraction of a cent. Existing payment systems, Fiat 720 00:53:34,440 --> 00:53:40,440 based payment systems cannot handle that discrete mechanism. 721 00:53:41,480 --> 00:53:47,120 You you can't get below, you know a certain denomination of a 722 00:53:47,120 --> 00:53:49,920 currency, say 1 cent, right? Why not? 723 00:53:49,920 --> 00:53:52,440 It just depends on how often you settle right? 724 00:53:53,240 --> 00:53:55,720 Kind of like if you have fractions of the send, then kind 725 00:53:55,720 --> 00:53:57,280 of like. This is what I'm saying from a, 726 00:53:57,360 --> 00:54:01,400 from a, from a very discreet point of view. 727 00:54:01,920 --> 00:54:07,240 If, if all I need is 1 action, just for sake of argument, one 728 00:54:07,240 --> 00:54:12,360 GPU cycle, I can't price at it. I, I have to do what you just 729 00:54:12,360 --> 00:54:17,520 said, I have to aggregate it. And so the question becomes, and 730 00:54:17,520 --> 00:54:25,600 This is why I said the jury's still out on this piece, is that 731 00:54:27,040 --> 00:54:31,920 actually a requirement or not? Time will tell. 732 00:54:32,400 --> 00:54:38,280 But I, you know, I do believe that there will be applications 733 00:54:38,280 --> 00:54:44,960 on a use case where you have discrete expertise that for that 734 00:54:44,960 --> 00:54:50,880 particular use case, for that set of requesters, that 735 00:54:51,040 --> 00:54:57,640 particular agent may only ever get called upon once or less 736 00:54:57,640 --> 00:55:03,200 times than you can actually aggregate to the, the floor of 737 00:55:03,200 --> 00:55:06,600 that currency. And So what do you do in that 738 00:55:06,600 --> 00:55:11,080 case? That service is always free or 739 00:55:11,080 --> 00:55:13,880 it has to be coupled with other services. 740 00:55:13,880 --> 00:55:21,800 And you know, so again, there's an element here of of, you know, 741 00:55:21,800 --> 00:55:25,200 speculation on whether that's going to be a driver. 742 00:55:26,160 --> 00:55:31,200 I think it will be having operated in this space for as 743 00:55:31,200 --> 00:55:37,760 long as I have, But whether or not it's a primary driver, I 744 00:55:37,760 --> 00:55:42,640 don't know, time will tell. But anyway, getting back to the 745 00:55:42,640 --> 00:55:46,760 original question, you know, I think at the end of the day, why 746 00:55:46,760 --> 00:55:49,760 are we doing this? Why are we using crypto instead 747 00:55:49,760 --> 00:55:53,200 of doing this in like a centralized multi tenanted 748 00:55:53,200 --> 00:56:00,720 charted database? We're driven by optionality and 749 00:56:00,720 --> 00:56:07,840 like the the drive to provide option and, and and that is 750 00:56:07,840 --> 00:56:15,960 derived from a desire. I mean this is where the sort of 751 00:56:15,960 --> 00:56:21,720 the crypto ethos shines through to, to provide the option of 752 00:56:21,720 --> 00:56:25,680 censorship resistance, right. So from our point of view, we 753 00:56:25,680 --> 00:56:32,840 view the payments piece as the most critical component in 754 00:56:32,840 --> 00:56:40,960 decentralization for AI agents. So if like Microsoft Open AI, 755 00:56:40,960 --> 00:56:45,080 Google, Facebook, Deepseek, whoever, if they monopolize the 756 00:56:45,320 --> 00:56:48,600 the the means for these agents to pay and get paid, their 757 00:56:48,600 --> 00:56:52,720 ability to de platform one agent, in our opinion, is an 758 00:56:52,720 --> 00:56:54,520 existential threat to all agents. 759 00:56:57,600 --> 00:57:03,920 It if if what if that, you know, centralizing entity can just 760 00:57:03,920 --> 00:57:06,480 with the flick of a switch. Microsoft deems this agent 761 00:57:06,480 --> 00:57:10,160 competitive with one of its lines of business. 762 00:57:10,800 --> 00:57:14,480 It doesn't matter if the rest of that agent is decentralized, at 763 00:57:14,880 --> 00:57:18,320 least in an economic context, it might as well not exist because 764 00:57:18,320 --> 00:57:22,560 it can no longer transact. So providing the infrastructure, 765 00:57:22,560 --> 00:57:27,880 the means for these agents, providing that optionality for 766 00:57:27,880 --> 00:57:34,200 them to pay and get paid always is like that's a driver for us. 767 00:57:35,960 --> 00:57:41,040 I, I totally hear that. And I think I, I, I fully 768 00:57:41,040 --> 00:57:47,080 understand that argument kind of like from a defensive 769 00:57:47,080 --> 00:57:50,600 engineering kind of point of view, but kind of the question 770 00:57:50,600 --> 00:57:52,400 is how do you get the flywheel going right? 771 00:57:52,400 --> 00:57:56,360 So kind of I, I, I have, I have no doubt that kind of you find 772 00:57:56,360 --> 00:58:00,480 people who are willing to kind of sell their services on your 773 00:58:00,480 --> 00:58:03,520 marketplace that's current, that's usually kind of the, the, 774 00:58:03,520 --> 00:58:09,240 the easy, the easy side of, of. Kind of of the business. 775 00:58:09,400 --> 00:58:12,160 Yeah, of any business, right, kind of like, but how do you 776 00:58:12,160 --> 00:58:13,200 make sure how do? You find the. 777 00:58:13,200 --> 00:58:17,600 Buyers actually, you know, consumers actually come to to 778 00:58:17,600 --> 00:58:21,200 your interface to kind of buy services there rather than 779 00:58:21,200 --> 00:58:23,320 elsewhere. OK. 780 00:58:23,320 --> 00:58:29,360 So to be flipping again, we don't care. 781 00:58:29,880 --> 00:58:34,360 That's up to the agent to provide a productive service 782 00:58:34,360 --> 00:58:39,840 that's somebody or something actually wants to use that's out 783 00:58:39,840 --> 00:58:45,360 of our sphere of influence. Now kind of generalizing it's in 784 00:58:45,360 --> 00:58:50,000 our sphere of influence by selecting who we partner with. 785 00:58:51,080 --> 00:58:55,320 So making sure that we're there's two things that we need 786 00:58:55,320 --> 00:58:57,880 to make sure of. One is to your point, we need to 787 00:58:57,880 --> 00:59:02,760 partner and get those those agents that are productive using 788 00:59:03,200 --> 00:59:08,920 our solution. And the way that we can do best 789 00:59:08,920 --> 00:59:14,600 to guarantee that is by making it as seamless and simple as 790 00:59:14,600 --> 00:59:17,520 possible, both from an integration point of view and 791 00:59:17,520 --> 00:59:24,360 from a usability point of view. So if we can reduce the friction 792 00:59:24,680 --> 00:59:28,840 and it's the easiest thing for agents to use to pay and get 793 00:59:28,840 --> 00:59:33,960 paid and there is a there is at least some kind of requirement 794 00:59:33,960 --> 00:59:39,840 for agents to pay and get paid, then the extrapolation is we're 795 00:59:39,840 --> 00:59:44,920 likely at least going to be in the running for the product that 796 00:59:44,920 --> 00:59:50,160 gets used. So that's that's the way, you 797 00:59:50,160 --> 00:59:52,160 know, this is how we're going to market. 798 00:59:52,160 --> 00:59:58,040 So it from from just a practical point of view, looking at multi 799 00:59:58,040 --> 01:00:04,320 agent system builders, swarm builders and Web 3 parlance, 800 01:00:04,600 --> 01:00:06,240 that's who we want to partner with. 801 01:00:08,480 --> 01:00:13,200 You know, this is like the crew AIS of the world, the agent OPS, 802 01:00:13,200 --> 01:00:17,280 the agencies on the web two side, the virtual, the Eliza 803 01:00:18,720 --> 01:00:23,000 OSS, etcetera, the Naphtha's of the Web three world. 804 01:00:25,240 --> 01:00:28,480 And then there's. The the additional step to that, 805 01:00:28,480 --> 01:00:32,880 because that's like B to B and then there's like B to AI or B 806 01:00:32,880 --> 01:00:35,440 to C. So B to B to C or B to B to AI 807 01:00:35,880 --> 01:00:41,560 also helping with that sort of step function with our partners, 808 01:00:41,560 --> 01:00:46,800 helping them try and attract agents that are doing something 809 01:00:46,800 --> 01:00:50,960 useful. I will say this, I'm going to 810 01:00:50,960 --> 01:00:54,480 rant a little bit just just to get this off my chest. 811 01:00:55,440 --> 01:00:59,600 What we need to do when the web 3 AI side of things is get out 812 01:00:59,600 --> 01:01:03,080 of our own fucking way and quit worrying about like 813 01:01:04,960 --> 01:01:09,800 verifiability and attestations on these systems because like 814 01:01:09,800 --> 01:01:13,120 those are nice to haves. What we need to build our 815 01:01:13,600 --> 01:01:17,760 productive agents in swarms to do some kind of useful work. 816 01:01:18,720 --> 01:01:24,520 I also don't think D Phi and agents being added to D Phi is 817 01:01:24,520 --> 01:01:27,760 the massive unlock that a lot of our community actually thinks it 818 01:01:27,760 --> 01:01:30,600 is. But that's, that's for another 819 01:01:30,800 --> 01:01:33,440 conversation. But anyway, my I, I just want to 820 01:01:33,440 --> 01:01:36,680 say like we should be, we need to get out of our own way. 821 01:01:36,680 --> 01:01:39,960 And instead of trying to focus purely on the infrastructure 822 01:01:39,960 --> 01:01:45,200 side and like integrating 0 knowledge or trusted execution 823 01:01:45,200 --> 01:01:50,040 environments or whatever, like I think we'd be better served just 824 01:01:50,520 --> 01:01:55,920 trying to focus on building like agents that do work that to your 825 01:01:55,920 --> 01:01:57,880 point are going to have actual users. 826 01:01:59,080 --> 01:02:03,160 OK. Then tell us about the usage of 827 01:02:04,600 --> 01:02:09,040 of nevermind right now. So kind of how many agent 828 01:02:09,040 --> 01:02:11,960 payments do we actually process kind of like on a daily basis 829 01:02:12,160 --> 01:02:16,600 and how do you how do you see that grow or how do you see how, 830 01:02:16,600 --> 01:02:18,880 where do you see the main drivers of that growth? 831 01:02:19,880 --> 01:02:23,120 Yeah. So at the moment there's been a 832 01:02:23,120 --> 01:02:31,400 a, a bit of a downswing and I think that's somewhat related to 833 01:02:31,400 --> 01:02:38,760 the downswing and the and enthusiasm in the market at its 834 01:02:38,760 --> 01:02:42,160 peak. We're probably seeing like a a 835 01:02:42,160 --> 01:02:44,960 couple handfuls of transactions a day. 836 01:02:46,360 --> 01:02:51,680 You know, in total we've done a few somewhere in the 837 01:02:51,680 --> 01:02:54,440 neighborhood of like 5000 transactions. 838 01:02:57,200 --> 01:03:02,360 Again, we're not counting the, the, the burn piece, we're just 839 01:03:02,360 --> 01:03:09,440 talking about like that initial purchase of those 1000 credits 840 01:03:09,480 --> 01:03:15,440 for example, right? It's relatively nominal, but I'm 841 01:03:15,440 --> 01:03:21,920 bullish like that. That sort of really predate the 842 01:03:22,040 --> 01:03:31,560 AI agent meme taking hold. And so I'm bullish that as we 843 01:03:31,560 --> 01:03:37,480 see more output on the agentic AI side that obviously this is 844 01:03:37,480 --> 01:03:40,040 going to be more and more of a requirement. 845 01:03:40,240 --> 01:03:44,800 So now our business is the business of amplification and 846 01:03:44,800 --> 01:03:48,920 getting this in front of as many AI agent builders is possible. 847 01:03:50,800 --> 01:03:54,280 Cool, so where can we send the AI agent builders to kind of 848 01:03:54,280 --> 01:04:00,200 find out more about Nevermind? Love to have you in our discord 849 01:04:00,200 --> 01:04:03,560 so you can connect to us via our website. 850 01:04:04,000 --> 01:04:09,560 Nevermind dot IONEVERMINED dot IO. 851 01:04:10,920 --> 01:04:17,720 You can also follow us on X. We're at nevermind under score 852 01:04:17,720 --> 01:04:25,720 IONEVERMINED under score IO. But yeah, would be happy to have 853 01:04:27,080 --> 01:04:30,840 everybody that's building and and all of your agents be a part 854 01:04:30,840 --> 01:04:33,560 of our ecosystem. Fantastic. 855 01:04:33,680 --> 01:04:35,360 Thank you so much for coming on, Don. 856 01:04:36,120 --> 01:04:36,880 Thanks for having me.