1 00:00:02,619 --> 00:00:04,639 [music] 2 00:00:14,520 --> 00:00:18,680 >> Matt Bloomberg, one of 3 00:00:17,000 --> 00:00:21,120 my distinguished guests on Pale Blue 4 00:00:18,680 --> 00:00:24,320 Nexus, has joined me today. Matt, I 5 00:00:21,120 --> 00:00:26,640 haven't figured out where you live from 6 00:00:24,320 --> 00:00:29,200 geography-wise, but it looks like maybe 7 00:00:26,640 --> 00:00:32,000 the northeast, but it well welcome 8 00:00:29,200 --> 00:00:33,360 welcome to the show and kindly give 9 00:00:32,000 --> 00:00:34,720 yourself an 10 00:00:33,360 --> 00:00:36,160 introduction. 11 00:00:34,720 --> 00:00:38,040 >> Thanks, Johan. Good to see you. I live 12 00:00:36,160 --> 00:00:40,000 just outside New York City. So, good 13 00:00:38,040 --> 00:00:41,840 call on the northeast. 14 00:00:40,000 --> 00:00:45,195 You would not know from my backyard that 15 00:00:41,840 --> 00:00:45,520 I live 7 miles from New York City. 16 00:00:45,195 --> 00:00:46,760 >> [gasps] 17 00:00:45,520 --> 00:00:48,720 >> I don't know what it was. I think it 18 00:00:46,760 --> 00:00:50,120 might have been the evergreens cuz I 19 00:00:48,720 --> 00:00:51,800 they're very familiar to me. 20 00:00:50,120 --> 00:00:52,720 >> Yes, it is that it is that time of the 21 00:00:51,800 --> 00:00:54,240 year. 22 00:00:52,720 --> 00:00:57,120 >> Yeah. [laughter] 23 00:00:54,240 --> 00:00:59,120 So, well, welcome. Glad to see a fellow 24 00:00:57,120 --> 00:01:01,680 northeasterner. I'm up here in Toronto. 25 00:00:59,120 --> 00:01:04,360 So, um good good to have you on the on 26 00:01:01,680 --> 00:01:08,880 the show today. Tell us about Mark up 27 00:01:04,360 --> 00:01:11,360 and and your role in it and it's it's 28 00:01:08,880 --> 00:01:12,560 it's an older company, isn't it? 29 00:01:11,360 --> 00:01:14,480 >> It is. Yeah, it's a really interesting 30 00:01:12,560 --> 00:01:16,440 story. So, I'm the CEO of the company's 31 00:01:14,480 --> 00:01:18,160 called Mark up AI. 32 00:01:16,440 --> 00:01:19,440 Um and I've only been at the business 33 00:01:18,160 --> 00:01:23,000 for about 34 00:01:19,440 --> 00:01:24,120 a year and a half. The company is 25 35 00:01:23,000 --> 00:01:25,760 years old. 36 00:01:24,120 --> 00:01:29,160 And it has a really really interesting 37 00:01:25,760 --> 00:01:30,920 story. Um we 38 00:01:29,160 --> 00:01:34,200 got started 39 00:01:30,920 --> 00:01:34,200 in 2002 40 00:01:34,600 --> 00:01:39,240 um as a carve out 41 00:01:36,880 --> 00:01:42,000 from something called the Berlin 42 00:01:39,240 --> 00:01:43,040 Artificial Intelligence Institute. 43 00:01:42,000 --> 00:01:45,480 >> Wow. 44 00:01:43,040 --> 00:01:48,000 >> 2002 is a long time ago and there was an 45 00:01:45,480 --> 00:01:50,720 academic institution in Germany 46 00:01:48,000 --> 00:01:52,120 called the AI Institute. 47 00:01:50,720 --> 00:01:54,560 Um and that's where the [clears throat] 48 00:01:52,120 --> 00:01:56,920 company's original technology was built. 49 00:01:54,560 --> 00:01:59,680 Um the company got carved out. It was 50 00:01:56,920 --> 00:02:01,400 called Acrolinx, and over the course of 51 00:01:59,680 --> 00:02:04,160 I don't know, 20-some odd years, built a 52 00:02:01,400 --> 00:02:06,400 really, really compelling business on 53 00:02:04,160 --> 00:02:08,679 top of a unique natural language 54 00:02:06,400 --> 00:02:11,040 processing stack. So, in fact, it was 55 00:02:08,679 --> 00:02:12,311 AI, but it was AI before large language 56 00:02:11,040 --> 00:02:12,680 models. 57 00:02:12,311 --> 00:02:14,920 >> [snorts] 58 00:02:12,680 --> 00:02:20,320 >> Um and um 59 00:02:14,920 --> 00:02:23,040 the company made a decision in uh 2024 60 00:02:20,320 --> 00:02:26,160 that um almost all of its customers were 61 00:02:23,040 --> 00:02:28,400 in the US. Uh and that the business 62 00:02:26,160 --> 00:02:31,920 really needed to pivot and become a 63 00:02:28,400 --> 00:02:33,600 contemporary agentic AI business. Um so, 64 00:02:31,920 --> 00:02:36,360 the board uh went through a search 65 00:02:33,600 --> 00:02:38,680 process and found me. Um and I joined in 66 00:02:36,360 --> 00:02:40,680 early 2025. I brought in 67 00:02:38,680 --> 00:02:42,840 um a lot of folks uh that I've worked 68 00:02:40,680 --> 00:02:45,160 with over the last uh 15 or 20 years 69 00:02:42,840 --> 00:02:46,200 across multiple companies uh in the tech 70 00:02:45,160 --> 00:02:48,440 space. 71 00:02:46,200 --> 00:02:50,400 And basically uh we uh built and 72 00:02:48,440 --> 00:02:52,640 actually just last week launched our new 73 00:02:50,400 --> 00:02:54,280 platform. Uh so, we renamed the company 74 00:02:52,640 --> 00:02:57,640 and the new platform is called Markup 75 00:02:54,280 --> 00:02:59,320 AI. Um it's a really powerful uh agentic 76 00:02:57,640 --> 00:03:01,480 AI platform 77 00:02:59,320 --> 00:03:03,880 uh to help the enterprises keep their 78 00:03:01,480 --> 00:03:05,959 content safe 79 00:03:03,880 --> 00:03:08,840 uh and ready to publish. 80 00:03:05,959 --> 00:03:12,120 Um so, we have agents that do brand, 81 00:03:08,840 --> 00:03:14,040 uh accuracy, compliance, and AI search 82 00:03:12,120 --> 00:03:15,920 visibility. 83 00:03:14,040 --> 00:03:16,840 Um and that's our platform. That's our 84 00:03:15,920 --> 00:03:19,320 business. 85 00:03:16,840 --> 00:03:21,440 Uh we cover any surface imaginable uh 86 00:03:19,320 --> 00:03:23,320 because we're available through MCP 87 00:03:21,440 --> 00:03:25,000 server, through API, 88 00:03:23,320 --> 00:03:26,040 uh and through a Chrome extension that 89 00:03:25,000 --> 00:03:28,040 integrates with just about every 90 00:03:26,040 --> 00:03:29,600 application you can imagine. 91 00:03:28,040 --> 00:03:31,040 Um so, [clears throat] we work with 92 00:03:29,600 --> 00:03:32,320 writers, we work with editors, we work 93 00:03:31,040 --> 00:03:33,519 with large companies across all 94 00:03:32,320 --> 00:03:36,160 verticals. 95 00:03:33,519 --> 00:03:38,360 Uh and again, the the the the story 96 00:03:36,160 --> 00:03:41,040 behind the business really is 97 00:03:38,360 --> 00:03:44,080 um content is the next code. 98 00:03:41,040 --> 00:03:45,280 Um the infrastructure layer around 99 00:03:44,080 --> 00:03:47,600 coding assistance is pretty well 100 00:03:45,280 --> 00:03:49,680 developed at this point. Content is 101 00:03:47,600 --> 00:03:50,959 where most enterprises are now turning 102 00:03:49,680 --> 00:03:52,600 their attention in terms of getting 103 00:03:50,959 --> 00:03:53,760 their true power of large language 104 00:03:52,600 --> 00:03:55,320 models 105 00:03:53,760 --> 00:03:57,200 um and the productivity benefits that 106 00:03:55,320 --> 00:03:59,600 that come with them. 107 00:03:57,200 --> 00:04:01,040 But like the early days of LLMs with 108 00:03:59,600 --> 00:04:01,680 coding, 109 00:04:01,040 --> 00:04:03,662 um 110 00:04:01,680 --> 00:04:03,800 there are lots of guardrails missing. 111 00:04:03,662 --> 00:04:05,400 >> [snorts] 112 00:04:03,800 --> 00:04:07,120 >> Um so that's what we're building and 113 00:04:05,400 --> 00:04:10,200 that's what we launched last week. And 114 00:04:07,120 --> 00:04:13,440 we serve about 120 enterprise customers. 115 00:04:10,200 --> 00:04:15,440 Mostly large Fortune 500 Global 2000. Um 116 00:04:13,440 --> 00:04:18,359 that we're now migrating from our NLP 117 00:04:15,440 --> 00:04:20,320 stack over the Aigentic stack. Um and we 118 00:04:18,359 --> 00:04:21,760 are uh open for business for smaller 119 00:04:20,320 --> 00:04:24,400 companies as well. The new stack is 120 00:04:21,760 --> 00:04:27,480 super easy self-serve. A single user can 121 00:04:24,400 --> 00:04:29,760 onboard and get going in minutes. 122 00:04:27,480 --> 00:04:32,360 >> Wow. So So tell me some of the benefits 123 00:04:29,760 --> 00:04:33,760 that your customers are experiencing for 124 00:04:32,360 --> 00:04:36,120 from using 125 00:04:33,760 --> 00:04:38,120 Markup in the Aigentic stack? 126 00:04:36,120 --> 00:04:42,360 >> Yeah, it's it's there is a productivity 127 00:04:38,120 --> 00:04:44,360 benefit. Um there is a um um we saw a 128 00:04:42,360 --> 00:04:46,760 very clear ROI benefit. There's a a real 129 00:04:44,360 --> 00:04:47,560 quality benefit. Um and there's some 130 00:04:46,760 --> 00:04:49,919 benefits [clears throat] that I always 131 00:04:47,560 --> 00:04:52,040 describe as the previously unthinkable. 132 00:04:49,919 --> 00:04:53,480 Which is really the magic of AI, right? 133 00:04:52,040 --> 00:04:54,720 It's It's not just that you do things 134 00:04:53,480 --> 00:04:56,400 faster, it's that you do things you just 135 00:04:54,720 --> 00:04:57,840 couldn't do before. 136 00:04:56,400 --> 00:04:59,200 So um 137 00:04:57,840 --> 00:05:01,080 let me give you a couple pieces of color 138 00:04:59,200 --> 00:05:02,600 around that. One of our clients 139 00:05:01,080 --> 00:05:04,560 um really succinctly put it the other 140 00:05:02,600 --> 00:05:06,000 day. She said, "Look, we fired all of 141 00:05:04,560 --> 00:05:09,600 our writers. 142 00:05:06,000 --> 00:05:12,400 We put it all in the hands of um LLMs. 143 00:05:09,600 --> 00:05:15,320 And now my team and I that are left 144 00:05:12,400 --> 00:05:17,560 spend 80% of our time fixing content 145 00:05:15,320 --> 00:05:19,760 that comes out of LLMs. 146 00:05:17,560 --> 00:05:21,440 Um so LLMs have turned people from 147 00:05:19,760 --> 00:05:21,960 writers into editors. 148 00:05:21,440 --> 00:05:24,840 >> Yeah. 149 00:05:21,960 --> 00:05:26,680 >> Um so using our platform, we really 150 00:05:24,840 --> 00:05:28,560 streamline that process. We can fix a 151 00:05:26,680 --> 00:05:31,480 tremendous amount of content. Our agents 152 00:05:28,560 --> 00:05:33,600 using deterministic rule sets. Um we 153 00:05:31,480 --> 00:05:35,160 know when to put uh the human in the 154 00:05:33,600 --> 00:05:36,600 loop uh if there's something that 155 00:05:35,160 --> 00:05:38,160 requires their inspection as opposed to 156 00:05:36,600 --> 00:05:39,320 just fixing. 157 00:05:38,160 --> 00:05:41,640 So all that's around quality and 158 00:05:39,320 --> 00:05:43,240 productivity. Um and the kind of thing 159 00:05:41,640 --> 00:05:45,440 you can do with us that you you couldn't 160 00:05:43,240 --> 00:05:47,480 really do before. Uh so what I call the 161 00:05:45,440 --> 00:05:49,840 previously unimaginable. 162 00:05:47,480 --> 00:05:52,393 Is you can use our API to check a 163 00:05:49,840 --> 00:05:52,520 repository of content of 164 00:05:52,393 --> 00:05:56,560 >> [clears throat] 165 00:05:52,520 --> 00:05:58,160 >> limitless size. Um so um I'll give you a 166 00:05:56,560 --> 00:06:00,360 couple of examples. Um 167 00:05:58,160 --> 00:06:01,640 one of uh one of our clients told us the 168 00:06:00,360 --> 00:06:03,000 other day, "Oh, we just did this big 169 00:06:01,640 --> 00:06:04,880 acquisition 170 00:06:03,000 --> 00:06:07,280 and the brand of the company we acquired 171 00:06:04,880 --> 00:06:10,360 is all is all off. It's not our brand. 172 00:06:07,280 --> 00:06:11,560 Um can you know, can we use you to fix 173 00:06:10,360 --> 00:06:14,000 the brand?" 174 00:06:11,560 --> 00:06:15,040 The answer is yes. We can scan 5 million 175 00:06:14,000 --> 00:06:17,320 pages 176 00:06:15,040 --> 00:06:20,400 um of support documentation of marketing 177 00:06:17,320 --> 00:06:22,040 for that brand um and flag everything 178 00:06:20,400 --> 00:06:24,520 that doesn't conform or needs to conform 179 00:06:22,040 --> 00:06:28,320 or fix it automatically. 180 00:06:24,520 --> 00:06:30,400 >> Amazing. And may I ask, do you do this 181 00:06:28,320 --> 00:06:33,240 in the cloud or do you do this on your 182 00:06:30,400 --> 00:06:34,640 own uh servers? 183 00:06:33,240 --> 00:06:36,720 >> Um 184 00:06:34,640 --> 00:06:39,160 we I mean we're we're we are native 185 00:06:36,720 --> 00:06:42,680 cloud-based application for sure. So we 186 00:06:39,160 --> 00:06:44,040 we run in AWS instance um and um and 187 00:06:42,680 --> 00:06:46,520 that that's where everything is is 188 00:06:44,040 --> 00:06:49,120 located, but obviously the um 189 00:06:46,520 --> 00:06:51,520 uh we do draw on the LLMs and we do most 190 00:06:49,120 --> 00:06:54,480 of that through AWS Bedrock. 191 00:06:51,520 --> 00:06:56,280 >> Okay. Yeah, I guess uh 192 00:06:54,480 --> 00:06:58,520 my one of the things I've been toying 193 00:06:56,280 --> 00:07:01,440 around recently is like with 194 00:06:58,520 --> 00:07:03,919 uh open-source models and things like 195 00:07:01,440 --> 00:07:06,440 that and I I don't know where you stand 196 00:07:03,919 --> 00:07:08,720 on that, but uh you sound like you're 197 00:07:06,440 --> 00:07:10,240 you might be using some frontier model. 198 00:07:08,720 --> 00:07:12,320 >> We are, yeah. Yeah, we actually use 199 00:07:10,240 --> 00:07:14,240 different models um 200 00:07:12,320 --> 00:07:15,560 uh we um 201 00:07:14,240 --> 00:07:18,120 uh you know, we have lots of different 202 00:07:15,560 --> 00:07:20,280 agents and different cases and um one of 203 00:07:18,120 --> 00:07:22,440 the um sort of value props we have to 204 00:07:20,280 --> 00:07:23,720 clients is um that we make sure that 205 00:07:22,440 --> 00:07:25,680 we're using the right model, the most 206 00:07:23,720 --> 00:07:27,040 efficient model and that anytime we 207 00:07:25,680 --> 00:07:29,240 switch models, 208 00:07:27,040 --> 00:07:31,080 um we are doing all the exhaustive eval 209 00:07:29,240 --> 00:07:33,400 work to make sure that the the output is 210 00:07:31,080 --> 00:07:34,680 consistent for the customer. So we have 211 00:07:33,400 --> 00:07:36,600 a lot of things running on Gemini, we 212 00:07:34,680 --> 00:07:37,840 have things running on Anthropic, um we 213 00:07:36,600 --> 00:07:40,800 have things running on on some of the 214 00:07:37,840 --> 00:07:43,080 open-source models as well. 215 00:07:40,800 --> 00:07:45,760 >> What kind of evidence 216 00:07:43,080 --> 00:07:49,520 do you gather uh 217 00:07:45,760 --> 00:07:50,440 to say like you're helping your customer 218 00:07:49,520 --> 00:07:51,400 with 219 00:07:50,440 --> 00:07:53,360 uh 220 00:07:51,400 --> 00:07:56,200 guarding their content, for example? 221 00:07:53,360 --> 00:07:59,520 Like how how do you prove that? 222 00:07:56,200 --> 00:08:01,440 >> Well, it's fairly um easy with our uh 223 00:07:59,520 --> 00:08:03,160 analytics and reporting module uh 224 00:08:01,440 --> 00:08:05,400 because everything that we do for a 225 00:08:03,160 --> 00:08:07,520 client's content, what we're doing is 226 00:08:05,400 --> 00:08:09,480 we're checking their content against a 227 00:08:07,520 --> 00:08:10,000 particular standard. 228 00:08:09,480 --> 00:08:11,520 >> Mhm. 229 00:08:10,000 --> 00:08:12,760 >> Um so, the reporting and analytics just 230 00:08:11,520 --> 00:08:14,720 comes back to how many things did we 231 00:08:12,760 --> 00:08:16,720 catch and fix against against a given 232 00:08:14,720 --> 00:08:18,720 standard. So, um I'll give you some 233 00:08:16,720 --> 00:08:22,440 examples. So, for Brand, which is one of 234 00:08:18,720 --> 00:08:24,040 our swim lanes, um the uh standards 235 00:08:22,440 --> 00:08:25,919 could be a company's terminology 236 00:08:24,040 --> 00:08:28,640 dictionary or glossary of preferred 237 00:08:25,919 --> 00:08:30,800 terms, or it could be their brand and 238 00:08:28,640 --> 00:08:32,320 style guide. And some corporations have 239 00:08:30,800 --> 00:08:33,680 brand and style guides that are hundreds 240 00:08:32,320 --> 00:08:33,919 and hundreds of pages long. 241 00:08:33,680 --> 00:08:35,560 >> Yeah. 242 00:08:33,919 --> 00:08:37,919 >> Even smaller ones have brand and style 243 00:08:35,560 --> 00:08:39,960 guides that are 20-30 pages long. 244 00:08:37,919 --> 00:08:41,719 So, anytime we do a pass at a document 245 00:08:39,960 --> 00:08:44,200 or a webpage and correct it, we can say, 246 00:08:41,719 --> 00:08:46,240 "Oh, we found 82 errors. We fixed them." 247 00:08:44,200 --> 00:08:47,360 Um same thing if we're doing compliance 248 00:08:46,240 --> 00:08:48,080 checking 249 00:08:47,360 --> 00:08:49,920 um 250 00:08:48,080 --> 00:08:51,600 for a particular law that the client 251 00:08:49,920 --> 00:08:52,960 cares about based on its jurisdiction or 252 00:08:51,600 --> 00:08:54,520 its vertical, 253 00:08:52,960 --> 00:08:55,760 how many things did we flag for human in 254 00:08:54,520 --> 00:08:58,400 the loop? 255 00:08:55,760 --> 00:09:00,000 Um similarly for AEO and GEO, sort of 256 00:08:58,400 --> 00:09:01,920 that AI search visibility, how many 257 00:09:00,000 --> 00:09:04,120 things did we find, flag, and fix that 258 00:09:01,920 --> 00:09:06,040 are going to improve um AI search 259 00:09:04,120 --> 00:09:08,720 visibility. And then the same thing for 260 00:09:06,040 --> 00:09:10,480 accuracy, uh when we do accuracy work, 261 00:09:08,720 --> 00:09:12,000 how many things did we find and flag 262 00:09:10,480 --> 00:09:13,800 against the deterministic source that 263 00:09:12,000 --> 00:09:14,960 needed to be corrected. 264 00:09:13,800 --> 00:09:16,120 >> Right. 265 00:09:14,960 --> 00:09:17,760 Well, I thought that 266 00:09:16,120 --> 00:09:19,960 >> And our system our system rates 267 00:09:17,760 --> 00:09:21,200 everything low, medium, high. Um so, 268 00:09:19,960 --> 00:09:22,640 there are a lot of things we find that 269 00:09:21,200 --> 00:09:23,839 like, all right, great, we put another 270 00:09:22,640 --> 00:09:24,920 comma in. 271 00:09:23,839 --> 00:09:26,440 Okay, you want your content to be 272 00:09:24,920 --> 00:09:28,720 consistent, but no one's going to get 273 00:09:26,440 --> 00:09:29,720 fired because the comma's missing. Um 274 00:09:28,720 --> 00:09:31,080 that's a little different than a 275 00:09:29,720 --> 00:09:33,280 pharmaceutical company putting out 276 00:09:31,080 --> 00:09:35,000 marketing that doesn't actually speak 277 00:09:33,280 --> 00:09:36,839 the truth about the drug, 278 00:09:35,000 --> 00:09:39,760 where someone could get fired or someone 279 00:09:36,839 --> 00:09:41,560 could die. So, uh you know, that's a uh 280 00:09:39,760 --> 00:09:43,320 an extreme example 281 00:09:41,560 --> 00:09:45,800 on either side. 282 00:09:43,320 --> 00:09:48,920 >> What do you think about the I talked 283 00:09:45,800 --> 00:09:53,320 about this on a on a taping yesterday, 284 00:09:48,920 --> 00:09:55,600 but like the agent-to-agent economy and 285 00:09:53,320 --> 00:09:57,240 people are talking about 286 00:09:55,600 --> 00:09:58,840 you're going to be talking and your 287 00:09:57,240 --> 00:09:59,960 brand is going to be represented by 288 00:09:58,840 --> 00:10:00,800 another 289 00:09:59,960 --> 00:10:04,560 uh 290 00:10:00,800 --> 00:10:06,840 is going to be assessed by a agent um in 291 00:10:04,560 --> 00:10:08,720 at some point in the future. 292 00:10:06,840 --> 00:10:10,640 So, it may not be a human reading the 293 00:10:08,720 --> 00:10:12,240 webpage anymore. 294 00:10:10,640 --> 00:10:14,040 >> Yeah, I mean it is 295 00:10:12,240 --> 00:10:17,120 it's an interesting world. Like they 296 00:10:14,040 --> 00:10:18,000 talk about the metaverse, I guess. 297 00:10:17,120 --> 00:10:19,760 The 298 00:10:18,000 --> 00:10:21,160 a lot of people are starting to talk 299 00:10:19,760 --> 00:10:22,440 about this sort of two-track internet, 300 00:10:21,160 --> 00:10:24,800 the human-to-human and 301 00:10:22,440 --> 00:10:26,560 machine-to-machine and and the crossover 302 00:10:24,800 --> 00:10:27,920 between those two. 303 00:10:26,560 --> 00:10:29,680 You know, the 304 00:10:27,920 --> 00:10:31,040 reality is you can do some things to 305 00:10:29,680 --> 00:10:33,880 optimize your content for machine 306 00:10:31,040 --> 00:10:35,640 readability, but most machines are 307 00:10:33,880 --> 00:10:37,440 trained to read human content and 308 00:10:35,640 --> 00:10:38,000 actually look for signals of human 309 00:10:37,440 --> 00:10:38,600 content. 310 00:10:38,000 --> 00:10:42,200 >> Right. 311 00:10:38,600 --> 00:10:44,800 >> Um so, we have agents like the AI search 312 00:10:42,200 --> 00:10:48,400 and AI visibility agents that know the 313 00:10:44,800 --> 00:10:50,640 rules for AEO and GEO and will will make 314 00:10:48,400 --> 00:10:52,640 the text conform to those rules, which 315 00:10:50,640 --> 00:10:54,360 sort of gets to machine readability. But 316 00:10:52,640 --> 00:10:56,520 if you really look underneath the hood 317 00:10:54,360 --> 00:10:59,160 with those rules, a lot of those rules 318 00:10:56,520 --> 00:11:01,640 are about the content being authentic, 319 00:10:59,160 --> 00:11:02,960 about being fresh, 320 00:11:01,640 --> 00:11:05,120 about 321 00:11:02,960 --> 00:11:07,680 being accurate, and all those things are 322 00:11:05,120 --> 00:11:09,360 are human rules. So, I'm not sure at the 323 00:11:07,680 --> 00:11:11,280 end of the day that this two-track 324 00:11:09,360 --> 00:11:13,640 internet is exactly going to unfold the 325 00:11:11,280 --> 00:11:14,960 way it sounds, 326 00:11:13,640 --> 00:11:16,120 but to the extent that it does, that 327 00:11:14,960 --> 00:11:18,920 just means that people are going to need 328 00:11:16,120 --> 00:11:20,880 to version their content A and B. 329 00:11:18,920 --> 00:11:22,240 >> Right. 330 00:11:20,880 --> 00:11:23,800 >> that's what our agents are designed to 331 00:11:22,240 --> 00:11:25,920 do. 332 00:11:23,800 --> 00:11:28,280 >> You mentioned the word metaverse. Did 333 00:11:25,920 --> 00:11:31,280 you go through that that whole cycle? 334 00:11:28,280 --> 00:11:33,080 >> I don't mean the metaverse. 335 00:11:31,280 --> 00:11:34,480 But it is it is uh let's call it a a 336 00:11:33,080 --> 00:11:37,120 multiverse. 337 00:11:34,480 --> 00:11:39,800 >> Right. Right. Uh did you did you go 338 00:11:37,120 --> 00:11:41,520 through that sort of cycle with NFTs and 339 00:11:39,800 --> 00:11:43,200 metaverse just randomly at 340 00:11:41,520 --> 00:11:44,680 >> Personally, I didn't engage that much in 341 00:11:43,200 --> 00:11:47,360 it. I mean, I know what all that is, but 342 00:11:44,680 --> 00:11:50,028 um uh I my my businesses have always 343 00:11:47,360 --> 00:11:51,200 been a little more boring than that. 344 00:11:50,028 --> 00:11:52,360 >> [laughter] 345 00:11:51,200 --> 00:11:54,680 >> Yeah, so 346 00:11:52,360 --> 00:11:57,720 uh speaking of which, um Matt, what do 347 00:11:54,680 --> 00:11:59,840 you do for for fun as a as a father and 348 00:11:57,720 --> 00:12:02,040 as a CEO? 349 00:11:59,840 --> 00:12:03,400 >> Um I have lots of fun in lots of 350 00:12:02,040 --> 00:12:06,120 different ways. I've got three great 351 00:12:03,400 --> 00:12:07,760 kids. They're late teens and uh I love 352 00:12:06,120 --> 00:12:09,640 having fun with them. Uh it's great to 353 00:12:07,760 --> 00:12:11,200 have everyone home for the summer from 354 00:12:09,640 --> 00:12:14,560 the different uh schools that they're at 355 00:12:11,200 --> 00:12:17,080 and uh um one of them likes to golf and 356 00:12:14,560 --> 00:12:18,600 so do I, so we do a lot of golf. Um 357 00:12:17,080 --> 00:12:23,400 a couple of them like to watch movies 358 00:12:18,600 --> 00:12:25,520 and uh we do a lot of that. Um and um 359 00:12:23,400 --> 00:12:27,360 love hiking. Uh love spending time with 360 00:12:25,520 --> 00:12:28,880 my wife. Um 361 00:12:27,360 --> 00:12:31,200 and uh I would say sort of the 362 00:12:28,880 --> 00:12:33,000 intellectual interests for me are have 363 00:12:31,200 --> 00:12:35,640 always been around uh American history 364 00:12:33,000 --> 00:12:37,160 and politics. Uh so I exercise that one 365 00:12:35,640 --> 00:12:39,200 in lots of different ways. I'm I'm doing 366 00:12:37,160 --> 00:12:40,040 a lot of vibe coding at the moment. Um 367 00:12:39,200 --> 00:12:43,080 so 368 00:12:40,040 --> 00:12:46,160 like uh like to keep busy. 369 00:12:43,080 --> 00:12:48,400 >> The thing about history is like I never 370 00:12:46,160 --> 00:12:50,120 watch historical movies because I know 371 00:12:48,400 --> 00:12:52,000 the the the movies are written by the 372 00:12:50,120 --> 00:12:54,400 winners and it's like I don't want to be 373 00:12:52,000 --> 00:12:57,493 lied to, right? So, how do you tell in 374 00:12:54,400 --> 00:12:58,080 history what's real and what's not? 375 00:12:57,493 --> 00:13:01,680 >> [snorts] 376 00:12:58,080 --> 00:13:02,480 >> Um you know, I I think with all content 377 00:13:01,680 --> 00:13:05,040 um 378 00:13:02,480 --> 00:13:07,319 and and increasingly so with with AI, 379 00:13:05,040 --> 00:13:07,360 all content um 380 00:13:07,319 --> 00:13:09,800 >> [clears throat] 381 00:13:07,360 --> 00:13:11,920 >> people need to bring check with healthy 382 00:13:09,800 --> 00:13:14,040 skepticism to it. Yeah. Um and really 383 00:13:11,920 --> 00:13:14,640 understand how to look for 384 00:13:14,040 --> 00:13:16,160 um 385 00:13:14,640 --> 00:13:18,960 you know, look for uh signs of 386 00:13:16,160 --> 00:13:20,800 credibility. Um 387 00:13:18,960 --> 00:13:23,520 it's one of the things I wish schools 388 00:13:20,800 --> 00:13:25,320 taught kids a little bit more. 389 00:13:23,520 --> 00:13:28,560 Cuz I think the internet has kind of 390 00:13:25,320 --> 00:13:30,600 done a lot to to break that um 391 00:13:28,560 --> 00:13:32,000 in in lots of ways. So, I you know, I 392 00:13:30,600 --> 00:13:34,000 feel like I have pretty finely tuned 393 00:13:32,000 --> 00:13:35,600 radar at this point, but I worry 394 00:13:34,000 --> 00:13:36,760 sometimes that that 395 00:13:35,600 --> 00:13:38,000 teenagers don't. Although, quite 396 00:13:36,760 --> 00:13:40,120 frankly, teenagers are pretty good at 397 00:13:38,000 --> 00:13:41,600 spotting video fakes 398 00:13:40,120 --> 00:13:46,000 in ways that 399 00:13:41,600 --> 00:13:47,920 you might not catch as as as an adult. 400 00:13:46,000 --> 00:13:49,720 >> Yeah, teenagers are good. My my my 401 00:13:47,920 --> 00:13:52,480 5-year-old isn't as good. He just 402 00:13:49,720 --> 00:13:54,480 watches these cat videos that the cats 403 00:13:52,480 --> 00:13:56,080 are doing all of these weird different 404 00:13:54,480 --> 00:13:57,600 things and you can you know it's AI, but 405 00:13:56,080 --> 00:13:58,360 he doesn't care. So. 406 00:13:57,600 --> 00:14:00,400 >> Yeah. 407 00:13:58,360 --> 00:14:03,040 >> It is it's it's a weird world. 408 00:14:00,400 --> 00:14:04,600 Back on track to to to 409 00:14:03,040 --> 00:14:07,480 to mark up. 410 00:14:04,600 --> 00:14:10,480 Um you guys lean on the Gartners 411 00:14:07,480 --> 00:14:10,480 Gartners 412 00:14:10,680 --> 00:14:14,280 sort of certification. Talk Talk about 413 00:14:12,200 --> 00:14:16,240 that and and and what that means for the 414 00:14:14,280 --> 00:14:18,400 future of content. 415 00:14:16,240 --> 00:14:20,760 >> Yeah, absolutely. 416 00:14:18,400 --> 00:14:23,320 You know, we So, Gartner sort of coined 417 00:14:20,760 --> 00:14:25,720 this term guardian agents a while back 418 00:14:23,320 --> 00:14:27,680 and it was it was as we were really 419 00:14:25,720 --> 00:14:29,120 building out our solution. We spent a 420 00:14:27,680 --> 00:14:30,960 lot of time with with Gartner. We're a 421 00:14:29,120 --> 00:14:33,440 client. We do a lot of briefings. We do 422 00:14:30,960 --> 00:14:34,560 a lot of inquiries there as well. And 423 00:14:33,440 --> 00:14:37,080 those guys are super smart about 424 00:14:34,560 --> 00:14:38,480 software world and you know, their their 425 00:14:37,080 --> 00:14:40,040 point 426 00:14:38,480 --> 00:14:42,440 and sort of the origin of the term 427 00:14:40,040 --> 00:14:45,680 guardian agent is that the only thing 428 00:14:42,440 --> 00:14:45,839 powerful enough to check AI is AI. 429 00:14:45,680 --> 00:14:48,400 >> Right. 430 00:14:45,839 --> 00:14:49,440 >> And which is I I always say like the 431 00:14:48,400 --> 00:14:51,000 only thing strong enough to cut a 432 00:14:49,440 --> 00:14:51,320 diamond is another diamond. 433 00:14:51,000 --> 00:14:52,280 >> Yep. 434 00:14:51,320 --> 00:14:54,360 >> And 435 00:14:52,280 --> 00:14:56,920 that's the the the the term guardian 436 00:14:54,360 --> 00:14:59,080 agent. So, AI to check the AI. 437 00:14:56,920 --> 00:15:00,720 And you know, the reality is our agents 438 00:14:59,080 --> 00:15:02,640 will check human content as well, not 439 00:15:00,720 --> 00:15:04,480 just AI content, but 440 00:15:02,640 --> 00:15:05,960 but the power comes in when you're 441 00:15:04,480 --> 00:15:07,600 checking 442 00:15:05,960 --> 00:15:09,640 you know, kind of hundreds or or 443 00:15:07,600 --> 00:15:11,520 thousands of things or millions of 444 00:15:09,640 --> 00:15:13,320 things all at once that you really need 445 00:15:11,520 --> 00:15:15,120 AI to do. 446 00:15:13,320 --> 00:15:17,480 >> And there's like a there's a concept of 447 00:15:15,120 --> 00:15:19,120 a deterministic trust score when you 448 00:15:17,480 --> 00:15:23,040 have AI 449 00:15:19,120 --> 00:15:23,800 track AI or or check on AI. Explain 450 00:15:23,040 --> 00:15:25,800 that. 451 00:15:23,800 --> 00:15:29,120 >> Yeah, I mean all of our agents have um 452 00:15:25,800 --> 00:15:30,880 algorithms that score. Um we uh most 453 00:15:29,120 --> 00:15:32,800 often, unless a client really wants to 454 00:15:30,880 --> 00:15:35,040 get granular and see a zero to 100 455 00:15:32,800 --> 00:15:37,440 score, um we will bucket things into 456 00:15:35,040 --> 00:15:38,760 low, medium, and high risk. Um and that 457 00:15:37,440 --> 00:15:40,280 that tends to do the job for most 458 00:15:38,760 --> 00:15:41,840 clients. The high risk has to be human 459 00:15:40,280 --> 00:15:45,320 in the loop. Medium and low risk just 460 00:15:41,840 --> 00:15:47,240 fix. Um but uh but you know, any any 461 00:15:45,320 --> 00:15:49,160 determinism can come with an algorithm 462 00:15:47,240 --> 00:15:50,800 that produces a score. 463 00:15:49,160 --> 00:15:53,160 Uh and uh you know, again, back to my 464 00:15:50,800 --> 00:15:55,480 earlier example, if a comma is missing, 465 00:15:53,160 --> 00:15:57,280 it's not going to be that big of a deal. 466 00:15:55,480 --> 00:15:58,640 Um if you are saying something that's 467 00:15:57,280 --> 00:16:00,800 out of compliance with a claim 468 00:15:58,640 --> 00:16:02,720 substantiation law, uh that's going to 469 00:16:00,800 --> 00:16:05,040 be a real problem. 470 00:16:02,720 --> 00:16:07,200 >> And is there like some way to check what 471 00:16:05,040 --> 00:16:09,760 you guys have done? Like is there like a 472 00:16:07,200 --> 00:16:10,480 a false negative or check or anything 473 00:16:09,760 --> 00:16:12,640 like that for you guys? 474 00:16:10,480 --> 00:16:13,360 >> have agents that check our own agents. 475 00:16:12,640 --> 00:16:15,320 Um 476 00:16:13,360 --> 00:16:17,200 uh yeah, so there is there is an extra 477 00:16:15,320 --> 00:16:18,560 loop on that. 478 00:16:17,200 --> 00:16:22,080 >> Amazing. Uh 479 00:16:18,560 --> 00:16:24,480 definitely you're you're you're 480 00:16:22,080 --> 00:16:26,480 marketing and selling safety and you're 481 00:16:24,480 --> 00:16:28,520 checking yourselves. Uh 482 00:16:26,480 --> 00:16:29,880 so I think that's amazing. What what are 483 00:16:28,520 --> 00:16:31,880 some of your future 484 00:16:29,880 --> 00:16:36,160 >> we're selling safety 485 00:16:31,880 --> 00:16:38,520 um and confidence. Um but we're also um 486 00:16:36,160 --> 00:16:40,240 you know, I think selling productivity. 487 00:16:38,520 --> 00:16:42,560 Um and in the case [clears throat] of of 488 00:16:40,240 --> 00:16:44,280 search AI search visibility, um you 489 00:16:42,560 --> 00:16:45,400 know, we're selling a revenue driver. Um 490 00:16:44,280 --> 00:16:47,000 if companies you know, if your your 491 00:16:45,400 --> 00:16:49,280 content can't get found, you can't get 492 00:16:47,000 --> 00:16:51,200 found. 493 00:16:49,280 --> 00:16:53,480 >> I absolutely believe in that. The the 494 00:16:51,200 --> 00:16:55,839 the sponsor of this show is is Huckabuy 495 00:16:53,480 --> 00:16:57,640 and they are in the search visibility 496 00:16:55,839 --> 00:16:59,240 phase. Let maybe let's talk about that a 497 00:16:57,640 --> 00:17:02,400 little bit. Um 498 00:16:59,240 --> 00:17:06,360 where do you think this space is going? 499 00:17:02,400 --> 00:17:10,560 Obviously, the old uh world of Google 500 00:17:06,360 --> 00:17:13,839 search is no longer there with with AI 501 00:17:10,560 --> 00:17:16,439 results coming above the fold um and and 502 00:17:13,839 --> 00:17:19,600 most people using AI to 503 00:17:16,439 --> 00:17:21,760 discover brands these days. What is your 504 00:17:19,600 --> 00:17:25,040 uh take on that? 505 00:17:21,760 --> 00:17:25,959 >> I mean that's that's here to stay. 506 00:17:25,040 --> 00:17:27,680 You know, it's interesting because you 507 00:17:25,959 --> 00:17:29,840 look at 508 00:17:27,680 --> 00:17:31,760 you know, you look at the different AI 509 00:17:29,840 --> 00:17:34,040 services and they really the quality 510 00:17:31,760 --> 00:17:35,720 varies pretty widely 511 00:17:34,040 --> 00:17:37,440 you know, when they default you to a an 512 00:17:35,720 --> 00:17:39,040 old model or if you decide you're going 513 00:17:37,440 --> 00:17:40,360 to use an old model because it's 514 00:17:39,040 --> 00:17:41,480 cheaper. 515 00:17:40,360 --> 00:17:42,840 You're going to find stuff that's out of 516 00:17:41,480 --> 00:17:44,920 date. 517 00:17:42,840 --> 00:17:46,800 People who use things like perplexity 518 00:17:44,920 --> 00:17:48,840 which are like up to the minute rag 519 00:17:46,800 --> 00:17:50,760 optimized are finding things that are 520 00:17:48,840 --> 00:17:52,640 really current. 521 00:17:50,760 --> 00:17:55,200 Freshness is a real test that you want 522 00:17:52,640 --> 00:17:57,080 to use with that. But you know, I think 523 00:17:55,200 --> 00:17:58,600 whether it's Google layering in Gemini's 524 00:17:57,080 --> 00:18:00,640 output or people going straight to 525 00:17:58,600 --> 00:18:03,120 perplexity or something else. People are 526 00:18:00,640 --> 00:18:04,640 just going to increasingly be 527 00:18:03,120 --> 00:18:06,880 you know, turning to AI to look for 528 00:18:04,640 --> 00:18:08,160 things and you know, I mean the results 529 00:18:06,880 --> 00:18:10,000 the difference in the results is 530 00:18:08,160 --> 00:18:11,440 incredible. 531 00:18:10,000 --> 00:18:12,920 You you ask perplexity a question you 532 00:18:11,440 --> 00:18:14,160 get back a table comparing five 533 00:18:12,920 --> 00:18:16,960 different products against the three 534 00:18:14,160 --> 00:18:18,480 things they know you care about. That's 535 00:18:16,960 --> 00:18:20,440 a pretty big difference than than 536 00:18:18,480 --> 00:18:22,720 getting a page of links which is sort of 537 00:18:20,440 --> 00:18:24,120 the old old school work. So Google's 538 00:18:22,720 --> 00:18:26,800 obviously figured this out too. It's 539 00:18:24,120 --> 00:18:28,560 sort of you know, a 10 up top. 540 00:18:26,800 --> 00:18:30,560 But I'm not sure that's exactly the best 541 00:18:28,560 --> 00:18:31,920 engine for that output at least today. 542 00:18:30,560 --> 00:18:33,080 Google's got plenty of resources. They 543 00:18:31,920 --> 00:18:35,200 can make that better and better and 544 00:18:33,080 --> 00:18:37,600 better. 545 00:18:35,200 --> 00:18:38,840 >> And you know, the old way is paid 546 00:18:37,600 --> 00:18:41,040 search. 547 00:18:38,840 --> 00:18:43,680 How does that get translate into this 548 00:18:41,040 --> 00:18:45,400 world? Do you see a path ahead for 549 00:18:43,680 --> 00:18:48,760 neutrality there? 550 00:18:45,400 --> 00:18:50,400 >> Yeah, I mean that's that's interesting 551 00:18:48,760 --> 00:18:53,720 when you think about it. You know, I 552 00:18:50,400 --> 00:18:55,040 would say the current regime is that 553 00:18:53,720 --> 00:18:56,520 paid search is getting a little less 554 00:18:55,040 --> 00:18:58,200 impactful. 555 00:18:56,520 --> 00:19:01,120 I expect that will change over time 556 00:18:58,200 --> 00:19:02,320 because economics dictate that it will. 557 00:19:01,120 --> 00:19:04,400 But I you know, I could be wrong. Look 558 00:19:02,320 --> 00:19:07,240 if everybody starts doing you know, $20 559 00:19:04,400 --> 00:19:08,880 a month subscriptions to these things 560 00:19:07,240 --> 00:19:11,640 you know, to use their preferred one and 561 00:19:08,880 --> 00:19:13,400 to use it in a memory focused mode. 562 00:19:11,640 --> 00:19:16,760 Maybe not. I mean the reason that paid 563 00:19:13,400 --> 00:19:19,320 search evolved is that search is free 564 00:19:16,760 --> 00:19:21,880 and chewed up a lot of compute and a lot 565 00:19:19,320 --> 00:19:23,600 of everything at Google and Yahoo and 566 00:19:21,880 --> 00:19:24,760 Microsoft. 567 00:19:23,600 --> 00:19:28,160 And 568 00:19:24,760 --> 00:19:30,440 you know, SEM was a really good way of 569 00:19:28,160 --> 00:19:31,640 of recouping the cost of that. 570 00:19:30,440 --> 00:19:33,000 You know, if the world moves to 571 00:19:31,640 --> 00:19:35,080 subscription model, maybe that doesn't 572 00:19:33,000 --> 00:19:36,200 need to be there. My guess is even if it 573 00:19:35,080 --> 00:19:38,080 moves to subscription model, you'll 574 00:19:36,200 --> 00:19:39,720 start seeing that in 575 00:19:38,080 --> 00:19:41,720 search results and you know, hopefully 576 00:19:39,720 --> 00:19:43,320 that will evolve the same way the SEO 577 00:19:41,720 --> 00:19:44,480 versus SEM world did where things are 578 00:19:43,320 --> 00:19:47,640 flagged and at least you know what's 579 00:19:44,480 --> 00:19:49,160 sponsored versus what's organic. 580 00:19:47,640 --> 00:19:51,800 >> I like that 581 00:19:49,160 --> 00:19:54,160 insight very much. I think like maybe 582 00:19:51,800 --> 00:19:55,760 everybody will have a $20 a month 583 00:19:54,160 --> 00:19:57,320 subscription just as they have car 584 00:19:55,760 --> 00:19:57,880 insurance and uh 585 00:19:57,320 --> 00:19:59,973 >> And 586 00:19:57,880 --> 00:20:00,000 and my guess is more than one. 587 00:19:59,973 --> 00:20:01,920 >> [snorts] 588 00:20:00,000 --> 00:20:03,600 >> Um you know, they'll have probably $20 589 00:20:01,920 --> 00:20:05,960 ones and then they'll have the $100 one 590 00:20:03,600 --> 00:20:09,360 for the max work. 591 00:20:05,960 --> 00:20:12,440 >> Right. Right. It it it certainly 592 00:20:09,360 --> 00:20:15,160 I do. I have couple and I have $20 ones 593 00:20:12,440 --> 00:20:19,240 too. So you're right. 594 00:20:15,160 --> 00:20:21,880 Talk about the world of I guess search 595 00:20:19,240 --> 00:20:23,840 in a year's time or two years' time. 596 00:20:21,880 --> 00:20:25,560 Where do you think it's going to go? 597 00:20:23,840 --> 00:20:27,440 I'm interested in this question as well. 598 00:20:25,560 --> 00:20:29,240 >> Yeah, I am you know, I'm not really an 599 00:20:27,440 --> 00:20:30,960 expert in search. So I don't know that 600 00:20:29,240 --> 00:20:33,040 I'm the best person to ask. I mean, I 601 00:20:30,960 --> 00:20:34,280 would say um 602 00:20:33,040 --> 00:20:36,680 uh 603 00:20:34,280 --> 00:20:38,480 the data would suggest the public data 604 00:20:36,680 --> 00:20:39,760 would suggest that Google's not losing 605 00:20:38,480 --> 00:20:42,160 share. 606 00:20:39,760 --> 00:20:43,680 Hmm. [snorts] I find that a little hard 607 00:20:42,160 --> 00:20:46,040 to believe. 608 00:20:43,680 --> 00:20:47,480 Um 609 00:20:46,040 --> 00:20:48,960 and I don't I can't reconcile that 610 00:20:47,480 --> 00:20:51,040 though. And the only reason I find it 611 00:20:48,960 --> 00:20:53,360 hard to believe is that there are a lot 612 00:20:51,040 --> 00:20:55,960 of people I know that are using the the 613 00:20:53,360 --> 00:20:57,680 different models not Gemini for search. 614 00:20:55,960 --> 00:21:00,320 Um and I know maybe I know a really 615 00:20:57,680 --> 00:21:01,880 small segment of the population, but um 616 00:21:00,320 --> 00:21:04,120 and Google certainly built a reflex over 617 00:21:01,880 --> 00:21:07,200 the last you know, 25 years 618 00:21:04,120 --> 00:21:08,160 that that included a verb Google. So 619 00:21:07,200 --> 00:21:10,000 um 620 00:21:08,160 --> 00:21:11,920 Uh so, I you know, a little unclear to 621 00:21:10,000 --> 00:21:13,600 me, but there there has to be some some 622 00:21:11,920 --> 00:21:15,800 fragmentation coming. 623 00:21:13,600 --> 00:21:17,640 Um but, you know, you never know with 624 00:21:15,800 --> 00:21:19,000 big tech. Um 625 00:21:17,640 --> 00:21:20,280 you know, Perplexity could start eating 626 00:21:19,000 --> 00:21:21,120 share, and all of a sudden someone buys 627 00:21:20,280 --> 00:21:23,320 them. 628 00:21:21,120 --> 00:21:24,960 Uh and that could be Google, you know, 629 00:21:23,320 --> 00:21:26,920 doubling down if they're allowed to uh 630 00:21:24,960 --> 00:21:28,800 with antitrust. Uh or it could be 631 00:21:26,920 --> 00:21:30,360 someone else that decides to um you 632 00:21:28,800 --> 00:21:32,960 know, to really to pick up that user 633 00:21:30,360 --> 00:21:34,200 base, pick up that technology, and and 634 00:21:32,960 --> 00:21:35,160 um 635 00:21:34,200 --> 00:21:37,080 it's I think it's I think it's 636 00:21:35,160 --> 00:21:38,280 impossible to tell. 637 00:21:37,080 --> 00:21:39,840 I think quite frankly, it's impossible 638 00:21:38,280 --> 00:21:41,160 to tell where anything is going beyond a 639 00:21:39,840 --> 00:21:42,480 few months out 640 00:21:41,160 --> 00:21:44,520 uh because the pace of change and the 641 00:21:42,480 --> 00:21:48,480 pace of development is like nothing I 642 00:21:44,520 --> 00:21:51,040 have seen in my career in technology. 643 00:21:48,480 --> 00:21:54,240 >> Well, and the one thing, 644 00:21:51,040 --> 00:21:56,600 you know, markup is it is is policing is 645 00:21:54,240 --> 00:21:58,880 is the it sounds to me like the 646 00:21:56,600 --> 00:22:00,960 avalanche of AI slop that that's that's 647 00:21:58,880 --> 00:22:04,480 coming out there, and and we all see it. 648 00:22:00,960 --> 00:22:06,880 It's there there's a bunch of stuff uh 649 00:22:04,480 --> 00:22:10,280 on your LinkedIn feeds, on your 650 00:22:06,880 --> 00:22:13,560 everywhere. And what is your, I guess, 651 00:22:10,280 --> 00:22:15,760 take on that? Like, how do you feel 652 00:22:13,560 --> 00:22:18,320 this AI slop problem 653 00:22:15,760 --> 00:22:20,720 uh will be 654 00:22:18,320 --> 00:22:22,920 I guess, dealt with? 655 00:22:20,720 --> 00:22:24,400 >> Um well, hopefully we deal with it. Uh 656 00:22:22,920 --> 00:22:26,280 and the world turns to us to deal with 657 00:22:24,400 --> 00:22:28,000 it. And uh you know, we built a business 658 00:22:26,280 --> 00:22:30,760 that's really easy to embed in lots of 659 00:22:28,000 --> 00:22:33,440 other solutions. We are not a monolithic 660 00:22:30,760 --> 00:22:35,160 stand-alone thing outside of workflows. 661 00:22:33,440 --> 00:22:36,640 Um but, um 662 00:22:35,160 --> 00:22:37,720 you know, uh but realistically, not 663 00:22:36,640 --> 00:22:39,400 everyone in the world is going to use 664 00:22:37,720 --> 00:22:43,600 us, and I don't know. I mean, it you 665 00:22:39,400 --> 00:22:44,960 know, AI automates bad behavior. Um and 666 00:22:43,600 --> 00:22:47,360 the reality is the internet automates 667 00:22:44,960 --> 00:22:49,520 bad behavior, too. Um and I'm not sure 668 00:22:47,360 --> 00:22:51,160 everything has been solved around that. 669 00:22:49,520 --> 00:22:54,720 And uh I don't think everything's going 670 00:22:51,160 --> 00:22:55,800 to be solved around AI slop, either. Um 671 00:22:54,720 --> 00:22:58,200 maybe the uh maybe the 672 00:22:55,800 --> 00:22:59,880 machine-to-machine reading will will 673 00:22:58,200 --> 00:23:03,080 either mean that it's not a problem 674 00:22:59,880 --> 00:23:04,640 because they can cut through slop, um or 675 00:23:03,080 --> 00:23:06,280 uh surfaces the problem more efficiently 676 00:23:04,640 --> 00:23:09,160 to solve. 677 00:23:06,280 --> 00:23:13,080 >> What do you think about AI influencers 678 00:23:09,160 --> 00:23:15,480 and user-generated content? 679 00:23:13,080 --> 00:23:16,920 >> Um, there are a couple of them that I 680 00:23:15,480 --> 00:23:19,720 follow, 681 00:23:16,920 --> 00:23:21,720 who have really smart things to say, and 682 00:23:19,720 --> 00:23:23,240 who definitely influence influence us 683 00:23:21,720 --> 00:23:24,600 here that are I would say they're you 684 00:23:23,240 --> 00:23:26,600 know sort of more 685 00:23:24,600 --> 00:23:27,880 power enterprise, you know, B2B 686 00:23:26,600 --> 00:23:30,960 influencers. 687 00:23:27,880 --> 00:23:32,400 I don't focus a lot on B2C AI 688 00:23:30,960 --> 00:23:33,920 influencers, and I don't know if there 689 00:23:32,400 --> 00:23:36,480 if there are any. There probably are 690 00:23:33,920 --> 00:23:38,040 some. Um, but you know, the thing with 691 00:23:36,480 --> 00:23:39,440 AI is 692 00:23:38,040 --> 00:23:41,960 again coming back to the previously 693 00:23:39,440 --> 00:23:43,560 unimaginable, it's really about power 694 00:23:41,960 --> 00:23:45,120 users exploring and pushing the 695 00:23:43,560 --> 00:23:46,360 boundaries and then recording what they 696 00:23:45,120 --> 00:23:47,916 did and what worked and what didn't 697 00:23:46,360 --> 00:23:49,000 work. So a couple of people 698 00:23:47,916 --> 00:23:52,400 [clears throat] that I pay attention to 699 00:23:49,000 --> 00:23:55,000 are like Camille Bach or Bank, B A N C. 700 00:23:52,400 --> 00:23:56,760 Um, Nate Jones. They both have really 701 00:23:55,000 --> 00:23:58,680 good substacks, 702 00:23:56,760 --> 00:23:59,960 um, where they just go into like either 703 00:23:58,680 --> 00:24:01,240 what they did something they did that 704 00:23:59,960 --> 00:24:02,560 worked, something they did that didn't 705 00:24:01,240 --> 00:24:03,800 work, or something a client of theirs 706 00:24:02,560 --> 00:24:06,200 tried. 707 00:24:03,800 --> 00:24:08,920 And those those end up being really 708 00:24:06,200 --> 00:24:08,920 really helpful. 709 00:24:09,200 --> 00:24:13,840 >> Matt, what's the next 710 00:24:11,680 --> 00:24:16,720 6 months to a year look like for Mark 711 00:24:13,840 --> 00:24:19,960 up, and what would you consider 712 00:24:16,720 --> 00:24:21,920 this 2026 being a great year capping 713 00:24:19,960 --> 00:24:23,480 off? What would happen there? 714 00:24:21,920 --> 00:24:26,120 >> Well, we launched our platform last 715 00:24:23,480 --> 00:24:27,480 week. So the rest of the year is all 716 00:24:26,120 --> 00:24:28,960 about adoption. 717 00:24:27,480 --> 00:24:30,640 And for us that's 718 00:24:28,960 --> 00:24:31,920 moving our legacy clients over as well 719 00:24:30,640 --> 00:24:33,800 as selling new clients. We're off to a 720 00:24:31,920 --> 00:24:35,600 great start on both. 721 00:24:33,800 --> 00:24:37,200 So it's early days and and therefore 722 00:24:35,600 --> 00:24:39,280 adoption matters and and obviously 723 00:24:37,200 --> 00:24:41,840 client impact and results. 724 00:24:39,280 --> 00:24:44,600 Um, and you know, our road map is is 725 00:24:41,840 --> 00:24:46,480 fairly straightforward. More surfaces, 726 00:24:44,600 --> 00:24:47,920 more agents and functionality, deeper 727 00:24:46,480 --> 00:24:50,640 memory, 728 00:24:47,920 --> 00:24:53,160 so that every client who uses us and 729 00:24:50,640 --> 00:24:56,320 adds another agent, another surface, or 730 00:24:53,160 --> 00:24:57,760 another employee to the team, uh 731 00:24:56,320 --> 00:24:59,960 capabilities are just better and better 732 00:24:57,760 --> 00:25:01,440 and better. 733 00:24:59,960 --> 00:25:02,760 >> Well, where can we keep up with you 734 00:25:01,440 --> 00:25:04,760 guys? 735 00:25:02,760 --> 00:25:06,920 What is your most active channel? And we 736 00:25:04,760 --> 00:25:07,840 definitely will have my audience follow 737 00:25:06,920 --> 00:25:09,600 you guys. 738 00:25:07,840 --> 00:25:13,440 >> Yeah, so Markup AI is the name of the 739 00:25:09,600 --> 00:25:15,280 company. My email is matt@markup.ai. 740 00:25:13,440 --> 00:25:16,640 LinkedIn is probably our most impactful 741 00:25:15,280 --> 00:25:19,600 channel, but we do have a blog that you 742 00:25:16,640 --> 00:25:22,480 can get off of the website markup.ai. 743 00:25:19,600 --> 00:25:25,400 And yeah, lots of lots of new product 744 00:25:22,480 --> 00:25:27,520 coming out every week. 745 00:25:25,400 --> 00:25:29,160 >> Thanks, Matt. It was a It was a pleasure 746 00:25:27,520 --> 00:25:31,160 having you on and I learned a lot from 747 00:25:29,160 --> 00:25:33,560 this interview. 748 00:25:31,160 --> 00:25:35,120 Please hang on after and we can have a 749 00:25:33,560 --> 00:25:36,080 little chat. 750 00:25:35,120 --> 00:25:38,400 >> Sounds good. Thanks for having me on, 751 00:25:36,080 --> 00:25:38,400 Johan.