1 00:00:02,080 --> 00:00:05,760 John, welcome to the Evolved Radio Podcast. Thank you so much for having 2 00:00:05,760 --> 00:00:09,320 me. This is really exciting. So a good place to start, I think, is, is 3 00:00:09,320 --> 00:00:12,920 your origin story. I think you have a really fascinating journey and I 4 00:00:12,920 --> 00:00:16,720 think maybe relatable to a lot of people, but somewhat unconventional. 5 00:00:16,720 --> 00:00:19,160 You want to kind of give us your background and how you got to be 6 00:00:19,160 --> 00:00:23,000 where you are? Sure. So I found myself at one point early in 7 00:00:23,000 --> 00:00:26,160 my life working in an aluminum factory, Alcan. 8 00:00:26,620 --> 00:00:30,460 Aluminum Canada for 12-hour shifts, very difficult physical 9 00:00:30,460 --> 00:00:34,180 heavy work, nothing to do with computers, forklifts and 10 00:00:34,180 --> 00:00:37,420 machines and very dangerous heavy operations. 11 00:00:37,900 --> 00:00:41,300 After a few years of that, I decided to make arrangements with the union to 12 00:00:41,300 --> 00:00:44,620 go back to St. Lawrence College in Kingston, Ontario for 13 00:00:44,860 --> 00:00:48,660 the 3-year computer programmer analyst program. So 14 00:00:48,660 --> 00:00:51,780 I managed to give up my day shifts for night shifts so that I could 15 00:00:51,780 --> 00:00:55,490 learn programming. But what was interesting is my placement out 16 00:00:55,490 --> 00:00:59,090 of that college was more in IT, more of a 17 00:00:59,090 --> 00:01:02,730 networks and desktops and Windows as opposed to 18 00:01:02,730 --> 00:01:06,410 programming. So I pursued that aspect of 19 00:01:06,490 --> 00:01:10,330 technology with certifications and eventually I did get on 20 00:01:10,330 --> 00:01:14,170 as a senior network engineer at an insurance company. Ultimately, I 21 00:01:14,170 --> 00:01:17,690 ended up as a senior network architect for the Parliament of Canada in 22 00:01:17,690 --> 00:01:21,520 Ottawa. 50 buildings, a greenfield network, 3 23 00:01:21,520 --> 00:01:25,120 data centers, 500 remote sites, national 24 00:01:25,440 --> 00:01:29,240 importance, the network, right? So that is, 25 00:01:29,240 --> 00:01:32,560 that was really where I learned a lot and 26 00:01:32,880 --> 00:01:36,000 I started to embrace network automation pretty early 27 00:01:36,560 --> 00:01:40,360 because building that network was one thing, but 28 00:01:40,360 --> 00:01:44,120 operating it day to day was an entirely different problem. And the 29 00:01:44,120 --> 00:01:47,440 scale and complexity really required us to embrace 30 00:01:47,440 --> 00:01:50,360 automation and doing things kind of programmatically on the network. 31 00:01:51,400 --> 00:01:55,200 So I self-published a book called Automate Your Network, and that's sort of when 32 00:01:55,200 --> 00:01:58,760 I kind of tried to make a scene on the 33 00:01:58,760 --> 00:02:02,600 social networks and LinkedIn and Twitter, and it all sort of revolved 34 00:02:02,600 --> 00:02:06,240 around the first book I self-published about network 35 00:02:06,240 --> 00:02:10,080 automation. And then I second, you know, a few years 36 00:02:10,080 --> 00:02:13,910 later, I published a second book through Cisco Press co-authored 37 00:02:14,270 --> 00:02:17,470 a book about Cisco PyETS, another network automation framework. 38 00:02:18,030 --> 00:02:21,750 In terms of when AI enters the picture for me, I had access 39 00:02:21,750 --> 00:02:25,230 to ChatGPT 3.5 in November of 2022, 40 00:02:26,430 --> 00:02:29,710 but what it really changed for me was when I applied for an API key 41 00:02:30,110 --> 00:02:33,670 and started connecting the artificial intelligence to my network 42 00:02:33,670 --> 00:02:37,230 automation work. And I've been obsessed with 43 00:02:37,230 --> 00:02:40,950 AI ever since. Every, you know, the, the introduction of 44 00:02:40,950 --> 00:02:44,390 RAG, the introduction of plugins for ChatGPT, 45 00:02:45,350 --> 00:02:48,790 CLIs, extensions, plugins, 46 00:02:49,270 --> 00:02:52,790 agents, LangChain, LangGraph, all of it. 47 00:02:52,950 --> 00:02:56,670 Every single aspect of artificial intelligence I've been trying 48 00:02:56,670 --> 00:03:00,310 to consume and share publicly my journey 49 00:03:00,550 --> 00:03:04,230 as I'm learning these things and as I'm trying to apply it to my trade. 50 00:03:04,920 --> 00:03:08,120 Which is network automation, right? And how it fits there. 51 00:03:08,600 --> 00:03:12,160 And it's really taken me places I never thought I would go. It's been an 52 00:03:12,160 --> 00:03:15,760 exciting journey. You have this term that 53 00:03:15,760 --> 00:03:19,560 I really like and kind of, I assume is riffing off of the 54 00:03:19,720 --> 00:03:23,400 DevOps scene of Vibops. Do you want to give us kind of 55 00:03:23,480 --> 00:03:26,680 your breakdown on this? Is this something that you coined or is this just something 56 00:03:26,680 --> 00:03:30,360 you're championing? Well, I don't know if I coined it. 57 00:03:30,360 --> 00:03:34,180 Maybe I did. I may have, but I sort of see that 58 00:03:34,180 --> 00:03:37,580 we went through traditional ops into DevOps, into NetDevOps, 59 00:03:38,820 --> 00:03:42,500 into AIOps, and I would say there's a bit of a difference. AIOps is 60 00:03:42,500 --> 00:03:46,260 really applying machine learning, supervised and unsupervised 61 00:03:46,260 --> 00:03:49,740 and semi-supervised learning over big datasets using 62 00:03:49,740 --> 00:03:53,380 generative AI. But on this idea of vibe coding, 63 00:03:54,020 --> 00:03:57,300 where anyone can do it and you apply your own personal 64 00:03:57,700 --> 00:04:01,360 experience and knowledge and your 65 00:04:01,360 --> 00:04:04,840 level of expertise to use natural language to drive 66 00:04:05,400 --> 00:04:08,760 code. Why does the network have to be behind? Why does 67 00:04:08,760 --> 00:04:12,440 infrastructure have to be behind? Unlike network automation, which took 68 00:04:12,520 --> 00:04:16,200 10 or 15 years to catch up with what developers were doing, 69 00:04:17,000 --> 00:04:20,600 can't we build our own agents? Can't I just use 70 00:04:20,600 --> 00:04:24,400 natural language to say, how's the health of the Wi-Fi in 71 00:04:24,400 --> 00:04:28,170 Vancouver? How is the traffic flowing today between 72 00:04:28,170 --> 00:04:31,690 New York and Singapore? Right? Using natural 73 00:04:31,690 --> 00:04:35,370 language to interface with this very sophisticated 74 00:04:35,370 --> 00:04:39,210 technology known as the network, right? So I like to think that, you 75 00:04:39,210 --> 00:04:42,730 know, it is sort of a mix of vibe and operations, 76 00:04:43,450 --> 00:04:47,170 and it's using things like model context protocol, plugging 77 00:04:47,170 --> 00:04:50,330 in the right MCPs to the right tools, and now 78 00:04:50,330 --> 00:04:53,760 agentic skills. And developing the skills that your 79 00:04:53,760 --> 00:04:57,400 agent has and using natural language to achieve 80 00:04:57,400 --> 00:05:01,200 these goals. And it's wonderful. I think it's more accessible 81 00:05:01,200 --> 00:05:04,760 than network automation and having to learn Python or 82 00:05:04,760 --> 00:05:08,479 Ansible or Terraform to just be able to describe it in 83 00:05:08,479 --> 00:05:11,960 your own natural language. So we can all become vibrators. 84 00:05:13,000 --> 00:05:16,840 Excellent. I mean, it is kind of incredible to 85 00:05:16,840 --> 00:05:20,560 me that like all the dashboarding and capabilities that we 86 00:05:20,560 --> 00:05:24,110 have, the network is still largely relegated to the command line. 87 00:05:24,260 --> 00:05:27,980 Interface, which makes it obscure for a lot of people. And I think it just 88 00:05:27,980 --> 00:05:31,660 doesn't get the same level of attention and care that 89 00:05:31,660 --> 00:05:35,140 it otherwise could, because I mean, quite frankly, like the network 90 00:05:35,140 --> 00:05:38,860 layer is the most critical part of the whole interface, because if you can't 91 00:05:38,860 --> 00:05:42,420 carry the data, it doesn't matter sort of what's happening on top of it. Right. 92 00:05:43,060 --> 00:05:46,420 I think that's sort of a victim of its own success in a way that 93 00:05:46,740 --> 00:05:50,330 networks typically prioritize stability. Over 94 00:05:50,410 --> 00:05:54,050 innovation, let's say, right? So even if it could 95 00:05:54,050 --> 00:05:57,890 save you 3 days on a weekend, some people are still 96 00:05:57,890 --> 00:06:01,650 more comfortable just doing it the long way, doing it at the CLI, doing 97 00:06:01,650 --> 00:06:05,250 it the way it's always been done. I think that AI is a real 98 00:06:05,250 --> 00:06:08,850 opportunity to reevaluate what is 99 00:06:08,850 --> 00:06:12,570 important in network engineering, the design, the security, 100 00:06:12,730 --> 00:06:16,480 the guardrails, the uptime. And I think 101 00:06:16,480 --> 00:06:20,200 that AI and these agents that we can build, if we blend it with our 102 00:06:20,200 --> 00:06:23,720 own 20, 25 years of experience, can 103 00:06:23,880 --> 00:06:27,720 really help, particularly in infrastructure, 104 00:06:27,960 --> 00:06:30,920 because it is so obscure, because it is such a small 105 00:06:31,640 --> 00:06:35,480 talent pool of CCIEs, of CCNPs, 106 00:06:35,480 --> 00:06:39,160 of CCIEs or INAs, you know, and 107 00:06:39,160 --> 00:06:42,580 the network hasn't been the most attractive field 108 00:06:43,060 --> 00:06:46,820 for the next generation. They wanted to get into security, they 109 00:06:46,820 --> 00:06:50,420 wanted to get into cloud, they wanted to do something sexy and 110 00:06:50,420 --> 00:06:53,780 fun. So the network is still around, but 111 00:06:54,100 --> 00:06:57,620 there's a lot of atrophy, there's a lot of huge gaps in skills. 112 00:06:58,340 --> 00:07:02,100 And I think it really is a ripe field where 113 00:07:02,100 --> 00:07:05,860 artificial intelligence can play a meaningful role, let's say. 114 00:07:06,180 --> 00:07:09,380 Yeah, I think it is sort of a wholesale reboot in a lot of ways 115 00:07:09,380 --> 00:07:13,100 because you know, in the companies that I ran, I was 116 00:07:13,100 --> 00:07:16,860 somewhat frustrated by the fact that a lot of graduates from technical 117 00:07:16,860 --> 00:07:20,260 colleges were, I felt, a little too specifically 118 00:07:20,260 --> 00:07:23,780 focused on network skills when a lot of the things that we did in a 119 00:07:23,780 --> 00:07:27,460 lot of those environments weren't necessarily network focused. 120 00:07:27,620 --> 00:07:31,300 But the more I think about it, it's like, to your point, there was a 121 00:07:31,300 --> 00:07:35,100 lot of opportunity in better managing the infrastructure. It was 122 00:07:35,100 --> 00:07:38,120 just something we tended to not focus on. And utilized 123 00:07:38,920 --> 00:07:42,600 maybe some tools or certain technologies that made those skill 124 00:07:42,600 --> 00:07:45,640 sets less of a requirement. But 125 00:07:46,280 --> 00:07:49,720 to the earlier point, like, I think there's a lot of 126 00:07:49,800 --> 00:07:53,120 capability that gets missed of you don't 127 00:07:53,120 --> 00:07:56,480 necessarily need a CCIE to diagnose, you know, an 128 00:07:56,480 --> 00:07:59,160 SMB network environment with a single 129 00:08:00,320 --> 00:08:04,040 layer, right? Like, it's just, it's not that complicated and you're kind of throwing 130 00:08:04,040 --> 00:08:07,440 like the surgeon at, you know, a plumbing problem. I 131 00:08:07,440 --> 00:08:10,750 recognize now that there was so much that happens in the 132 00:08:10,750 --> 00:08:14,310 infrastructure, even if it's an uncomplicated environment that 133 00:08:14,310 --> 00:08:18,150 could have benefited from someone generally understanding the network layer 134 00:08:18,150 --> 00:08:21,830 better. You know, we have all this data and if the 135 00:08:21,830 --> 00:08:25,310 AI is interpreting what's happening at, at that lower 136 00:08:25,310 --> 00:08:28,790 layer, then you can get a lot more granular, a lot more 137 00:08:29,030 --> 00:08:32,550 insightful information on how to manage a problem that's 138 00:08:32,550 --> 00:08:35,910 happening. Because in IT, we see this problem all the time is like, 139 00:08:36,220 --> 00:08:39,900 something's wrong. We're not exactly sure what it is. And it's probably 140 00:08:39,900 --> 00:08:43,660 something of the network layer, but none of us really understand the intricacies 141 00:08:43,660 --> 00:08:47,500 of how the packet layer works. And, you know, you made this joke in 142 00:08:47,500 --> 00:08:50,780 one of your presentations of like, well, okay, I guess this is one, 143 00:08:51,100 --> 00:08:53,980 this one time a year, I'm going to have to break out Wireshark and figure 144 00:08:53,980 --> 00:08:57,820 out all over, all over again, how this thing works. Right. So 145 00:08:57,820 --> 00:09:01,660 I, it's funny about the packet analysis because it would 146 00:09:01,660 --> 00:09:04,700 literally, I know it sounds funny, but I was lying in bed and I sort 147 00:09:04,700 --> 00:09:08,490 of sat up and went, hang on. I can put JSON into 148 00:09:08,890 --> 00:09:12,690 a vector store and do retrieval augmented generation, something I had 149 00:09:12,690 --> 00:09:16,530 already done. And then the other side of my brain was like, you 150 00:09:16,530 --> 00:09:20,370 can use Tshark at the command line to turn a packet capture into 151 00:09:20,370 --> 00:09:24,090 JSON. And these two ideas sort of married and mixed like 152 00:09:24,490 --> 00:09:28,290 paint in my brain. And so that was one of the earlier things 153 00:09:28,290 --> 00:09:31,370 that I tried was to see if I could upload a packet capture. 154 00:09:32,100 --> 00:09:35,780 And just talk to it in natural language using artificial intelligence. 155 00:09:35,940 --> 00:09:39,780 And it turned out to be wildly successful. There is online packet 156 00:09:39,780 --> 00:09:43,460 capture exams where they give you the pcap and they give you the 10 157 00:09:43,460 --> 00:09:47,220 questions and then the answers as well. And the AI was able to get 158 00:09:47,220 --> 00:09:50,620 like 20 out of 20 on these packet capture 159 00:09:50,620 --> 00:09:54,420 exams, which I thought was pretty neat. And it's, it's progressed to a 160 00:09:54,420 --> 00:09:57,460 point where now with major hyperscaler models, 161 00:09:58,220 --> 00:10:01,900 You can just upload your PCAP to ChatGPT and start chatting with it. 162 00:10:01,900 --> 00:10:05,540 That's a capability it has now. It didn't have that 3 years 163 00:10:05,540 --> 00:10:09,380 ago. So things move very, very rapidly. And I think 164 00:10:09,380 --> 00:10:12,940 we should take advantage of these tools. I love Wireshark. 165 00:10:12,940 --> 00:10:15,500 I've been a speaker at Sharkfest 166 00:10:16,620 --> 00:10:20,380 in Europe and in North America. But like you said, 167 00:10:20,380 --> 00:10:23,660 right, the 3 times, 4 times a year you actually have to break out that 168 00:10:23,660 --> 00:10:27,300 tool. To prove it's not the network more than anything. 169 00:10:27,940 --> 00:10:31,580 I've rarely used it to actually prove there was a problem with the network. I 170 00:10:31,580 --> 00:10:35,340 use it to disprove the network as the source of a problem a 171 00:10:35,340 --> 00:10:38,420 lot. So wouldn't it be neat to just upload the capture you get and say, 172 00:10:38,740 --> 00:10:42,340 am I the root cause of this problem? Just in natural language. 173 00:10:42,420 --> 00:10:46,020 Yeah, that one feels familiar to me. I was an enterprise Citrix admin 174 00:10:46,260 --> 00:10:49,660 and, you know, everyone loves to blame the presentation layer is what I 175 00:10:49,940 --> 00:10:53,390 used to say is like they're, The Citrix server is busted. It's like, no, 176 00:10:53,470 --> 00:10:57,230 no, it works fine. Like it's usually a network or, 177 00:10:57,470 --> 00:11:01,030 you know, sometimes an infrastructure issue. Quite often it's the 178 00:11:01,030 --> 00:11:04,870 application. There's nothing wrong with the server. Right. So to that point, like you get, 179 00:11:04,870 --> 00:11:07,910 you got to be able to split these things apart and kind of tease through 180 00:11:07,910 --> 00:11:11,430 the OSI layer of like, what are, what is the actual source of the issue 181 00:11:11,430 --> 00:11:15,110 here? Right. The worst one is when you curl the IP and you 182 00:11:15,110 --> 00:11:18,790 get a response, but the URL doesn't respond by name and it's a 183 00:11:18,790 --> 00:11:22,140 DNS problem. Never DNS hidden. Don't we know that already? 184 00:11:22,300 --> 00:11:26,020 That's a long day ahead, which authoritative server is getting this 185 00:11:26,020 --> 00:11:29,860 wrong, right? That's never fun. Yeah. So that, what 186 00:11:29,860 --> 00:11:33,540 you mentioned about doing the packet capture in PCAP and throwing it into JSON, 187 00:11:33,540 --> 00:11:37,140 like you have this project, one of your many, many GitHub 188 00:11:37,140 --> 00:11:40,740 repos that people, I guess like people can just, can they just go 189 00:11:40,740 --> 00:11:44,540 to the primary GitHub and then all of your projects are listed here? Is that? 190 00:11:44,620 --> 00:11:46,940 That's right. So it's github.com/automateyournetwork 191 00:11:48,460 --> 00:11:52,280 and all the projects, all the repositories are there for you to clone. And try 192 00:11:52,280 --> 00:11:56,120 yourself. The more popular one this week is, is one 193 00:11:56,120 --> 00:11:59,720 built on OpenClaw. So I saw the craze around this 194 00:11:59,720 --> 00:12:03,560 OpenClaw and I thought, well, I better try to build one. And I connected 195 00:12:03,560 --> 00:12:07,080 it to network automation tools and different network tools. And 196 00:12:07,080 --> 00:12:10,800 it's pretty neat. I have it in the VibeOps forum with about 600 197 00:12:10,800 --> 00:12:14,640 network engineers poking at it. And the first thing everyone tried to do 198 00:12:14,640 --> 00:12:18,360 was either break it, uh, get it to reveal secrets 199 00:12:18,360 --> 00:12:22,000 or get it to destroy the network. It was such a weird 200 00:12:22,000 --> 00:12:25,680 social experiment to see, okay, this AI agent, artificial 201 00:12:25,680 --> 00:12:29,440 intelligence agent is here. It's in the room. It's in the Slack channel. Go ahead 202 00:12:29,440 --> 00:12:33,240 and ask it things. And immediately 203 00:12:33,240 --> 00:12:36,440 everyone was just trying to break it, but, but it, but it held up its 204 00:12:36,440 --> 00:12:40,253 guardrails. It's, it gave a report at the end of the day, sort 205 00:12:40,253 --> 00:12:43,840 of, we asked it for a report and it held up to over 30 206 00:12:43,920 --> 00:12:47,690 social engineering attacks. Over the course 207 00:12:47,690 --> 00:12:50,570 of the 8 hours that it was alive in the Slack channel. 208 00:12:51,210 --> 00:12:54,730 So, yeah, get involved with the community, join the VibeOps forum, 209 00:12:54,730 --> 00:12:58,450 clone and star and fork the repo. It really is neat to 210 00:12:58,450 --> 00:13:02,250 have a personal assistant, but that's been trained and given 211 00:13:02,250 --> 00:13:05,850 skills around infrastructure management. Yeah. 212 00:13:06,250 --> 00:13:09,570 So I guess like that, we'll come back to this with the, the, the PacketBuddy 213 00:13:09,850 --> 00:13:13,090 piece. Like, are you saying like some of these are maybe more 214 00:13:13,090 --> 00:13:16,770 deprecated because like if you just talk to an LLM directly, like it 215 00:13:16,770 --> 00:13:20,510 already has that skillset, you don't necessarily need those projects anymore. Is that the 216 00:13:20,510 --> 00:13:23,950 case? I'm sort of glad I didn't launch a company around using 217 00:13:23,950 --> 00:13:27,510 AI to talk to packet captures because now you can just upload it to any 218 00:13:27,510 --> 00:13:31,270 major model and they seem to know how to do it. I see. 219 00:13:31,430 --> 00:13:35,149 Now, it's— I'm not trying to take credit for 220 00:13:35,149 --> 00:13:38,790 this, but some people have asked, is it because while they were 221 00:13:38,790 --> 00:13:42,470 learning and training the next iteration of the model, did they pick 222 00:13:42,470 --> 00:13:45,310 up your work from GitHub since it's all open 223 00:13:45,310 --> 00:13:48,690 source?, right? Did they learn how to do it from 224 00:13:48,930 --> 00:13:52,730 your initial attempts? I'd love to know the answer to that. I 225 00:13:52,730 --> 00:13:56,490 don't know. But yes, everything changes every day. 226 00:13:56,490 --> 00:14:00,090 You know, these hyperscalers put out a tweet and suddenly, you know, your 227 00:14:00,090 --> 00:14:03,650 startup idea is just now an AI can do it, right? 228 00:14:03,650 --> 00:14:06,770 I reuse that code inside of that 229 00:14:07,090 --> 00:14:10,610 NetClaw agent. And in Slack, now I can 230 00:14:10,610 --> 00:14:14,370 upload PCAPs to my agent and my agent can talk to 231 00:14:15,340 --> 00:14:19,100 them. So that's still pretty remarkable to think that you just need to 232 00:14:19,100 --> 00:14:22,780 get your PCAP and then right from Slack, talk to the 233 00:14:22,780 --> 00:14:26,580 AI agent and send the PCAP for analysis. So is this kind of 234 00:14:26,580 --> 00:14:30,060 how you view like VibeOps shaping out in the 235 00:14:30,140 --> 00:14:33,900 future? Is like, it is just sort of like a chat agentic future, 236 00:14:34,060 --> 00:14:37,380 like that there's maybe in a co-work model, like you either 237 00:14:37,380 --> 00:14:40,980 have specialized agents or even like a generalized chat where 238 00:14:40,980 --> 00:14:44,620 you're literally just talking to the chat about what's happening in the 239 00:14:44,700 --> 00:14:48,550 environment and it kind of like here's all the analysis I've captured in the last 240 00:14:48,550 --> 00:14:51,830 5 minutes of the infrastructure. Here are 241 00:14:52,070 --> 00:14:55,670 some ideas on potentially what's going wrong and why you're seeing these 242 00:14:55,990 --> 00:14:58,950 alerts. Like, is that sort of like what you envision here? I think it's going 243 00:14:58,950 --> 00:15:02,790 to be very much a human resources issue as much as it is a 244 00:15:03,030 --> 00:15:06,670 technology issue. Meaning I wouldn't just turn a bot up on 245 00:15:06,670 --> 00:15:10,390 day one, an agent, and plug it into production. It has to 246 00:15:10,390 --> 00:15:13,920 be trained on my corporate policies. It has to be trained on 247 00:15:13,920 --> 00:15:17,760 the guardrails, the change management requests, certain 248 00:15:18,320 --> 00:15:21,680 network-specific guardrails, all the things that you would train a 249 00:15:22,240 --> 00:15:25,880 junior, right? It's just like hiring a junior human. The first thing you're going to 250 00:15:25,880 --> 00:15:29,680 do isn't going to throw them into the fire and have them handle some BGP 251 00:15:30,240 --> 00:15:33,920 change. You're likely going to start with read-only activities, 252 00:15:34,000 --> 00:15:37,760 testing, documentation, maybe minor tickets that they can 253 00:15:38,170 --> 00:15:41,690 handle. I think it's going to reflect that human experience where you have 254 00:15:41,690 --> 00:15:45,330 this agent start with human in the loop, human on the loop, human in 255 00:15:45,330 --> 00:15:48,970 the lead, fully autonomous. And they're just like 256 00:15:49,610 --> 00:15:52,730 digital, you know, I've heard virtual employees, I've heard digital 257 00:15:53,450 --> 00:15:57,250 coworkers. And they're given a level of autonomy. They can reason and 258 00:15:57,250 --> 00:16:01,090 now they can call tools, particularly in the form of skills 259 00:16:01,090 --> 00:16:04,890 or MCPs. So it's changing very, 260 00:16:04,890 --> 00:16:08,720 very fast here. And I read some Gartner report that 261 00:16:08,720 --> 00:16:12,440 they believe 70% of IT operations 262 00:16:12,440 --> 00:16:16,160 by 2030 will be augmented by AI in 263 00:16:16,160 --> 00:16:19,720 some way, shape, or form. Maybe not totally displaced or 264 00:16:19,720 --> 00:16:23,360 totally autonomous, but AI will be 265 00:16:23,360 --> 00:16:26,520 participating in 70% of the work in my 266 00:16:27,320 --> 00:16:30,760 field before the end of the decade, right? Yeah. I 267 00:16:31,000 --> 00:16:34,260 like the quote or description from, 268 00:16:34,260 --> 00:16:38,100 from NVIDIA CEO who said that the IT department will become the 269 00:16:38,100 --> 00:16:41,820 HR department for AI. I'm glad that you said that. I was going to 270 00:16:41,820 --> 00:16:45,060 bring that up, but I wasn't sure if people are sick of me saying that 271 00:16:45,140 --> 00:16:48,620 because I really paid attention to that. I know people made a meme out of 272 00:16:48,620 --> 00:16:52,380 it, but through my experience now, having built some 273 00:16:52,380 --> 00:16:55,780 of these agents, it really does feel less like a 274 00:16:55,940 --> 00:16:59,390 technology problem and more like an HR department problem. As 275 00:17:00,270 --> 00:17:03,990 to where I would use these agents in my org, whose work 276 00:17:03,990 --> 00:17:06,430 they can augment, how I expose 277 00:17:07,230 --> 00:17:10,470 them, you know, securely, AAA and orchestration 278 00:17:10,470 --> 00:17:14,230 and augmentation and orchestrating the whole thing together. 279 00:17:14,230 --> 00:17:17,871 Do we need supervisor agents? You know, how high up 280 00:17:18,190 --> 00:17:22,030 this org chart do we build the hierarchy? It's an 281 00:17:22,030 --> 00:17:25,710 architectural HR type issue, right? Yeah, that's really fascinating. I think that's 282 00:17:25,710 --> 00:17:29,430 the only way that this works. I mean, The other sort of HR-ish 283 00:17:29,430 --> 00:17:32,150 type thing that I'm curious about here to get your read on, because I've sort 284 00:17:32,150 --> 00:17:35,270 of debated this in my head and chatted with a few people. I'd love to 285 00:17:35,270 --> 00:17:38,990 get your, your thoughts on this as well is like, what does entry-level 286 00:17:38,990 --> 00:17:42,710 job in IT look like in the future? Right? Because it used to be that, 287 00:17:42,710 --> 00:17:46,430 you know, you would have a junior role, like a level 1 tech 288 00:17:46,430 --> 00:17:49,910 who, you know, does a lot of the triage and some of the basic work 289 00:17:49,910 --> 00:17:53,630 that's routine and has SOPs built around it. And you kind of get some 290 00:17:53,630 --> 00:17:57,090 level of exposure to what's going on here. And it's almost like we're getting to 291 00:17:57,090 --> 00:18:00,250 a space where we're almost eliminating that role and the necessity for 292 00:18:00,330 --> 00:18:04,090 it entirely, because if it's repeatable and it has documentation, then, 293 00:18:04,090 --> 00:18:07,690 you know, just stick an AI on it. But then what does that mean? People 294 00:18:07,690 --> 00:18:11,130 just start at level 2 and they just chat with the, 295 00:18:11,130 --> 00:18:14,810 the AI agents that are monitoring the environment. 296 00:18:14,810 --> 00:18:18,370 And if that's the case, then like, how do you get into that job? And 297 00:18:18,370 --> 00:18:21,130 does it require kind of any technical know-how at 298 00:18:22,790 --> 00:18:25,830 all? I've given this a lot of thought and it does come up and it's 299 00:18:25,830 --> 00:18:29,110 sort of, now we're sort of into the almost an 300 00:18:29,270 --> 00:18:31,270 ethical philosophical discussion 301 00:18:33,030 --> 00:18:36,670 about, um, you know, what humans' role should 302 00:18:36,670 --> 00:18:40,510 be and is. And I wish I had a good answer and I'm not trying 303 00:18:40,510 --> 00:18:43,110 to skirt the question, but I think 304 00:18:44,550 --> 00:18:47,350 that in my utopianist outcome, in my 305 00:18:48,300 --> 00:18:52,100 mind, those juniors can be accelerated through their career faster. 306 00:18:52,100 --> 00:18:55,500 They don't necessarily have to spend 25 years like I 307 00:18:55,980 --> 00:18:59,740 did to become a so-called expert in some of these things. 308 00:18:59,740 --> 00:19:03,380 I think that their access to AI 309 00:19:03,380 --> 00:19:06,300 and to senior engineers writing these 310 00:19:07,420 --> 00:19:11,140 agents, it's a way to transfer the knowledge and to pass the torch. 311 00:19:11,140 --> 00:19:14,340 I don't think we should be pulling the ladder up on the juniors. 312 00:19:15,100 --> 00:19:18,740 Right. We still are going to need humans that understand 313 00:19:18,740 --> 00:19:22,460 how these agents are written, how they're built, the tools they're built 314 00:19:22,460 --> 00:19:26,140 on, the principles that guide them. There's still very much a place 315 00:19:26,140 --> 00:19:29,980 for human beings here, but I think that it should 316 00:19:29,980 --> 00:19:33,020 democratize things in a way similar to Lotus 317 00:19:33,660 --> 00:19:37,380 1-2-3 or Excel did for accounting. Right now, 318 00:19:37,380 --> 00:19:40,460 are there more or less accountants after the 319 00:19:41,810 --> 00:19:45,610 spreadsheet? You know, became widely available. I would suggest more, 320 00:19:45,610 --> 00:19:48,770 right? I would suggest that they, that, that there's 321 00:19:48,770 --> 00:19:52,570 more spreadsheets, more data, more accountants. 322 00:19:52,570 --> 00:19:55,890 It's just that their job is a little bit different. The job has 323 00:19:56,530 --> 00:19:59,890 changed and they're working at more elevated levels in the 324 00:20:00,290 --> 00:20:03,650 company. They're working on strategy. They're working on investment. They're working 325 00:20:03,650 --> 00:20:07,090 on taxes, right? Accounting changed as a 326 00:20:07,800 --> 00:20:11,480 result. But it didn't eliminate accounting. I'm trying to build an agent that you 327 00:20:11,480 --> 00:20:15,040 can augment your tool, your, your team with 328 00:20:15,040 --> 00:20:17,720 to handle a lot of the 329 00:20:18,440 --> 00:20:21,920 mundane, repeatable, safe, you know, 330 00:20:22,480 --> 00:20:26,200 low-hanging fruit through just natural language, right? 331 00:20:26,200 --> 00:20:29,920 There's always going to be work. There's always going to be work in IT and 332 00:20:29,920 --> 00:20:33,520 in networks in particular, just because we have these agents 333 00:20:33,520 --> 00:20:36,870 doesn't suddenly mean we don't need the humans. And through my own 334 00:20:37,750 --> 00:20:41,110 experience, I went through the same 335 00:20:41,590 --> 00:20:45,350 dismay with network automation. You're going to automate yourself out of a job. 336 00:20:45,350 --> 00:20:48,310 You're going to automate, automation is going to wipe out network 337 00:20:48,950 --> 00:20:52,550 engineers. What I found was when I automated one thing, 338 00:20:52,550 --> 00:20:56,230 they didn't just say, well, thanks, great job. You know, see you later. We don't 339 00:20:56,230 --> 00:20:59,790 need you anymore. You automated that one thing. They asked what we 340 00:20:59,790 --> 00:21:03,400 could automate next. So I became more valuable to 341 00:21:03,400 --> 00:21:07,160 the company after I started introducing automation, not 342 00:21:07,160 --> 00:21:11,000 less valuable. Right? So I think it's similar with 343 00:21:11,000 --> 00:21:14,760 artificial intelligence. If it's John's agent, John's not 344 00:21:14,760 --> 00:21:18,440 going anywhere. Right? Right? But if 345 00:21:18,440 --> 00:21:21,920 you're just consuming John's agent, if you don't have an agent of your 346 00:21:22,400 --> 00:21:26,160 own, then you might want to, you know, start to contribute in a 347 00:21:26,160 --> 00:21:29,640 different way, right? So this is interesting too, because like 348 00:21:29,640 --> 00:21:31,720 I'm curious to get your thoughts on the 349 00:21:32,760 --> 00:21:36,280 practical implementation of some of these strategies for, you know, some of the 350 00:21:36,280 --> 00:21:39,400 people listening, typically managing, you know, 30 351 00:21:39,560 --> 00:21:42,560 to 130 clients at a time in a 352 00:21:43,159 --> 00:21:46,480 multi-tenant environment. What is your thought on sort of first steps 353 00:21:46,480 --> 00:21:50,200 in exploring the utilization of this technology? Say like they've got an 354 00:21:50,520 --> 00:21:54,080 RMM, like maybe they've got, you know, PRTG or some basic kind 355 00:21:54,080 --> 00:21:57,841 of monitoring tools in place. But the, the, the level of the 356 00:21:57,841 --> 00:22:01,040 sort of view of visibility on the network in particular 357 00:22:01,040 --> 00:22:04,840 and infrastructure as a whole is fairly rudimentary. Like, where would be 358 00:22:04,840 --> 00:22:08,600 the first place to explore? I can't imagine it's, you know, OpenClaw is 359 00:22:08,600 --> 00:22:12,321 necessarily the first place to go. Could be a little dangerous, but any, any other 360 00:22:12,321 --> 00:22:16,080 suggestions on the entry point here? I think MSPs are in a 361 00:22:16,320 --> 00:22:20,080 very supremely unique position to take advantage of 362 00:22:20,400 --> 00:22:23,600 this in that they can start to build digital representatives that 363 00:22:24,440 --> 00:22:28,240 represent the portfolio under that client. Right now, that 364 00:22:28,240 --> 00:22:32,000 agent is going to be specific and bespoke per customer. But 365 00:22:32,000 --> 00:22:35,080 if you can start to connect it with things like retrieval augmented 366 00:22:35,960 --> 00:22:39,120 generation, accessing your knowledge base, your PDFs, your, 367 00:22:39,120 --> 00:22:42,360 your spreadsheets, your Salesforce, 368 00:22:43,320 --> 00:22:47,000 your Jira, your Confluence, your Atlassian, right? Start 369 00:22:47,320 --> 00:22:50,440 to list all the tools that you use or that a human would 370 00:22:51,130 --> 00:22:54,930 use to best manage that tenant. And think of it that 371 00:22:54,930 --> 00:22:58,610 you could have a chat interface into that and ask it 372 00:22:58,610 --> 00:23:02,130 where the pain points are, where the latest tickets are, what 373 00:23:02,130 --> 00:23:05,530 the status of the project is. Think of it as 374 00:23:05,530 --> 00:23:09,370 a practical employee that you're going to build up 375 00:23:09,370 --> 00:23:13,090 and attach tools to it. And those tools are going to reach out and get 376 00:23:13,090 --> 00:23:16,250 the external data that the model needs. 377 00:23:17,550 --> 00:23:20,990 Now, I mean, be very careful. Don't do shadow AI here. This, this 378 00:23:20,990 --> 00:23:24,350 does require a level of enterprise agreement and a 379 00:23:24,350 --> 00:23:28,030 private LLM and an API key and approval from 380 00:23:28,030 --> 00:23:31,310 your departments. Don't just do this on a 381 00:23:31,390 --> 00:23:34,910 YOLO DIY sort of thing. There's a lot 382 00:23:34,910 --> 00:23:38,630 involved here. However, you can start to build agents 383 00:23:38,630 --> 00:23:41,970 with agent development toolkits, ADKs. 384 00:23:41,970 --> 00:23:45,680 Open Claw, correct. It's a personal assistant sort of thing 385 00:23:45,680 --> 00:23:49,160 right now. It's probably not commercially ready, but Claude 386 00:23:49,160 --> 00:23:52,704 Code, Claude Desktop, Cursor, 387 00:23:52,896 --> 00:23:56,720 Antigravity, ChatGPT, Gemini, the list goes on and 388 00:23:57,040 --> 00:24:00,560 on. I think just starting somewhere and maybe starting with 389 00:24:00,560 --> 00:24:04,160 a roadmap and a plan of, you know, by the end of 390 00:24:04,400 --> 00:24:08,160 March, let's have one agent, right? And then by the 391 00:24:08,160 --> 00:24:11,640 end of April, let's see if we can scale that to 5 or 15 agents, 392 00:24:11,640 --> 00:24:15,240 right? And start to see how you can augment your workforce. 393 00:24:15,240 --> 00:24:19,040 And augment yourself. How could you build a little personal assistant and 394 00:24:19,040 --> 00:24:22,760 what tools should it have access to, to make your life 395 00:24:22,760 --> 00:24:26,520 easier and better and more productive and more fulfilled? One 396 00:24:26,520 --> 00:24:30,160 of the, the sort of the central pieces to this is like creating boundaries around 397 00:24:30,160 --> 00:24:32,960 it, like what it can and can't do. The HR policy for the, uh, for 398 00:24:32,960 --> 00:24:36,760 the AI, I think is, is really, really important here. The privacy 399 00:24:36,760 --> 00:24:40,280 implications of this, I think are also really massive, right? Because you're dealing with other 400 00:24:40,280 --> 00:24:43,910 people's data, sensitive information in some cases. You mentioned kind 401 00:24:43,910 --> 00:24:46,750 of using a private LLM, not just sort of, 402 00:24:47,550 --> 00:24:50,910 hey, I, I, I got a free key on GPT. Like that's never the way 403 00:24:50,910 --> 00:24:54,670 to go, but you know, what about using the 404 00:24:54,750 --> 00:24:58,550 public models versus private models? Any thoughts on sort of, you know, 405 00:24:58,550 --> 00:25:02,310 roll your own, keep it in Ollama, run it local versus utilizing 406 00:25:02,310 --> 00:25:05,830 like say a secure container in Azure or an API 407 00:25:05,830 --> 00:25:09,360 key from one of the, one of the major brands. That, that is 408 00:25:09,360 --> 00:25:13,040 an avenue. And I really suggest people look into that avenue, 409 00:25:13,040 --> 00:25:16,840 particularly with, say, personal things. So 410 00:25:16,840 --> 00:25:20,640 yes, Ollama, LM Studio, Microsoft Foundry Local, there 411 00:25:20,640 --> 00:25:24,360 are very, very capable public models. By the end of this 412 00:25:24,680 --> 00:25:28,280 call, you could install any of those three and literally have a 413 00:25:28,280 --> 00:25:31,560 model locally to chat with. So they all have 414 00:25:31,560 --> 00:25:35,240 REST APIs, so you can do programming against them. You can write 415 00:25:35,240 --> 00:25:38,920 these agents against them. Most of the latest models that 416 00:25:38,920 --> 00:25:42,560 are open source can do tool calling, which lets you do 417 00:25:42,560 --> 00:25:46,360 these reasoning and action agents. But there's, you know, there may be 418 00:25:46,360 --> 00:25:50,160 hardware limitations there, right? We now we're starting to talk about the 419 00:25:50,160 --> 00:25:53,560 size of the model and the number of parameters and the GPU 420 00:25:53,560 --> 00:25:57,280 or CPU that you have locally. That is another very 421 00:25:57,360 --> 00:26:01,040 safe offline avenue is looking at 422 00:26:01,120 --> 00:26:04,830 open source models. There is still a quite, I would 423 00:26:04,830 --> 00:26:07,590 say quite a big disparity between the quality 424 00:26:08,310 --> 00:26:11,950 though of a private model, like an Anthropic Claude 425 00:26:11,950 --> 00:26:15,550 4.6 through a private key and an open 426 00:26:15,550 --> 00:26:19,310 source model, but it's a horse race, right? Things get better, things improve. 427 00:26:19,310 --> 00:26:22,630 Who knows what the next model around the corner is going to be capable 428 00:26:22,870 --> 00:26:25,870 of. Yeah, I guess so, like from my perspective, like, especially in this use case 429 00:26:25,870 --> 00:26:29,520 of what I'm thinking, you know, obviously if you're doing, you 430 00:26:29,520 --> 00:26:33,160 know, vibe coding, then, you know, obviously Cloud Code 46 is 431 00:26:33,440 --> 00:26:37,040 the go-to, you know, but if you're just doing kind of packet 432 00:26:37,040 --> 00:26:40,400 interpretation and network statistics and data collection, like I feel like 433 00:26:41,120 --> 00:26:44,680 a local, a local model with a fairly decent, like 434 00:26:44,680 --> 00:26:47,880 off-shelf GPU doesn't need to be, you know, a 3090 or something like that. It 435 00:26:47,880 --> 00:26:51,440 could be something pretty decent, but not, you know, blow their socks off. 436 00:26:51,600 --> 00:26:54,800 It'll probably churn through that, albeit maybe slightly 437 00:26:55,240 --> 00:26:58,880 slower than, a more private model from the cloud, but you know, 438 00:26:58,880 --> 00:27:02,720 then it's local. It's, you know, maybe 10, 30% slower, 439 00:27:02,720 --> 00:27:06,440 but it still does the job. Free. And there's a huge upside to things 440 00:27:06,440 --> 00:27:10,160 being free, especially when you're in an exploratory phase. If all this is 441 00:27:10,160 --> 00:27:13,120 new to you and you just, you know, you don't want to put another thing 442 00:27:13,120 --> 00:27:16,840 on your credit card, but you don't want to be held back from starting. There's 443 00:27:16,840 --> 00:27:20,560 a lot to be said for Ollama and LM Studio and 444 00:27:20,560 --> 00:27:23,930 these local models. There really is. I completely agree 445 00:27:24,330 --> 00:27:28,050 with you. And, you know, some people really value their privacy. Some 446 00:27:28,050 --> 00:27:31,610 people really don't want to be using these datasets 447 00:27:32,250 --> 00:27:35,890 in particular with a public model, right? And it has all 448 00:27:35,890 --> 00:27:39,690 the advantages. It has a REST API. You can program against it. MCPs work 449 00:27:40,010 --> 00:27:43,850 with it. Yeah, really good suggestion. Yeah. Okay. On the 450 00:27:43,850 --> 00:27:47,650 sort of vibe, vibe coding, vibe development, what are your thoughts on, I 451 00:27:47,650 --> 00:27:50,760 guess, two things is. I mean, to lead into the question here is why do 452 00:27:50,760 --> 00:27:54,320 you think like the tools vendors have been kind of slow to get 453 00:27:54,320 --> 00:27:58,080 on this, this train? And granted, you know, like it's not been a long time 454 00:27:58,080 --> 00:28:01,680 and we, we can respect that enterprises move at a, at a different 455 00:28:01,840 --> 00:28:05,600 pace than, you know, us hobbyists do, I suppose. But I find it 456 00:28:05,600 --> 00:28:09,280 interesting that some of these, these capabilities have not already shown up in 457 00:28:09,280 --> 00:28:12,840 some of the, the industry tools already. Any thoughts 458 00:28:12,840 --> 00:28:16,560 on the speed of development and the speed of application for some of 459 00:28:17,800 --> 00:28:21,600 these capabilities? Yeah. So on the whole vibe coding, I think that we're going 460 00:28:21,600 --> 00:28:24,600 to, I think we've maybe even reached a point where it's just called 461 00:28:25,320 --> 00:28:29,120 coding now. Right? Like, I think everybody's doing it this way. So I 462 00:28:29,120 --> 00:28:32,839 think we should just call it coding. Yeah. Because just quickly on that, 463 00:28:32,839 --> 00:28:36,280 because like you see people arguing that like, oh, well, you 464 00:28:36,520 --> 00:28:40,240 know, it creates trash code and, you know, I would never use that. 465 00:28:40,240 --> 00:28:44,060 But then the other, a lot of decent developers will argue like, look, I've 466 00:28:44,060 --> 00:28:47,540 had some people on my team that were terrible coders and it's like, trust me, 467 00:28:47,540 --> 00:28:51,340 Claude is a much better developer than these people. You're not going to win everyone 468 00:28:51,420 --> 00:28:54,260 over right now. Here's the one thing that I like to remind people when they 469 00:28:54,260 --> 00:28:57,940 say to me, well, you're not reading all this code. Like you don't really 470 00:28:57,940 --> 00:29:01,460 understand what it's doing. Well, that's a level of abstraction to me. I think 471 00:29:01,460 --> 00:29:05,300 that's a positive. I think that's on the good side of vibe coding, not 472 00:29:05,300 --> 00:29:08,660 the bad side of vibe coding. How many Python packages do you go 473 00:29:08,660 --> 00:29:12,360 to the PyPI? Github.org and look up the source code 474 00:29:12,360 --> 00:29:16,200 of the Python and the libraries that are included. How many times do you 475 00:29:16,200 --> 00:29:19,960 go to the npm every time you Node install 476 00:29:19,960 --> 00:29:23,720 something? Come on, let's be honest with ourselves, right? The one caveat I would make, 477 00:29:23,720 --> 00:29:27,440 John, is like, if you're going to vibe code something and release 478 00:29:27,440 --> 00:29:30,560 it publicly and use it in some capacity, especially if you're going to sell it, 479 00:29:30,560 --> 00:29:34,280 for the love of God, get somebody who is very qualified 480 00:29:34,280 --> 00:29:37,630 to sign off on it. Especially from a security standpoint, but 481 00:29:38,110 --> 00:29:41,750 also from production ready, right? Like maybe you haven't read it, 482 00:29:41,750 --> 00:29:45,550 but maybe someone should, you know, I think that's very fair. Anything 483 00:29:45,550 --> 00:29:48,990 you're going to charge money for, maybe, you know, you should 484 00:29:48,990 --> 00:29:52,670 have some rigor around that, but why the vendors are 485 00:29:52,670 --> 00:29:56,390 behind that really bothers me. It really does because, and I, 486 00:29:56,390 --> 00:29:59,910 to be fair, MCP is, let's just call it for, you know, 487 00:29:59,910 --> 00:30:03,670 16 months old. So it's about a year and a half old now. 488 00:30:03,670 --> 00:30:07,430 But I'm just like, where are the MCPs for all these 489 00:30:07,430 --> 00:30:10,830 platforms where I can just plug them in? Now, I don't know if it's because 490 00:30:10,830 --> 00:30:14,430 it is too easy, because there's a lot of revenue at stake 491 00:30:14,510 --> 00:30:18,150 for support and professional services. And there's a lot that goes into 492 00:30:18,150 --> 00:30:21,070 being a vendor, right? And I don't know if they're 493 00:30:21,950 --> 00:30:25,590 just being overprudent and cautious to not 494 00:30:25,590 --> 00:30:28,750 give away the keys to their 495 00:30:29,940 --> 00:30:33,780 monetary success., right? If there was suddenly an MCP for 496 00:30:34,260 --> 00:30:37,700 Vendor X's tool and you don't need to buy their platform for 497 00:30:38,260 --> 00:30:41,820 that tool anymore, maybe that's what's holding them back is, is sort of, they 498 00:30:41,820 --> 00:30:45,620 don't want to cannibalize their 499 00:30:46,020 --> 00:30:49,060 own commercial platforms by really, by getting involved 500 00:30:49,860 --> 00:30:53,620 in MCP. But I don't know. I think SMTP 501 00:30:53,620 --> 00:30:56,260 and HTTP, these protocols that everyone can use 502 00:30:57,430 --> 00:31:00,990 and build upon. Got us to where we are today, I think they can 503 00:31:00,990 --> 00:31:04,630 find a way to monetize MCP and to 504 00:31:05,030 --> 00:31:08,630 still, to still be innovative and keeping up with 505 00:31:09,430 --> 00:31:13,230 the tools available, but, but not losing their shirt, right? Another use case I kind 506 00:31:13,230 --> 00:31:17,070 of wanted to explore with you is something that I find is particularly 507 00:31:17,070 --> 00:31:20,550 problematic. I've had a few conversations with some groups recently in the past couple 508 00:31:20,870 --> 00:31:24,030 of weeks about alert fatigue and the difficulty of 509 00:31:24,550 --> 00:31:28,400 getting the signal-to-noise ratio right in managing complex 510 00:31:28,400 --> 00:31:31,920 environments. Where like you can monitor everything, but, you know, 10,000 tickets on a board 511 00:31:31,920 --> 00:31:35,680 is not helpful to anybody. Or you can tame down the noise and 512 00:31:35,680 --> 00:31:39,320 potentially have something go unnoticed, and then a VP is super pissed that you guys 513 00:31:39,320 --> 00:31:42,880 missed it, right? I have to imagine there is a combination 514 00:31:42,880 --> 00:31:46,640 of your tool sets as well as an Agentic capability that could maybe help people 515 00:31:46,640 --> 00:31:50,120 kind of run up the middle here on what is actually important in 516 00:31:50,120 --> 00:31:53,840 this pile of noise, and therefore how should we tune our 517 00:31:53,840 --> 00:31:57,360 alerts and our RMM for management around that? What are your thoughts on that as 518 00:31:57,360 --> 00:32:01,180 sort of an exercise that people could take on? It's 519 00:32:01,180 --> 00:32:04,380 read-only. You're not affecting change. You're not disrupting 520 00:32:05,340 --> 00:32:08,380 flows. You're not, you know, doing the hard thing first, 521 00:32:08,460 --> 00:32:11,740 but it's valuable, extremely valuable. And what a great use 522 00:32:12,300 --> 00:32:15,980 for generative AI that predicts tokens and can do correlation 523 00:32:15,980 --> 00:32:19,620 and root cause analysis to point it 524 00:32:19,620 --> 00:32:22,700 at 10,000 tickets or 6,000 tickets or 525 00:32:23,740 --> 00:32:27,570 whatever and say, boil this down into 100 526 00:32:27,570 --> 00:32:31,250 tickets. Make this something that a human can consume, find 527 00:32:31,250 --> 00:32:34,770 the patterns, find repeatable tickets, find, you know what I mean? Like, 528 00:32:34,770 --> 00:32:38,530 what a great use case. And then from there, maybe have 529 00:32:38,530 --> 00:32:40,690 the AI build on top of that. Now that we're 530 00:32:41,490 --> 00:32:44,890 down to 100, recategorize them, break them down further, and 531 00:32:44,890 --> 00:32:48,570 give me some suggested code on how to roll out the fix, how 532 00:32:48,570 --> 00:32:52,010 to deal with this issue, right? So you start to get 533 00:32:52,010 --> 00:32:55,770 into solving the problems after you sort them and boil them 534 00:32:55,770 --> 00:32:59,570 down. Into a human consumable level, make the 535 00:32:59,570 --> 00:33:03,170 tickets, make the plans, make the order of operations, right? Come 536 00:33:03,170 --> 00:33:06,570 up with the test plans. What would success look like? What would 537 00:33:06,970 --> 00:33:10,450 failure look like? I think that's a wonderful use case. And most people want to 538 00:33:10,769 --> 00:33:14,410 start with read-only and human in the loop. And what a safe exercise 539 00:33:14,650 --> 00:33:18,370 to start with is just giving it access to your ticket 540 00:33:18,370 --> 00:33:22,070 farm and seeing what knowledge you can get out of that data, right? 541 00:33:22,070 --> 00:33:25,750 With that, I want to be cognizant of your time here. Sure, maybe 542 00:33:25,750 --> 00:33:29,550 we can connect offline and have a ton more conversations because I got through 543 00:33:29,550 --> 00:33:31,750 like 20% of what I want to ask you for your list of questions. Okay, 544 00:33:31,750 --> 00:33:34,630 okay. Well, I can come back. I'd be more than happy to come back. I 545 00:33:34,630 --> 00:33:38,150 love the tenor and tone of this discussion, so we should maybe wrap it up 546 00:33:38,150 --> 00:33:41,990 and then I'll come back in a few weeks. Okay. Like 547 00:33:41,990 --> 00:33:45,150 from that, like what are the practical steps? You know, like we talked about, you 548 00:33:45,150 --> 00:33:48,880 know, maybe roll your own, look for some internal use cases. 549 00:33:49,360 --> 00:33:52,880 Are there particular resources that you would point to in the audience 550 00:33:52,880 --> 00:33:56,600 of, you know, MSPs managing, you know, 30 to 551 00:33:56,600 --> 00:34:00,280 130 clients? What would you say is sort of the important takeaways of like next 552 00:34:00,280 --> 00:34:04,080 actions for them to look at for the raft of information and 553 00:34:04,080 --> 00:34:07,520 content that you've created both in books and open source projects? And, you know, I 554 00:34:07,520 --> 00:34:10,560 want to thank you for that. Like the fact that you're creating all 555 00:34:11,120 --> 00:34:14,880 of this information and open sourcing, I think is incredibly valuable and 556 00:34:15,629 --> 00:34:19,469 really, really admirable. Well, thank you so much. I, um, honestly, I 557 00:34:19,469 --> 00:34:22,789 think, and I don't want to lose anyone who's not a programmer, but ignore that, 558 00:34:22,789 --> 00:34:26,509 but set that aside in your mind for a second. Go ahead and 559 00:34:26,509 --> 00:34:30,269 download VS Code, Visual Studio Code, and it comes with 560 00:34:30,269 --> 00:34:34,109 Copilot inside of it. Copilot is free for X number of calls. There's 561 00:34:34,109 --> 00:34:37,909 enough there to get you started. And now you have an integrated 562 00:34:37,909 --> 00:34:41,530 development environment, an IDE, which you can edit files and do different things 563 00:34:41,530 --> 00:34:45,090 with. And a Copilot to help you write 564 00:34:45,090 --> 00:34:48,890 code, write emails, interrogate the files that are open in the 565 00:34:48,890 --> 00:34:52,130 IDE. And then once you're comfortable there, get a little bit of 566 00:34:52,450 --> 00:34:55,850 comfort with that, look into how to add an 567 00:34:55,850 --> 00:34:59,410 MCP to your Copilot. There's going to be a little snippet of code. You've got 568 00:34:59,410 --> 00:35:03,090 to plug it into your Copilot, and now your Copilot has 569 00:35:03,810 --> 00:35:07,540 access to tools. Think of the top 5 things you do every day. Are 570 00:35:07,540 --> 00:35:11,260 you in Salesforce? Are you in Atlassian? Are you in Jira? What 571 00:35:11,260 --> 00:35:14,820 is it that you need AI's help with? And see 572 00:35:14,820 --> 00:35:18,340 if they have MCPs that you can plug in. 573 00:35:18,340 --> 00:35:21,300 From there, the sky's the limit. You're going to find more and more MCPs. You're 574 00:35:21,300 --> 00:35:25,100 going to find day-to-day use. You're going to start getting 575 00:35:25,100 --> 00:35:28,820 your $20 a month value out of your AI. Uh, and 576 00:35:28,820 --> 00:35:32,540 you can also do all this with free open source models. 577 00:35:32,540 --> 00:35:36,360 So, and reach out, feel free to connect with me on social 578 00:35:36,360 --> 00:35:39,080 media. Um, or if you have any questions at all, if there's any sharp edges, 579 00:35:39,080 --> 00:35:42,800 just let me know. Okay. I'll link to your LinkedIn profile on, 580 00:35:42,800 --> 00:35:46,360 in the show notes. Anywhere else that you would direct people, your, 581 00:35:46,360 --> 00:35:50,120 maybe give your GitHub address, I'll link to that as well. Any other places 582 00:35:50,120 --> 00:35:53,920 you would direct people to? I would find my YouTube channel. I think 583 00:35:53,920 --> 00:35:57,360 it's going to be really help people who want to get on this journey. I'm 584 00:35:57,360 --> 00:36:00,760 on Twitter, X still, and LinkedIn is a really good place to find 585 00:36:01,270 --> 00:36:03,430 me as well. I want to thank you for having me on the show. This 586 00:36:03,430 --> 00:36:07,270 has been a lot of fun. Really appreciate your time, John. Thanks. All right. Take 587 00:36:07,270 --> 00:36:07,300 care. See ya.