1 00:00:02,800 --> 00:00:05,680 Michael, Mark, welcome to the Evolved Radio podcast. 2 00:00:06,960 --> 00:00:10,760 Thanks for having us. Thanks Todd. So 3 00:00:10,760 --> 00:00:14,400 glad to have you here. You know, there's been a lot of talk in the 4 00:00:14,400 --> 00:00:17,920 industry lately about everyone calling for 5 00:00:18,320 --> 00:00:21,920 the doom cycle of AI. There's a lot of conflicting 6 00:00:22,000 --> 00:00:25,720 opinions about whether or not this is working. So we 7 00:00:25,720 --> 00:00:29,520 wanted to kind of get into some of the details of like why that 8 00:00:29,520 --> 00:00:33,330 is potentially a debate. And Michael, maybe start 9 00:00:33,330 --> 00:00:36,970 with you. You have sort of an interesting perspective on this. Some of 10 00:00:36,970 --> 00:00:40,650 the things that you saw in your travels in the medical 11 00:00:40,650 --> 00:00:44,370 realm where you were before joining thread, do you want to give us kind of 12 00:00:44,370 --> 00:00:47,610 a bit of your backstory and how AI is related? 13 00:00:48,490 --> 00:00:52,210 Sure. To your point that where I've been in the past, 14 00:00:52,210 --> 00:00:55,690 I've spent the last 10 years or so inside the digital health space which 15 00:00:56,250 --> 00:01:00,010 has its own set of challenges outside of the AI piece. But 16 00:01:01,080 --> 00:01:04,840 yeah, for the past prior five years before joining thread, I was working 17 00:01:05,000 --> 00:01:07,560 at a, in a healthcare organization that was building 18 00:01:08,360 --> 00:01:11,800 agents on a kind of a more classic rules based 19 00:01:12,440 --> 00:01:16,160 structure versus what is really more commonplace 20 00:01:16,160 --> 00:01:18,920 today based on the large language models and etc. 21 00:01:20,200 --> 00:01:23,720 But yeah, look, 22 00:01:23,880 --> 00:01:27,560 setting aside kind of the unique challenges of operating in the health 23 00:01:27,560 --> 00:01:30,930 care space and kind of the 24 00:01:33,010 --> 00:01:36,770 different set of rigidity, our clients today in the space that Mark and I 25 00:01:36,770 --> 00:01:40,570 operate on have a set of standards and requirements that we need to 26 00:01:40,570 --> 00:01:44,290 make. They're just kind of uniquely different in healthcare. But I think to your point 27 00:01:44,290 --> 00:01:46,450 about at the beginning where you mentioned 28 00:01:48,450 --> 00:01:52,170 this doomsday and is AI really delivering on kind 29 00:01:52,170 --> 00:01:55,710 of the value prop? I think it really depends on how you 30 00:01:55,710 --> 00:01:59,470 frame the question of that. I think where it may be 31 00:01:59,470 --> 00:02:03,150 failing or perceived to be failing is where the promises 32 00:02:04,270 --> 00:02:07,950 were ahead of its ability to begin with. And 33 00:02:09,790 --> 00:02:13,470 the shiny promise is really easy to sell and that's easy to sell 34 00:02:13,470 --> 00:02:16,430 in almost any industry regardless of 35 00:02:18,350 --> 00:02:22,110 being healthcare or in more of a core technology space, 36 00:02:23,950 --> 00:02:27,070 a place where there's no doubt it's creating value is 37 00:02:27,950 --> 00:02:31,710 some of the back end systems, some of the analysis side of things. 38 00:02:32,350 --> 00:02:36,150 Where I think some of the challenges have come in both in healthcare 39 00:02:36,150 --> 00:02:39,670 and outside of healthcare, is where the 40 00:02:39,670 --> 00:02:43,470 rubber meets the road on real specificity of problem solving, 41 00:02:43,550 --> 00:02:47,110 solution solving, having the 42 00:02:47,110 --> 00:02:50,790 AI or in our case an agent being able to take 43 00:02:50,790 --> 00:02:54,470 an action that brings to resolution. And I think if you're able 44 00:02:54,470 --> 00:02:58,310 to frame the ability of the product and 45 00:02:58,310 --> 00:03:01,230 set the right promise rather than this, 46 00:03:02,430 --> 00:03:06,030 you know, North Star. Always looking at this North Star 47 00:03:06,750 --> 00:03:10,550 type situation, I think it's in a lot of cases been 48 00:03:10,550 --> 00:03:13,990 more successful than maybe we Give it credit or some of the recent 49 00:03:13,990 --> 00:03:17,510 articles. And also, no matter how you look at the 50 00:03:17,510 --> 00:03:21,260 speed at which it's, at which it's moving, I mean, 51 00:03:21,260 --> 00:03:24,500 Mark can certainly speak to, you know, what 52 00:03:24,740 --> 00:03:28,420 thread has been able to develop in other agentic companies 53 00:03:28,740 --> 00:03:32,380 in a much shorter period of time with a much 54 00:03:32,380 --> 00:03:36,100 higher degree of fidelity, a much higher degree of solutioning 55 00:03:36,260 --> 00:03:39,940 for the end user than we would have seen five years 56 00:03:39,940 --> 00:03:43,700 ago. So kind of to end that rant. Yeah, 57 00:03:43,700 --> 00:03:47,110 I think the promises have been too huge 58 00:03:47,350 --> 00:03:49,910 and kind of inset the wrong expectation. 59 00:03:51,110 --> 00:03:54,910 But the underlying, I think if you, you kind of said before we 60 00:03:54,910 --> 00:03:57,390 got on here, you have a little bit different view. And I think I'm with 61 00:03:57,390 --> 00:04:00,950 you. If the expectation is set correctly, it is. Has been incredible. 62 00:04:01,750 --> 00:04:05,430 The advances and the ability to solve day 63 00:04:05,430 --> 00:04:08,910 to day work problems are. It's quite 64 00:04:08,910 --> 00:04:12,710 remarkable. Yeah, it's, you know, the, 65 00:04:13,760 --> 00:04:17,160 the Gartner hype cycle sort of comes to mind here. We had chatted about this 66 00:04:17,160 --> 00:04:20,560 before. We were talking about coming on and like, I don't know that we're 67 00:04:20,560 --> 00:04:24,240 necessarily in the trough of disillusionment yet. Right. I think 68 00:04:24,240 --> 00:04:27,960 it's maybe on the downward trend. Like I think to your point, there 69 00:04:27,960 --> 00:04:31,440 was a lot of sort of unrealistic, unrealistic 70 00:04:31,440 --> 00:04:35,240 expectations of what, what these things were capable of. But at the 71 00:04:35,240 --> 00:04:39,040 same time, the part that I marry up with that is like how much we've 72 00:04:39,040 --> 00:04:42,510 blown past even the most basic expectations 73 00:04:42,670 --> 00:04:46,270 of what an AI should be capable of. Like I heard this conversation on a 74 00:04:46,270 --> 00:04:50,070 podcast and they were talking about like, hey, remember the Turing Test? Like no 75 00:04:50,070 --> 00:04:53,710 one, like that was hot for about five seconds as we ripped past 76 00:04:53,710 --> 00:04:57,470 it and no one seemed to really notice. Right. So like, like 77 00:04:57,470 --> 00:05:01,190 be in a way, like we're actually beyond, I think what the peak 78 00:05:01,190 --> 00:05:04,750 of inflated expectation would have been like. We shot so 79 00:05:04,750 --> 00:05:08,560 far past it that the, the, the drop to the trough of disillusionment 80 00:05:08,560 --> 00:05:12,120 is sort of like almost passing through the workable 81 00:05:12,120 --> 00:05:15,960 cycles, the workable workflows. But we're in this sort of 82 00:05:15,960 --> 00:05:19,560 weird, messy middle of. We expect. Maybe it should be capable of 83 00:05:19,560 --> 00:05:23,160 more because it's not AGI. I don't know. Like, Mark, 84 00:05:23,320 --> 00:05:26,920 what's your take on this? We had a great AI 85 00:05:26,920 --> 00:05:30,640 service unleashed weekly meeting today and I 86 00:05:30,640 --> 00:05:33,920 love your point about Turing Test because we've, 87 00:05:34,080 --> 00:05:37,720 we've now progressed. We are now at the Will Smith 88 00:05:37,720 --> 00:05:41,320 eating spaghetti test and it's 89 00:05:41,320 --> 00:05:45,160 insane. And we've all been like, oh, okay, it's okay, it's conscious and 90 00:05:45,160 --> 00:05:48,919 like spaghetti doesn't crunch and there's steam coming off 91 00:05:48,919 --> 00:05:52,680 the pasta. So we've, we've definitely blown 92 00:05:52,680 --> 00:05:55,600 past that significantly. 93 00:05:56,640 --> 00:06:00,480 What. And I think about this a lot and there's 94 00:06:00,480 --> 00:06:04,200 like two big concepts I want to convey. This is not the 95 00:06:04,200 --> 00:06:06,920 first AI cycle in human history. 96 00:06:08,120 --> 00:06:11,840 This is the second AI cycle. We have 97 00:06:11,840 --> 00:06:15,479 had a nuclear winter for AI after 98 00:06:15,479 --> 00:06:19,240 the first big realization, right. I forget the 99 00:06:19,240 --> 00:06:22,600 guy's name, but the first learning of neural 100 00:06:22,600 --> 00:06:26,360 networks came from a biologist that was also mathematician and he figured 101 00:06:26,360 --> 00:06:30,010 out the neural network and we made a massive leap 102 00:06:30,010 --> 00:06:33,250 and then it just went quiet. Right. And 103 00:06:33,650 --> 00:06:37,130 you know, to Michael's point, like we had to do a lot of 104 00:06:37,130 --> 00:06:40,850 rule based conversation flows 105 00:06:43,170 --> 00:06:46,730 and it went quiet and then the transformer 106 00:06:46,730 --> 00:06:49,090 model from Google came out. 107 00:06:49,890 --> 00:06:53,410 Hilarious that they did not monetize or productize that 108 00:06:53,830 --> 00:06:57,270 and OpenAI did. So I think we're on the cusp of another 109 00:06:57,990 --> 00:06:59,750 big, big 110 00:07:02,150 --> 00:07:05,950 uphill and just like this 111 00:07:05,950 --> 00:07:09,510 explosive growth with gen AI now 112 00:07:10,390 --> 00:07:14,070 there's a lot here. Humans always 113 00:07:14,150 --> 00:07:17,270 overspend. If anyone was around during.com 114 00:07:17,590 --> 00:07:21,220 bubble, man, did we overspend for 115 00:07:21,220 --> 00:07:24,860 pets.com to do something that was probably not 116 00:07:24,860 --> 00:07:28,620 worth it, right? But what, what 117 00:07:28,620 --> 00:07:31,980 was left behind that was infrastructure. What 118 00:07:31,980 --> 00:07:35,460 happened was we went from a thousand dollar T1 119 00:07:35,460 --> 00:07:39,100 lines to 50 DSL lines 120 00:07:40,060 --> 00:07:43,580 and that infrastructure is why Google exists, is 121 00:07:43,580 --> 00:07:47,030 why Amazon exists. So we are definitely 122 00:07:47,030 --> 00:07:50,630 overspending. The question is what's going to be the 123 00:07:50,630 --> 00:07:54,310 infrastructure that's left behind for us to build 124 00:07:54,310 --> 00:07:57,910 upon? Yeah, no, I think that's a really good 125 00:07:57,910 --> 00:08:00,870 prediction that like, you know, maybe we're kind of 126 00:08:00,870 --> 00:08:04,670 misunderstanding what the point is right now. Right. And this kind of 127 00:08:04,670 --> 00:08:07,670 goes to like, I, I kind of bring this up every time we're, we're talking 128 00:08:07,670 --> 00:08:11,470 about AI on the podcast, but you know, I won't do the full detail 129 00:08:11,550 --> 00:08:15,250 of this, but my, my AI sort of prediction timeline 130 00:08:15,250 --> 00:08:19,090 is now three or four years into this. And I 131 00:08:19,090 --> 00:08:22,810 predicted kind of three to five years was usable workflows 132 00:08:22,810 --> 00:08:25,490 and we're well into that. Like there's a lot of very useful 133 00:08:26,690 --> 00:08:30,330 creations, you know, thread certainly being one of them is a 134 00:08:30,330 --> 00:08:34,130 very reusable work workflow use case. And it provides value. Like 135 00:08:34,130 --> 00:08:37,890 those things absolutely exist and it moves on to sort 136 00:08:37,890 --> 00:08:41,640 of later on it becomes something that starts to potentially eat service 137 00:08:41,640 --> 00:08:45,080 desks after five to 10 years. And I think there's a lot of things that 138 00:08:45,080 --> 00:08:48,800 you guys are cooking up that lend to that vision. And 139 00:08:48,800 --> 00:08:51,720 it's not a doom and gloom type thing of like everyone's going to lose their 140 00:08:51,720 --> 00:08:54,800 jobs because AI is taking over. It Definitely pushes people to 141 00:08:55,200 --> 00:08:59,000 higher value work, right? It's commoditization of rote 142 00:08:59,000 --> 00:09:02,760 work within those ecosystems. So if that's the case, 143 00:09:02,760 --> 00:09:06,510 like that's part of what I'm curious about here and I have, I have some, 144 00:09:06,510 --> 00:09:09,510 some opinions I can definitely lend to this. But I'm sure 145 00:09:10,550 --> 00:09:14,350 your guys's perspective on, you know, like the HBR study that came out that 146 00:09:14,350 --> 00:09:17,950 said 81% of AI projects 147 00:09:17,950 --> 00:09:21,509 in business are considered a failure, right? Like, like 148 00:09:21,509 --> 00:09:25,190 why, if we're sort of seeing the early 149 00:09:25,270 --> 00:09:28,790 stages of this golden age of AI, why is everyone 150 00:09:28,870 --> 00:09:32,390 predicting that, you know, we've overstated this and you know, 151 00:09:32,470 --> 00:09:36,150 now everyone's like well crap, maybe we shouldn't have fired all those people 152 00:09:36,310 --> 00:09:39,330 at sales and Microsoft because now we have to hire 153 00:09:39,570 --> 00:09:43,290 80% of them back, right? Like what is the tension between those, 154 00:09:43,290 --> 00:09:47,010 like is it, is it oversold or you know, is, 155 00:09:47,090 --> 00:09:50,090 are we sort of missing the mark on what we're actually supposed to be achieving 156 00:09:50,090 --> 00:09:53,810 right now? I think it's probably a little bit of both 157 00:09:53,810 --> 00:09:57,170 and we have to be careful to not be a little bit of the 158 00:09:57,490 --> 00:10:01,090 self fulfilling in that. Right? If you're a big major company, 159 00:10:01,970 --> 00:10:05,730 leave them nameless. But any of them, or any company, large or small, who 160 00:10:06,210 --> 00:10:09,950 you make a commitment and you go out and you get ahead of your 161 00:10:09,950 --> 00:10:13,630 understanding of the practical application of 162 00:10:13,710 --> 00:10:17,230 the technology in this case, you know, agentic that we're talking about, 163 00:10:17,710 --> 00:10:21,310 then there's this pressure to live up 164 00:10:21,310 --> 00:10:24,990 to. Oh, you said it was going to replace, pick a department. You said it 165 00:10:24,990 --> 00:10:27,870 was going to replace your legal department. Well, 166 00:10:28,590 --> 00:10:32,430 you know, if you're a public company, you better show 167 00:10:32,430 --> 00:10:36,110 some evidence that that occurred because you've suggested it 168 00:10:36,510 --> 00:10:39,390 happened rather than, because I think because of the, 169 00:10:40,720 --> 00:10:44,360 some of the stuff we talked about, the promise is so 170 00:10:44,360 --> 00:10:47,120 huge and the reality is 171 00:10:48,240 --> 00:10:52,080 some of the stuff is so amazing that it's really easy 172 00:10:52,400 --> 00:10:55,920 to take a look and say look, it fundamentally can do this. Well, 173 00:10:56,320 --> 00:11:00,120 we can jump all the way to there. It's like, it's 174 00:11:00,120 --> 00:11:03,920 like almost, it's, it's almost human nature 175 00:11:05,890 --> 00:11:08,130 to kind of jump to the end of the, right to the end of the 176 00:11:08,130 --> 00:11:10,770 rainbow or, or whatever it is rather than 177 00:11:11,570 --> 00:11:15,170 understanding that it's not understanding, you know, being methodical 178 00:11:15,170 --> 00:11:18,690 that everything is a step process, right? You've got to, 179 00:11:20,210 --> 00:11:23,850 your legal team's got to know how to use it and where the value creation 180 00:11:23,850 --> 00:11:27,570 is before you could not have one or, or 181 00:11:27,570 --> 00:11:31,410 to the point you guys made, you said earlier, how do you up level 182 00:11:31,410 --> 00:11:35,060 skill set, how do you get These people in any department doing 183 00:11:35,140 --> 00:11:37,220 higher value creation things 184 00:11:38,980 --> 00:11:41,460 rather than kind of time, 185 00:11:42,580 --> 00:11:46,100 workforce burning less optimal things. 186 00:11:47,780 --> 00:11:51,460 And I love that. Can I, can I add on? I don't know if you. 187 00:11:52,660 --> 00:11:55,460 Man, there's two big things that are happening 188 00:11:55,860 --> 00:11:58,910 economically. We went crazy on 189 00:11:58,910 --> 00:12:01,630 cryptocurrency. We went through Covid. 190 00:12:02,510 --> 00:12:06,110 We've inflated the market. We 191 00:12:06,110 --> 00:12:09,870 needed a way to generate more value. I don't know what 192 00:12:09,870 --> 00:12:12,110 would have happened if we didn't get the gen 193 00:12:13,150 --> 00:12:16,270 inflection point. So I think big business 194 00:12:16,750 --> 00:12:20,270 needed a soft landing. Right. Or they needed a way 195 00:12:20,430 --> 00:12:23,550 of, of progressing forward. And I think, 196 00:12:24,240 --> 00:12:27,960 I think we've saved all the crypto farms because I'm pretty sure all the GPUs 197 00:12:27,960 --> 00:12:31,640 are now used for AI. But also there's 198 00:12:31,640 --> 00:12:35,360 a lot of inflation in the economy. There's a lot. And so we're trying 199 00:12:35,360 --> 00:12:39,080 to figure out unemployment 200 00:12:39,080 --> 00:12:42,640 is low, salaries are high. So what do we do? 201 00:12:43,040 --> 00:12:46,720 How do we, how do we fix the labor market? And I think there's this 202 00:12:46,800 --> 00:12:50,400 big demand for us right now to figure out the digital work 203 00:12:51,580 --> 00:12:55,340 because if we can do will heal our economy, at least in 204 00:12:55,340 --> 00:12:58,620 the United States, it will heal our economy. So I think with 205 00:12:59,420 --> 00:13:03,180 Harvard Business Review, they're obviously going after the large companies. So I think 206 00:13:03,180 --> 00:13:06,060 that's why we're seeing that pressure there. 207 00:13:08,060 --> 00:13:11,100 Yeah, I think, look, some of the challenges 208 00:13:11,820 --> 00:13:15,500 right now, are you looking at a very high level. 209 00:13:15,820 --> 00:13:19,660 There's a couple dozen companies driving this perceived. If 210 00:13:19,660 --> 00:13:23,120 you look at the market, this market explos. Whereas the 211 00:13:23,120 --> 00:13:26,680 reality for those major companies that have 212 00:13:26,680 --> 00:13:30,440 their point, that have their AI plan mapped out and 213 00:13:30,680 --> 00:13:34,400 they're building the data centers and they're building the hardware and 214 00:13:34,400 --> 00:13:37,960 they're building the processing power. Those are very clear. 215 00:13:38,120 --> 00:13:41,720 The vast majority of the companies in the US 216 00:13:42,200 --> 00:13:45,880 around the world, and certainly our customers, they're figuring out 217 00:13:46,120 --> 00:13:49,960 how to deploy it really effectively. And everyone uses it and 218 00:13:49,960 --> 00:13:53,400 touches it on their computer, they do searches with it, they do whatever. But from 219 00:13:53,400 --> 00:13:57,240 practical workflow utilization. What do you mean? We're taking my wedding 220 00:13:57,240 --> 00:14:00,400 photos and making them anime is not productive to the economy. 221 00:14:01,040 --> 00:14:04,760 Yeah, it's my favorite activity. We're still like the 222 00:14:04,760 --> 00:14:08,600 average company. The business model isn't as clear as an 223 00:14:08,600 --> 00:14:12,360 Nvidia or somebody who's building the data centers. In 224 00:14:12,360 --> 00:14:15,960 addition to that, we're still settling in. Mark mentioned the broader 225 00:14:15,960 --> 00:14:19,480 economy. Like what is it doing? Right. As these data centers 226 00:14:19,480 --> 00:14:22,520 absorb the electricity and power, there'll be 227 00:14:23,480 --> 00:14:27,200 other investments in new sources of power at the same time. 228 00:14:27,200 --> 00:14:30,800 Until those are there, we all might see our Electric bills go 229 00:14:30,800 --> 00:14:34,120 up. Right? Because supply and demand, right? 230 00:14:34,280 --> 00:14:37,960 So if there's that, or terraforming the. Planet 231 00:14:38,200 --> 00:14:40,760 to be a compositional brain. It's wild. 232 00:14:42,370 --> 00:14:46,090 It's insane. I do have one thing I really like. 233 00:14:46,090 --> 00:14:49,410 I need to get off my chest. Let him do it. 234 00:14:51,090 --> 00:14:54,530 There's a big difference between a proof of concept 235 00:14:55,250 --> 00:14:59,090 and production rate. There has not been 236 00:14:59,250 --> 00:15:02,370 that big of a difference in traditional software development. 237 00:15:02,850 --> 00:15:06,570 Like, sure, you haven't thought about all the workflows. You haven't thought about 238 00:15:06,570 --> 00:15:09,800 all the use cases, obviously. However, 239 00:15:10,520 --> 00:15:13,800 this is the first time we're talking about indeterministic 240 00:15:14,280 --> 00:15:17,640 code. This is code that's making its decisions. 241 00:15:18,760 --> 00:15:22,520 And it's a completely different game. It might look 242 00:15:22,520 --> 00:15:26,320 great, but every time when we first started to 243 00:15:26,320 --> 00:15:29,720 QA our voice agent, every time we'd pick up, 244 00:15:29,960 --> 00:15:33,680 what would we say? I have a printer problem. I have a printer problem. 245 00:15:33,680 --> 00:15:36,860 I have a printer. I was like, this is awesome. 246 00:15:37,580 --> 00:15:41,260 It's so good. But I never thought maybe it's just 247 00:15:41,420 --> 00:15:45,260 good at printer problems. Because when you get 248 00:15:45,260 --> 00:15:48,140 into production and it's not a printer problem, 249 00:15:50,620 --> 00:15:54,340 it's a lack of connectivity, it's 250 00:15:54,340 --> 00:15:58,100 a broken dll. It just acts 251 00:15:58,100 --> 00:16:01,740 completely differently. And so I think the 252 00:16:01,740 --> 00:16:05,500 combination of the, the. The pressure economically to 253 00:16:05,500 --> 00:16:09,300 find the next way to heal, but 254 00:16:09,300 --> 00:16:12,980 also seeing those little POCs, they're like, yeah, I'm 255 00:16:12,980 --> 00:16:15,780 Clara, baby. I'm gonna fire 700 people. 256 00:16:18,260 --> 00:16:22,060 And it's not going so well. Maybe we should reverse 257 00:16:22,060 --> 00:16:25,380 out of that. Because people actually have hard, 258 00:16:26,170 --> 00:16:29,610 hard questions. And it's not all about, like, I need a refund. 259 00:16:31,050 --> 00:16:34,690 Yeah. And this, this sort of gets to, like, my feeling. I have two things 260 00:16:34,690 --> 00:16:38,250 that I think are sort of at the core of. Of those. Those numbers 261 00:16:38,250 --> 00:16:41,970 of the percentages of failed AI projects in business. And 262 00:16:41,970 --> 00:16:45,290 one, I think is. Is sort of the majority of it is 263 00:16:45,770 --> 00:16:49,450 the historical percentage of failed 264 00:16:50,090 --> 00:16:53,240 IT projects in general is. Is 265 00:16:53,240 --> 00:16:57,040 catastrophically high. Right. So, like, sure 266 00:16:57,200 --> 00:17:01,040 to say 81% fail. But if we're starting from a base baseline of 267 00:17:01,040 --> 00:17:04,680 64% of IT projects fail, which is. Is a 268 00:17:04,680 --> 00:17:08,440 general stat, then we're really not that far off the mark. So I don't feel 269 00:17:08,440 --> 00:17:11,560 like this is as. As sort of as 270 00:17:11,560 --> 00:17:15,280 diminishing to AI as a project as a whole. In business, 271 00:17:15,440 --> 00:17:19,120 it's more to the point of AI, of IT is messy, 272 00:17:19,570 --> 00:17:23,250 and the implementation often doesn't go the way that we think because of poor project 273 00:17:23,250 --> 00:17:27,090 management, because of scope creep, because of people issues. 274 00:17:27,490 --> 00:17:31,250 And I think where that gets married. Up with it's always people issues. 275 00:17:33,970 --> 00:17:37,730 Project Management is as simple as who does what by when. And then you add 276 00:17:37,730 --> 00:17:41,530 people and things get complicated, right? So like I think 277 00:17:41,530 --> 00:17:45,010 the other core of this is that a lot of the AI, 278 00:17:46,230 --> 00:17:49,910 like well, I'm using air quotes here for the audio listeners. Like 279 00:17:50,150 --> 00:17:53,830 AI deployments and projects were largely just sort of, 280 00:17:53,830 --> 00:17:57,110 hey, we turned on Copilot for you guys. Go forth and prosper, 281 00:17:57,350 --> 00:18:01,150 right? And it was, or you know, we subscribed to Gemini. You 282 00:18:01,150 --> 00:18:04,230 know, you guys, you guys now have access to these tools and then you go 283 00:18:04,230 --> 00:18:07,670 back and you check with people who have not had any level of training 284 00:18:07,750 --> 00:18:11,510 and have no concept of how to utilize these things in their day to 285 00:18:11,510 --> 00:18:15,190 day operations. Like, hey, is that, that, how's that been? It's been 286 00:18:15,190 --> 00:18:18,590 amazing, right? Like it's totally revolutionized your work. And they're like, 287 00:18:19,150 --> 00:18:22,950 I got nothing. Like I didn't find it that useful. They're like, failed 288 00:18:22,950 --> 00:18:26,750 project, I guess. AI didn't save our business, right? So I think those 289 00:18:26,750 --> 00:18:30,430 are the things that are largely at work. And I've been even kind of amazed. 290 00:18:30,430 --> 00:18:34,110 You know, I've been working on some, some AI projects and 291 00:18:34,110 --> 00:18:37,390 as a part of this I've been interviewing a lot of MSPs and 292 00:18:37,390 --> 00:18:41,040 technologists and saying if you had an AI army to 293 00:18:41,040 --> 00:18:44,600 build some workflows for you to automate some things, to insert some 294 00:18:44,600 --> 00:18:48,360 agentic capabilities in your day to day operations to save you money 295 00:18:48,360 --> 00:18:51,520 and time, what would you do? And the level 296 00:18:51,840 --> 00:18:55,199 of sort of sophistication and creativity in those 297 00:18:55,199 --> 00:18:58,840 responses from highly technical people that are very aware of what their 298 00:18:58,840 --> 00:19:02,640 business is have been very routine. It's like really like 299 00:19:02,640 --> 00:19:05,780 we need to scratch a little deeper and think about some things that like that 300 00:19:05,780 --> 00:19:09,500 we could really sort of leverage and blow the doors off of the capabilities 301 00:19:09,500 --> 00:19:13,340 and allow people to do more. So if you assume 302 00:19:13,340 --> 00:19:17,020 that the IT people that are immersed in this all day and have a pretty 303 00:19:17,020 --> 00:19:20,780 good sense of what this stuff does, they can't come up with great use cases 304 00:19:21,020 --> 00:19:24,540 for how to deploy it and how to you how to leverage it for workflows. 305 00:19:24,700 --> 00:19:28,140 How are you going to expect that like Sarah from accounting 306 00:19:28,300 --> 00:19:32,020 and Jim in the marketing group are any 307 00:19:32,020 --> 00:19:35,860 further ahead in figuring out like exactly what they need from 308 00:19:35,860 --> 00:19:38,600 a workflow standpoint. And this is where I think like 309 00:19:39,480 --> 00:19:43,200 the core of this really comes down to is a lot of this kind 310 00:19:43,200 --> 00:19:46,800 of needs to be programmatic and determined because just to say to 311 00:19:46,800 --> 00:19:50,520 people you have AI, it will be amazing for your work 312 00:19:50,680 --> 00:19:54,280 ends up as a dead end, right? So I'll sort of 313 00:19:54,280 --> 00:19:57,360 end my rant there and leave you guys to sort of jump back in on 314 00:19:57,360 --> 00:20:01,080 this. What do you think, Mark? I haven't earned. We 315 00:20:01,080 --> 00:20:03,410 haven't earned secret here. 316 00:20:05,090 --> 00:20:08,610 The reason for our success is because we 317 00:20:08,610 --> 00:20:12,370 pair workflows with AI. 318 00:20:13,250 --> 00:20:17,050 It's not about telling people to use AI when they need it. You 319 00:20:17,050 --> 00:20:20,890 need to put it into the workflow. Right. The difference 320 00:20:20,890 --> 00:20:24,610 between telling accounts payable 321 00:20:24,610 --> 00:20:28,050 person, hey, I got you a Claude subscription. It's 322 00:20:28,050 --> 00:20:31,680 awesome. And telling them that every 323 00:20:31,680 --> 00:20:35,280 invoice that comes in will process added to the 324 00:20:35,280 --> 00:20:39,040 database. And I'll tell you at the end of the 325 00:20:39,040 --> 00:20:42,600 day what is the breakdown of invoices 326 00:20:42,600 --> 00:20:45,560 from vendors. Let me help you prioritize. 327 00:20:46,520 --> 00:20:50,200 So the key to success here is a workflow 328 00:20:51,160 --> 00:20:54,280 that's deterministic. Like this is a real use case. 329 00:20:55,130 --> 00:20:57,930 I need to figure like msps, man. 330 00:20:58,090 --> 00:21:00,170 Reconciling vendor 331 00:21:01,210 --> 00:21:05,050 invoices to billing to your customers is 332 00:21:05,050 --> 00:21:08,850 insane. Yep. And of course you can give them 333 00:21:08,850 --> 00:21:12,690 Claude, but you need to ingrain yourself in 334 00:21:12,690 --> 00:21:16,450 that workflow. You need to get every invoice 335 00:21:16,450 --> 00:21:19,970 and automatically process it and read the 336 00:21:19,970 --> 00:21:23,800 invoice and break it down. Like that's been, that's 337 00:21:23,800 --> 00:21:27,520 been so instrumental and I think that's what a lot of people aren't 338 00:21:27,520 --> 00:21:30,920 doing. You're right. It's so, it's hilarious. Co pilot. 339 00:21:31,240 --> 00:21:33,160 Great, Claude. Awesome. 340 00:21:34,920 --> 00:21:38,720 But it's not part of a workflow. And what Mark's talking about, when it really 341 00:21:38,720 --> 00:21:42,520 becomes just peripheral through society. 342 00:21:42,840 --> 00:21:46,560 Right. To where it's just, it's when somebody's 343 00:21:46,560 --> 00:21:50,180 used it all day and they don't realize they've used it because 344 00:21:50,980 --> 00:21:54,420 it's just culturally part where it is not. 345 00:21:54,500 --> 00:21:58,260 I'm taking a specific action that 346 00:21:58,260 --> 00:22:01,780 will engage this AI thing. 347 00:22:01,780 --> 00:22:05,460 Somebody said to me from my prior 348 00:22:05,540 --> 00:22:08,340 career, healthcare moves in decades, 349 00:22:10,020 --> 00:22:13,740 not in quarters or not in years. In 350 00:22:13,740 --> 00:22:17,250 a lot of ways. Everybody we talk to is still early 351 00:22:17,250 --> 00:22:19,850 adopter. Right. I mean to the average 352 00:22:20,330 --> 00:22:24,050 330 plus some a million Americans out there, our. Industry is 353 00:22:24,050 --> 00:22:27,650 an early adopter. Yeah. And the subset of people. 354 00:22:27,650 --> 00:22:31,410 Yeah. And if you stick to the decade is what creates 355 00:22:31,410 --> 00:22:35,130 change. I mean I think just about my kids, they're 356 00:22:35,370 --> 00:22:38,730 adults now and simple things 357 00:22:39,610 --> 00:22:43,460 like they want to know why I'd call 358 00:22:43,460 --> 00:22:47,060 instead of just text them. I mean it was only 359 00:22:47,140 --> 00:22:50,820 15 years that that wouldn't have been an option. So 360 00:22:50,820 --> 00:22:53,540 maybe not quite the decade. Mark's kids 361 00:22:54,340 --> 00:22:57,700 will not know a world where 362 00:22:57,700 --> 00:23:01,460 AI isn't influencing almost 363 00:23:01,620 --> 00:23:04,260 every piece of technology they touch. 364 00:23:05,700 --> 00:23:07,860 They won't. For good and for bad both. 365 00:23:10,260 --> 00:23:13,860 But again, kind of to my early where I started that 366 00:23:13,940 --> 00:23:17,700 it's about kind of to where 367 00:23:17,700 --> 00:23:19,380 Mark was going in that 368 00:23:22,580 --> 00:23:26,340 it's embedding the solution using AI solution 369 00:23:26,580 --> 00:23:30,220 in your day to day workflow. So it's just part of what 370 00:23:30,220 --> 00:23:33,700 you're doing. And that's when 371 00:23:34,180 --> 00:23:37,900 those big promises that maybe that article was suggesting the 372 00:23:37,900 --> 00:23:41,500 81% aren't achieving. That's when we start to. We see the 373 00:23:41,500 --> 00:23:45,180 reverse of that and it's 80% of them are achieving and maybe 374 00:23:45,180 --> 00:23:48,940 20% or not. It's also last thing I'll say on 375 00:23:48,940 --> 00:23:52,500 that. If we go really long term and we talk 376 00:23:52,500 --> 00:23:55,540 about Mark, you were talking about the way people think about 377 00:23:55,940 --> 00:23:59,380 planning a workflow or planning a solution. His kids 378 00:24:00,050 --> 00:24:03,730 at their age, they won't know that that's just how they will think. 379 00:24:05,090 --> 00:24:08,290 So everyone knows my kids are four and five. Oh. So yes. 380 00:24:09,010 --> 00:24:12,610 Minor adults. Yeah, but right. You're absolutely right. 381 00:24:12,690 --> 00:24:16,410 You're absolutely right. My kids like okay, you're tired of telling 382 00:24:16,410 --> 00:24:19,490 me stories. Can you have the AI tell me a story about Sonic? 383 00:24:20,530 --> 00:24:24,050 I'm like. That'S crazy. 384 00:24:24,290 --> 00:24:26,930 Like how did you come to that conclusion? 385 00:24:28,370 --> 00:24:31,890 You know, because it just. But I would say like this is really important. 386 00:24:32,210 --> 00:24:35,890 Like this is really important. It's not 387 00:24:35,890 --> 00:24:39,730 just AI, you need to get into workflow. If you 388 00:24:39,730 --> 00:24:42,210 don't get it into workflow, you will fail. 389 00:24:43,490 --> 00:24:45,730 It needs to be part of a workflow. 390 00:24:47,250 --> 00:24:49,890 So let's build on that because like, like, 391 00:24:51,170 --> 00:24:54,930 I mean I understand what I. Well maybe I'll preface that. I 392 00:24:54,930 --> 00:24:58,610 think I understand what you mean. Right. Of like, like we actually map out a 393 00:24:58,610 --> 00:25:02,210 full process. We understand what instigates a certain action 394 00:25:02,370 --> 00:25:06,050 or some. Something, something. An activity 395 00:25:06,210 --> 00:25:10,010 and then it leads through a bunch of actions and creates in some type 396 00:25:10,010 --> 00:25:12,770 of endpoint. Right. Like 397 00:25:13,650 --> 00:25:17,290 is that sort of all we're thinking about in a workflow? And understanding that 398 00:25:17,290 --> 00:25:20,970 enough is being more detailed and understanding what are the 399 00:25:20,970 --> 00:25:24,180 actions we take on initiation of something 400 00:25:24,500 --> 00:25:28,220 towards an end goal. Is that kind of what. What you're thinking about when you're 401 00:25:28,220 --> 00:25:31,300 talking about workflows? Yes. 402 00:25:32,020 --> 00:25:35,460 Very high level. Right? Yeah. Business process mining. 403 00:25:35,540 --> 00:25:39,220 Yes. Yeah. Right. Like there's a trigger. It's typically, 404 00:25:39,540 --> 00:25:43,340 you know, someone reaching out. There's a communication layer. But I 405 00:25:43,340 --> 00:25:46,740 think what the most important part that people aren't realizing 406 00:25:46,900 --> 00:25:50,620 is that if you map out, if you just forget AI if you 407 00:25:50,620 --> 00:25:54,140 just map out the whole business process, it starts to 408 00:25:54,140 --> 00:25:57,820 branch heavily towards the end. Right. Always it 409 00:25:57,820 --> 00:26:01,660 becomes wild. Oh, this customer is on this 410 00:26:01,660 --> 00:26:04,500 plan. Oh, this customer. We promised them something. 411 00:26:05,300 --> 00:26:09,100 We always promise. Right. Like everyone, everyone promises a customer 412 00:26:09,100 --> 00:26:12,780 something. And so what I have found is that the 413 00:26:12,780 --> 00:26:15,720 first two, three, four steps of a business process 414 00:26:16,360 --> 00:26:20,200 are typically linear, but then what 415 00:26:20,200 --> 00:26:23,880 happens is you hire a middle manager to be that, 416 00:26:24,680 --> 00:26:28,240 that branching point, like, oh, this needs to go here, or 417 00:26:28,240 --> 00:26:31,920 oh, we need to do this. What we're finding 418 00:26:31,920 --> 00:26:34,760 is you can, you can automate 1, 2, 3, 4 419 00:26:35,480 --> 00:26:38,600 and 5 is where you put your AI. 420 00:26:39,400 --> 00:26:42,790 5 is where you say, oh, we promised this to a 421 00:26:42,790 --> 00:26:46,590 customer. Do not bill them for this. 5 is where 422 00:26:46,590 --> 00:26:50,430 you say, this is a security issue. Get 423 00:26:50,430 --> 00:26:53,190 it over to the security team as soon as humanly possible. 424 00:26:54,470 --> 00:26:58,310 And I love William. I will give him a shout 425 00:26:58,310 --> 00:27:02,030 out forever. William from Integratech, he said a 426 00:27:02,030 --> 00:27:05,870 great line. AI is here. And we want to use 427 00:27:05,870 --> 00:27:09,600 AI to make service more human. It's not 428 00:27:09,600 --> 00:27:13,320 about responding back to the other person that's triggered the 429 00:27:13,320 --> 00:27:17,160 workflow, but it's about getting it to the right person 430 00:27:17,160 --> 00:27:19,840 in the right context with the right information. 431 00:27:20,800 --> 00:27:23,840 So like, that's been like our, our pillar. 432 00:27:25,120 --> 00:27:28,400 So. And this kind of goes to like, I think the way that 433 00:27:28,720 --> 00:27:32,440 maybe people are misusing the technology in a lot of ways. Like, I see this 434 00:27:32,440 --> 00:27:36,070 as very similar to sort of 435 00:27:36,070 --> 00:27:39,790 the other tools that get used and how people migrate away from them. 436 00:27:39,790 --> 00:27:43,310 We talked about sort of the generational changes of how people utilize 437 00:27:43,310 --> 00:27:47,110 technology. Right. And you know, I use Perplexity all day 438 00:27:47,110 --> 00:27:50,510 and I barely use Google anymore, and that's a fairly recent 439 00:27:50,510 --> 00:27:54,310 transition for me. But I know a lot of people, you know, their, their, 440 00:27:54,310 --> 00:27:58,110 their default is to, to still Google stuff. You know, I have these conversations with, 441 00:27:58,360 --> 00:28:02,080 with my wife and instead of her doing. I had the same conversation with my 442 00:28:02,080 --> 00:28:05,840 wife, by the way. It's hilarious. Yeah, like she'll Google stuff and then ask 443 00:28:05,840 --> 00:28:09,480 me to ask the AI. It's like, well, you know, you can do this too, 444 00:28:09,480 --> 00:28:13,000 right? Like, so I think there's a lot of that of people are somewhat 445 00:28:13,000 --> 00:28:16,640 comfortable with the way that they operate, the way that they deploy these technologies, the 446 00:28:16,640 --> 00:28:20,360 way that they think about the activities that they need. And honestly, one of the 447 00:28:20,360 --> 00:28:24,020 biggest shifts that I've had with AI was, you 448 00:28:24,020 --> 00:28:27,660 know, for years I was still using it like 449 00:28:27,740 --> 00:28:31,580 a search engine. Like I was sort of one or three shot prompting everything. 450 00:28:31,980 --> 00:28:35,780 And then a person made a very sort of simple suggestion to 451 00:28:35,780 --> 00:28:39,460 me. They're like, treat it like a co worker. And that was a huge unlock 452 00:28:39,460 --> 00:28:42,580 for me. And just the way that I interacted with AI and made it a 453 00:28:42,580 --> 00:28:46,180 lot more capable because the expectation was 454 00:28:46,180 --> 00:28:49,910 just slightly different about how I engaged with it and 455 00:28:50,310 --> 00:28:54,150 as I sort of explained this and we've had this conversation, I think 456 00:28:54,230 --> 00:28:58,070 I'm more, maybe more convinced, like, you guys opinion on this. 457 00:28:58,070 --> 00:29:01,830 Like, maybe we are on the slope of enlightenment and the Gartner 458 00:29:01,830 --> 00:29:05,350 hype cycle because now we have a better understanding of how to interact 459 00:29:05,350 --> 00:29:08,950 with and how to prescribe the technology's use, 460 00:29:09,190 --> 00:29:12,950 right? To just sort of say, here's a tool and you know 461 00:29:12,950 --> 00:29:16,030 you're going to do amazing work, right? Like, if you think of it in a 462 00:29:16,030 --> 00:29:18,870 very sort of, sort of pedestrian way of like, 463 00:29:19,780 --> 00:29:23,500 you know, here's a, here's a wrench, right, To a person who 464 00:29:23,500 --> 00:29:27,340 needs a hammer. And they're like, great, I don't really understand what this does 465 00:29:27,340 --> 00:29:29,580 for me. So I'm just going to go back and do what I was doing 466 00:29:29,580 --> 00:29:33,300 before. And like, they're like, he doesn't understand, doesn't understand how to use this 467 00:29:33,300 --> 00:29:36,900 tool, right? Like, this has been a total failure. Why wasn't this, 468 00:29:37,700 --> 00:29:41,020 you know, creating massive outcomes the way that we saw it? I think there's a 469 00:29:41,020 --> 00:29:44,860 lot of similarity to that. So I think your point about like, the definition of 470 00:29:44,860 --> 00:29:48,400 understanding how the work actually flows through the organization 471 00:29:48,400 --> 00:29:51,840 and where are the intersections where we can actually start to insert 472 00:29:52,080 --> 00:29:55,520 some of these agentic actions and utilize 473 00:29:55,760 --> 00:29:59,560 AI more as a co worker or a team member than 474 00:29:59,560 --> 00:30:02,960 just sort of a tool unto itself. Any thoughts on that? 475 00:30:02,960 --> 00:30:06,520 Yes. I have thoughts. 476 00:30:06,520 --> 00:30:10,080 Michael, you want to go first or should I? You go, I'll do cleanup on 477 00:30:10,080 --> 00:30:13,900 this one. Yeah, that's absolutely 478 00:30:13,900 --> 00:30:17,580 right. And I think, you know, we just had our keynote 479 00:30:19,020 --> 00:30:22,660 and we get this question a lot because Thread 480 00:30:22,660 --> 00:30:26,300 is becoming a little notorious for automating, dispatching work. 481 00:30:27,820 --> 00:30:31,100 We are for better or for worse. And 482 00:30:31,420 --> 00:30:34,820 folks are asking us like, well, are you scared to talk to 483 00:30:34,820 --> 00:30:38,230 MSPs? Like, aren't their dispatchers scared to get Thread? 484 00:30:39,590 --> 00:30:42,710 And we had this deep reflection where we realized 485 00:30:44,150 --> 00:30:47,510 where AI is today. We're not replacing 486 00:30:47,510 --> 00:30:51,270 100% of any job, of any role. 487 00:30:52,070 --> 00:30:55,790 What we're seeing is that we 488 00:30:55,790 --> 00:30:59,430 are moving to a supervisor relationship with the AI. 489 00:30:59,430 --> 00:31:03,270 It is a coworker, but it's very important to understand it's a supervisor 490 00:31:03,590 --> 00:31:07,390 relationship. And when I found out that Waymo was doing, 491 00:31:07,390 --> 00:31:10,870 it's all, it's, you know, autonomous driving, but 492 00:31:10,870 --> 00:31:14,670 behind all those cars, there's a person with an 493 00:31:14,670 --> 00:31:17,510 Xbox controller looking at nine cars. 494 00:31:18,550 --> 00:31:22,350 And if any of them get stuck, they use an Xbox controller to get 495 00:31:22,350 --> 00:31:26,190 out. Now, is it better than 496 00:31:26,190 --> 00:31:28,390 hiring nine drivers? Absolutely, 497 00:31:29,600 --> 00:31:33,240 absolutely. So I think the role, what's, what's Happening. 498 00:31:33,240 --> 00:31:36,240 And again, like I said, like, we'll see 499 00:31:36,960 --> 00:31:40,720 how many, how many more inflection points we have. But the relationship 500 00:31:40,720 --> 00:31:44,400 is a co working relationship, but it's a supervisory 501 00:31:44,480 --> 00:31:48,240 relationship. We're not going to get rid of all the dispatchers. We're going to 502 00:31:48,240 --> 00:31:52,000 make the dispatchers that are the best, the supervisors. 503 00:31:52,320 --> 00:31:55,880 We're not going to get rid of the technicians. We're going to make the best 504 00:31:55,880 --> 00:31:59,600 technicians supervise their little army, 505 00:31:59,760 --> 00:32:03,400 their digital workforce. Because you're right, they can reason, 506 00:32:03,400 --> 00:32:07,160 they can talk, they can't innovate. And the world has 507 00:32:07,160 --> 00:32:10,679 entropy. So how do we handle entropy? Humans are great at 508 00:32:10,679 --> 00:32:14,400 entropy. So I believe it's 509 00:32:14,400 --> 00:32:18,200 really important for us to recognize this. I 510 00:32:18,200 --> 00:32:21,840 believe that jobs will be automated, like jobs to be done, 511 00:32:22,160 --> 00:32:25,680 like prioritizing, categorizing or driving someone 512 00:32:26,160 --> 00:32:28,960 a mile. But I don't think roles 513 00:32:30,400 --> 00:32:33,520 will be fully destroyed. I think roles are going to become 514 00:32:33,520 --> 00:32:36,640 supervisory roles. I like that vision. 515 00:32:37,600 --> 00:32:41,040 I think, look to your point about where are we on the curve? If we 516 00:32:41,040 --> 00:32:42,080 started that curve, 517 00:32:45,120 --> 00:32:48,330 we're, we're really early because 518 00:32:49,450 --> 00:32:53,250 even the really early adopters still, I'll still 519 00:32:53,250 --> 00:32:57,050 get on an engineering stand up with Mark and the team 520 00:32:57,050 --> 00:32:59,770 and there'll be a new release from one of the large models. 521 00:33:00,970 --> 00:33:04,610 And we're still saying that group who live and breathe 522 00:33:04,610 --> 00:33:07,610 it. Holy crap. Look at, look at what this one does. 523 00:33:08,570 --> 00:33:11,610 You know, just the release three months ago 524 00:33:12,640 --> 00:33:16,080 didn't do X in 18 months. Before that, 525 00:33:16,960 --> 00:33:20,400 you know, seemed like an, you know, it's the difference between an 526 00:33:20,400 --> 00:33:24,240 infant to a teenager and kind of its, its 527 00:33:24,240 --> 00:33:26,880 capabilities. And that is, that's people 528 00:33:27,840 --> 00:33:31,520 like Mark's engineering and product team and the company's team 529 00:33:31,840 --> 00:33:35,680 living and breathing it. Whereas look, 530 00:33:35,680 --> 00:33:39,330 I, I think about, you know, the 531 00:33:39,330 --> 00:33:43,010 masses of people think about, just go through my family. I've got a lawyer 532 00:33:43,010 --> 00:33:46,650 and my brother's a school teacher and one's a lawyer and one's a 533 00:33:46,970 --> 00:33:50,250 works in retail. And like they are not, 534 00:33:52,010 --> 00:33:55,490 they are. Not medical industry radiologists. 535 00:33:55,490 --> 00:33:59,330 Yeah. They're just not what a radiologist organization standpoint. Like they're 536 00:33:59,330 --> 00:34:02,330 not, they don't, they don't yet have the 537 00:34:02,330 --> 00:34:05,970 opportunity to. And if they do, it's it's out of human 538 00:34:05,970 --> 00:34:09,530 curiosity. Right. You know what I mean? And to your point about. 539 00:34:10,010 --> 00:34:13,330 Yeah. Tinkering and to point about, about specific 540 00:34:13,330 --> 00:34:16,770 workflows like my brother who has been a 541 00:34:16,770 --> 00:34:20,610 schoolteacher for years, you know, we all know about the 542 00:34:20,610 --> 00:34:24,170 challenges of teachers. Like did the paper get written? Did they write themselves? And 543 00:34:24,250 --> 00:34:27,450 you know, so there's a whole other set of challenges. So certainly 544 00:34:28,010 --> 00:34:31,860 he is. He is teaching himself and 545 00:34:31,860 --> 00:34:35,500 educating himself on those practical applications, on how 546 00:34:35,500 --> 00:34:39,260 it's impacting his environment and being sure that 547 00:34:39,260 --> 00:34:42,660 like, you know, kids still have license 548 00:34:42,900 --> 00:34:46,620 to learn and that kind of thing. But again, it's. It's fairly 549 00:34:46,620 --> 00:34:49,900 isolated to the more not the 550 00:34:49,900 --> 00:34:53,700 advanced stuff. So to your point of the book, I 551 00:34:53,940 --> 00:34:57,540 Super early I think in that. In that curve 552 00:34:57,700 --> 00:35:01,540 maybe not as early for the marks of the world even 553 00:35:01,540 --> 00:35:04,900 for me. Todd, I want to flip it back on you. I have a question 554 00:35:04,900 --> 00:35:08,500 for you. I have. I want to know your thoughts on this, if that's 555 00:35:08,500 --> 00:35:11,980 okay. Sure. You know, Jensen is 556 00:35:11,980 --> 00:35:15,300 obviously the CEO of Nvidia is. Yeah. 557 00:35:16,820 --> 00:35:20,020 Fixing shovels. Win any gold rush. Yep. 558 00:35:20,830 --> 00:35:24,510 And he said a quote and I was like, yes, we are not forgotten. 559 00:35:25,550 --> 00:35:29,390 The IT department of every company is going 560 00:35:29,390 --> 00:35:33,150 to be the HR department of AI 561 00:35:33,150 --> 00:35:36,750 agents. Ooh, I like that. 562 00:35:36,910 --> 00:35:40,510 Yeah. TNHR have been new 563 00:35:40,510 --> 00:35:44,230 user onboarding termination. I. I just. I would love to 564 00:35:44,230 --> 00:35:48,070 hear your thoughts about that. Yeah. In Jensen's 565 00:35:48,070 --> 00:35:51,910 a brilliant dude. There's a lot I disagree with about Jensen's 566 00:35:51,910 --> 00:35:55,710 management style, which is a separate podcast seems to have worked 567 00:35:55,710 --> 00:35:59,390 for him. But holy hell, the. That 568 00:35:59,390 --> 00:36:03,070 image I think is really smart because if I 569 00:36:03,070 --> 00:36:06,790 pull that back to something that's a bit more familiar right now, and that being 570 00:36:06,790 --> 00:36:10,550 the extrapolation from IT is outsourcing has become a lot 571 00:36:10,550 --> 00:36:14,270 more practical for IT organizations post Covid because 572 00:36:14,270 --> 00:36:17,870 it just removed all the geographic barriers on where people work 573 00:36:17,870 --> 00:36:21,510 from. And one of the things that I insisted very early on 574 00:36:21,510 --> 00:36:25,030 when I encouraged people to consider outsourcing was 575 00:36:25,030 --> 00:36:28,870 that they said, oh, I've tried that. It doesn't work. I'm like, yeah, 576 00:36:28,870 --> 00:36:32,670 but you probably abdicated a lot of the responsibility. And you 577 00:36:32,670 --> 00:36:36,470 know, hey, we hired this person. They should work remotely. They're good 578 00:36:36,470 --> 00:36:40,110 at what they do. And you know, none of that happened. But 579 00:36:40,110 --> 00:36:43,790 you never spent any time with them. You didn't treat them like a staff member. 580 00:36:43,790 --> 00:36:47,510 So of course the results were different. So the 581 00:36:47,510 --> 00:36:51,150 extrapolation of that I think is now that we'll have an 582 00:36:51,150 --> 00:36:54,830 AI workforce, we have to understand the. 583 00:36:55,150 --> 00:36:58,870 The workflow analysis and kind of figuring out 584 00:36:58,870 --> 00:37:02,590 like what the business patterns are and therefore how we can insert AI into it. 585 00:37:02,750 --> 00:37:06,210 So I think there's a very natural marriage there of like the business 586 00:37:06,290 --> 00:37:09,890 consulting aspect of it is becoming more prevalent because 587 00:37:10,050 --> 00:37:13,810 the, you know, the just the break fix and the support requirements 588 00:37:13,810 --> 00:37:17,610 are lower. And you know, once the AI start to 589 00:37:17,610 --> 00:37:21,370 live inside the. The business's analysis gets complete and figure out 590 00:37:21,370 --> 00:37:25,050 sort of where the agentic capabilities are and you know, building on 591 00:37:25,050 --> 00:37:28,690 those workflows, there still needs someone to kind of manage and operate 592 00:37:28,690 --> 00:37:32,210 those like from that supervisory position. Right. To hire them. 593 00:37:32,610 --> 00:37:36,370 Yeah, exactly. Like we, we understand sort of like which agent 594 00:37:36,370 --> 00:37:40,050 is particularly good at this type of role. And you know, 595 00:37:40,450 --> 00:37:44,170 we manage some of the exceptions based on sort of a decision tree 596 00:37:44,170 --> 00:37:47,890 because stuff happens, right. Work and people are messy and 597 00:37:48,050 --> 00:37:51,810 you know, how do we step in when, when someone requires a bit of intervention 598 00:37:51,810 --> 00:37:55,410 on that when it's dealing with, you know, an AI 599 00:37:55,890 --> 00:37:59,330 team member. So I, I think that's, I think that's a cool vision. I really 600 00:37:59,330 --> 00:38:00,770 like that as an, as an idea. 601 00:38:03,750 --> 00:38:07,550 I couldn't agree more. Right. And there's so much opportunity for us. This 602 00:38:07,550 --> 00:38:11,390 is why there's so much excitement in the MSP industry. Like we 603 00:38:11,390 --> 00:38:14,790 are going to be the HR for, for agents. 604 00:38:16,789 --> 00:38:20,590 And I, I just want to, you know, Todd, please, let's keep going. 605 00:38:20,590 --> 00:38:24,150 But I just love the relationship between IT 606 00:38:24,230 --> 00:38:27,960 and hr. Who's working and what tools do 607 00:38:27,960 --> 00:38:31,760 they have to get leverage in their work. Where the tools 608 00:38:31,920 --> 00:38:35,640 where the fire. Currently. I agree. Yeah. But look 609 00:38:35,640 --> 00:38:39,040 what's happening in the industry. Rippling deal just 610 00:38:39,040 --> 00:38:42,640 works. They're all starting to launch little like 611 00:38:42,880 --> 00:38:46,240 device management, onboarding and off board. 612 00:38:47,120 --> 00:38:50,920 I think that's going to be apart from the AI curve. I think that 613 00:38:50,920 --> 00:38:54,440 the marriage between IT and HR is going to be an interesting one. 614 00:38:55,240 --> 00:38:59,040 So I mean, obviously I could talk with you guys for hours on 615 00:38:59,040 --> 00:39:02,120 this. I want to look. To sort of close it up. I'll sort of drop 616 00:39:02,120 --> 00:39:05,720 one more here. That is really outside of sort of the 617 00:39:05,800 --> 00:39:09,160 basics of what we're talking about here. But Mark, I'm sure you've put some thought 618 00:39:09,160 --> 00:39:12,040 to this because I know that you live, eat and breathe this stuff. 619 00:39:12,760 --> 00:39:16,520 What are your thoughts on RL AI like, especially 620 00:39:16,520 --> 00:39:20,040 as we start to embody AI into RL 621 00:39:20,920 --> 00:39:24,720 real life. Like, like robots basically being 622 00:39:24,720 --> 00:39:28,200 human AI and starting to interact with, with the real world. 623 00:39:29,160 --> 00:39:32,880 You know, I, I've been thinking a lot about this recently. You have any, 624 00:39:32,880 --> 00:39:36,680 any thoughts on, on sort of the, the putting AI out 625 00:39:36,680 --> 00:39:39,240 into the world and having it interact in that fashion? 626 00:39:41,960 --> 00:39:45,400 We've, we've run out of training data. There's a big problem. 627 00:39:45,640 --> 00:39:48,750 That's why I'm saying the last AI winter 628 00:39:49,070 --> 00:39:52,190 perhaps may not be the only one 629 00:39:52,510 --> 00:39:55,910 we've run out of training data because the 630 00:39:55,910 --> 00:39:59,630 LLMs that we have today have learned everything they possibly can 631 00:40:00,270 --> 00:40:03,990 from all of our Reddit posts and all the junk we've put online. It's 632 00:40:03,990 --> 00:40:06,910 petabytes. It's great, but we're running out of data. 633 00:40:07,950 --> 00:40:11,430 I think there's two angles here. I think that AI needs 634 00:40:11,430 --> 00:40:15,250 eyes. It needs to, it needs to. You 635 00:40:15,250 --> 00:40:18,570 know, our eyesight consumes more information 636 00:40:19,370 --> 00:40:22,330 per second than any other of our senses. 637 00:40:23,050 --> 00:40:26,730 So I think there's, there will be an effort to get 638 00:40:26,730 --> 00:40:30,330 more information. So I think it's a natural progression of, of giving 639 00:40:30,330 --> 00:40:34,130 it senses in a certain sense. And 640 00:40:34,130 --> 00:40:37,210 I think that's really interesting. Now 641 00:40:37,930 --> 00:40:41,400 what I'm really excited about is I think, 642 00:40:41,400 --> 00:40:44,440 manufacturing. And 643 00:40:45,400 --> 00:40:49,160 I'll go back like after we had a disaster in New York, 644 00:40:50,760 --> 00:40:54,440 you know, the MSP I worked for, we bought the Sprint Data center 645 00:40:54,440 --> 00:40:57,400 in Purchase, New York, and we ran a data center 646 00:40:59,560 --> 00:41:02,600 and there's like predictive hardware failures 647 00:41:03,160 --> 00:41:06,930 and you just need to swap out a hard drive from a SATA drive. It's 648 00:41:06,930 --> 00:41:10,690 like we could figure it out. We just need someone to do it. 649 00:41:12,370 --> 00:41:16,130 So I, I think that on the manufacturing side and on data center 650 00:41:16,130 --> 00:41:19,330 side, I think we're going to see some of those, 651 00:41:20,450 --> 00:41:24,090 some of those things automated that are not as important as, for 652 00:41:24,090 --> 00:41:27,690 example, figuring out the energy problem or figuring out how to get 653 00:41:27,690 --> 00:41:31,250 smaller in our silicon manufacturing or get to 654 00:41:31,250 --> 00:41:35,040 quantum computing. But I, I 655 00:41:35,040 --> 00:41:38,760 think there's a desire to feed more information into the 656 00:41:38,760 --> 00:41:42,600 AI because we've run out. We have a big problem. We've run out. We're 657 00:41:42,600 --> 00:41:46,000 now using synthetic data, but also to help on that 658 00:41:46,000 --> 00:41:48,800 manufacturing side. Michael, what do you think? 659 00:41:50,000 --> 00:41:52,560 I'm gonna go more human, philosophical. 660 00:41:54,480 --> 00:41:58,040 That's what we work together, Michael. Perfect. I 661 00:41:58,040 --> 00:42:01,560 think if those robots, humanoids, 662 00:42:01,560 --> 00:42:05,120 real life, play a role in where Mark 663 00:42:05,120 --> 00:42:08,360 went, where the agent, where it's inserting and improving, 664 00:42:09,880 --> 00:42:13,640 obviously huge value there. I think on a human level, 665 00:42:14,040 --> 00:42:17,880 I think we learned a little bit from COVID in those years. We 666 00:42:17,880 --> 00:42:21,720 as people get a lot from interaction and being around each other. 667 00:42:22,520 --> 00:42:26,200 So if it over a decade or more would create an environment 668 00:42:26,280 --> 00:42:30,130 where we're doing less interaction with each 669 00:42:30,130 --> 00:42:33,890 other, I think that's not great for 670 00:42:33,890 --> 00:42:36,930 us just as people and as 671 00:42:37,090 --> 00:42:40,130 relationship. I've kept referring to kids in the generational. 672 00:42:41,250 --> 00:42:44,690 I mean, I'm not that old, but I mean, I can remember when my 673 00:42:44,690 --> 00:42:48,050 daytime activity was go outside and play with your friends. 674 00:42:48,610 --> 00:42:52,250 Like that was what after school was. And then one 675 00:42:52,250 --> 00:42:55,910 generation back there, the video games and all, which is, which is great. 676 00:42:55,910 --> 00:42:59,310 And it's developed our minds in different ways. But 677 00:43:00,750 --> 00:43:03,150 on the human level, I hope we deploy them 678 00:43:05,870 --> 00:43:09,470 where it makes us better and not 679 00:43:09,470 --> 00:43:13,230 the other way around. Yeah, I share that sentiment. 680 00:43:13,390 --> 00:43:16,910 As a technologist and someone who is very excited about this stuff, 681 00:43:17,070 --> 00:43:20,750 I learned in Covid that despite the fact that I consider myself 682 00:43:20,750 --> 00:43:24,600 an introvert once it was no longer an option to interact with 683 00:43:24,600 --> 00:43:28,280 people, it affected me. I think you're absolutely right that 684 00:43:28,280 --> 00:43:32,040 there. There needs to be. Still needs to be humanity in this. And I 685 00:43:32,040 --> 00:43:35,720 think that'll be a tricky. A tricky tightrope for us as well as we start. 686 00:43:35,960 --> 00:43:39,560 I love how. I love the angle that Michael took. And 687 00:43:39,560 --> 00:43:43,080 Michael knows this. I don't have TVs. My kids have 688 00:43:43,560 --> 00:43:47,320 zero screen time. I 689 00:43:47,320 --> 00:43:50,870 understand. Considering how nerdy you are. Yeah, no, I understand. 690 00:43:52,390 --> 00:43:54,950 I understand the problem. Yeah. 691 00:43:56,710 --> 00:44:00,470 But I also, you know, like, maybe we can manufacture 692 00:44:00,470 --> 00:44:04,110 stuff easier without people getting their fingers cut off. That's where 693 00:44:04,110 --> 00:44:07,950 I'm like, that's important. Or a humanoid going down 694 00:44:07,950 --> 00:44:11,670 and just sort of, like, pulling dead GPUs, slamming in a new one to the 695 00:44:11,670 --> 00:44:15,350 array and moving down the line, basically. Definitely. Yeah, definitely. 696 00:44:16,160 --> 00:44:18,880 Cool. Well, this has been really fun, you guys. Appreciate you coming on. 697 00:44:19,840 --> 00:44:23,520 We'll link to everything thread in the show notes and 698 00:44:24,000 --> 00:44:27,320 link to you guys on LinkedIn if anyone wants to reach out, but appreciate your 699 00:44:27,320 --> 00:44:30,480 time. Thank you. Thanks so much. All right, people.