1 00:00:00,000 --> 00:00:03,030 Chase Douglas: There's a lot of interesting stuff happening in AI 2 00:00:03,030 --> 00:00:06,960 where we're starting to realize on the coding front, like what does it mean? 3 00:00:07,680 --> 00:00:13,320 To code efficiently with AI when you get into spec driven, test driven development. 4 00:00:19,410 --> 00:00:21,360 Corey Quinn: Welcome to Screaming in the Cloud. 5 00:00:21,480 --> 00:00:26,400 I'm Cory Quinn, and I am joined by dear friend of mine and yours, chase 6 00:00:26,400 --> 00:00:29,250 Douglas, CEO at Archodex these days. 7 00:00:29,620 --> 00:00:30,970 Chase, how have you been? 8 00:00:31,390 --> 00:00:32,590 Chase Douglas: I, I've been good. 9 00:00:32,590 --> 00:00:33,580 I've been good. 10 00:00:33,700 --> 00:00:36,130 Corey Quinn: You've gone through an interesting series of evolutions. 11 00:00:36,130 --> 00:00:39,280 You were the CEO at a startup called Stackery before you did that. 12 00:00:39,280 --> 00:00:39,970 No one knows. 13 00:00:39,970 --> 00:00:42,220 No one has any idea what your past is, 14 00:00:42,220 --> 00:00:42,610 Chase Douglas: just to be clear. 15 00:00:42,610 --> 00:00:43,390 CTO. 16 00:00:43,420 --> 00:00:45,790 'cause I'd worked alongside some amazing people. 17 00:00:46,000 --> 00:00:47,260 Corey Quinn: Ah, my apologies, I'll. 18 00:00:48,180 --> 00:00:49,919 Those three letter titles are strange. 19 00:00:49,919 --> 00:00:54,419 CTO at Stackery and co-founder, let's forget that part of it. 20 00:00:54,425 --> 00:00:54,614 Chase Douglas: Exactly. 21 00:00:54,809 --> 00:00:55,230 Exactly. 22 00:00:55,230 --> 00:00:57,239 Corey Quinn: Which honestly means your job is whatever needs to be done in 23 00:00:57,239 --> 00:01:00,239 the, in that moment, you have five things that must get done today. 24 00:01:00,239 --> 00:01:02,160 Chase Douglas: On a good day, you'll get to, here's a thousand question 25 00:01:02,160 --> 00:01:07,229 questionnaire for compliance about printers, about your SaaS product. 26 00:01:07,259 --> 00:01:08,100 Corey Quinn: AI can do it. 27 00:01:08,429 --> 00:01:08,550 Yeah. 28 00:01:09,030 --> 00:01:10,695 So then you got acquired. 29 00:01:11,369 --> 00:01:15,059 By AWS, which was interesting because the way I found out that you got 30 00:01:15,059 --> 00:01:18,690 acquired was I got an email as a Stacky customer, and then I emailed 31 00:01:18,690 --> 00:01:19,740 a few of you be like, what the hell? 32 00:01:19,740 --> 00:01:21,990 Like, hey, that we've been like that for months now and we haven't 33 00:01:21,990 --> 00:01:23,070 been allowed to tell anyone. 34 00:01:23,190 --> 00:01:24,359 Thanks for noticing. 35 00:01:24,750 --> 00:01:26,460 So yeah, that was. 36 00:01:26,759 --> 00:01:31,350 Amazon Marketing, once again, taking a vow of silence for no discernible reason. 37 00:01:31,530 --> 00:01:32,789 You were there for a few years. 38 00:01:32,789 --> 00:01:34,229 You got a few things out the door. 39 00:01:34,229 --> 00:01:38,130 The second coming of Stackery, but it's a UI focused product at AWS. 40 00:01:38,130 --> 00:01:41,520 So it's of course doomed that company cannot build a good UX to save it's 41 00:01:41,520 --> 00:01:46,169 life and then decided that, alright, I'm gonna go do something else. 42 00:01:46,320 --> 00:01:48,630 What makes me unhappy unreasonably. 43 00:01:48,630 --> 00:01:50,759 So, and then you just had to go tilted a windmill. 44 00:01:50,970 --> 00:01:51,720 What windmill is it? 45 00:01:51,870 --> 00:01:54,230 Chase Douglas: I'm building a project, it's called Archodex. 46 00:01:54,510 --> 00:01:58,110 Uh, it's me and a fellow principal engineer from AWS. 47 00:01:58,530 --> 00:02:03,660 There's a lot of really interesting, uh, fun stuff happening with AI. 48 00:02:04,020 --> 00:02:09,030 Uh, when companies go out there and they say like, we have a great idea for how. 49 00:02:09,565 --> 00:02:14,035 We can give people an amazing feature, an amazing solution, a product, whatever, 50 00:02:14,035 --> 00:02:15,774 and is gonna use AI on the backend. 51 00:02:16,195 --> 00:02:19,015 Uh, a lot of times you get these product teams, they dive in, they do 52 00:02:19,015 --> 00:02:22,795 some cool stuff, some POCs, and then they're like, we're ready to launch. 53 00:02:22,855 --> 00:02:25,584 And then they start talking with their lawyers, their compliance teams, 54 00:02:25,584 --> 00:02:29,635 especially in enterprise departments with like five different security teams. 55 00:02:30,055 --> 00:02:35,155 And the skids, uh, happen and you have to figure out. 56 00:02:35,770 --> 00:02:39,640 What do we do to solve a lot of compliance stuff, and especially because 57 00:02:39,670 --> 00:02:43,780 AI, like people are sending all kinds of crazy stuff into these systems, 58 00:02:44,050 --> 00:02:46,510 uh, freeform text and, and whatnot. 59 00:02:46,780 --> 00:02:52,210 And so it's a, it's a challenging space, but what's really interesting 60 00:02:52,240 --> 00:02:55,240 is leading companies doing AI. 61 00:02:55,240 --> 00:02:57,760 So these are companies like A DFS themselves. 62 00:02:58,035 --> 00:03:00,135 Have their own AI products. 63 00:03:00,165 --> 00:03:04,905 Not like Bedrock is an AI product, but I mean like more like their DevOps 64 00:03:04,935 --> 00:03:10,605 agent, like it's using AI inter intercom with their, their fin, especially AI 65 00:03:10,635 --> 00:03:13,575 enabled customer, uh, service solutions. 66 00:03:13,815 --> 00:03:16,725 Sentry with, again, their kind of DevOps focused AI stuff. 67 00:03:16,725 --> 00:03:20,445 I've talked with people at companies like these and there's a lot of 68 00:03:20,445 --> 00:03:22,515 common things where they'll say, okay. 69 00:03:23,370 --> 00:03:27,090 To solve for security and governance and compliance. 70 00:03:27,120 --> 00:03:30,120 We have this kind of like central pipeline that we send 71 00:03:30,120 --> 00:03:33,480 all our LLM AI interactions to. 72 00:03:33,660 --> 00:03:36,000 And from there we figure out what do we need to do, audit 73 00:03:36,000 --> 00:03:39,690 logging, where do we send the raw logs, uh, analytics, how do we. 74 00:03:40,165 --> 00:03:44,035 Anonymize these things and send it to the right analytics warehouse 75 00:03:44,065 --> 00:03:48,864 or operational o hotel service, vendor of choice, potentially 76 00:03:48,864 --> 00:03:50,665 anomaly detection for security. 77 00:03:50,994 --> 00:03:53,274 But the point is there's a central. 78 00:03:54,060 --> 00:03:59,790 Pipeline where this is all governed and managed because of how radioactive 79 00:03:59,790 --> 00:04:02,429 it could be if this were mishandled. 80 00:04:02,579 --> 00:04:02,850 Corey Quinn: Yes. 81 00:04:02,910 --> 00:04:04,679 Even now at a, we're, we're a small startup. 82 00:04:04,679 --> 00:04:08,515 We're building software at Duck Bill, and in fact, this is probably a good. 83 00:04:09,135 --> 00:04:10,905 Point to wind up doing the sponsored read. 84 00:04:10,905 --> 00:04:13,785 Since we are now the only sponsor of this after years I have gotten 85 00:04:13,785 --> 00:04:17,655 it down there, uh, we are helping companies solve their cloud contracts 86 00:04:17,655 --> 00:04:19,755 with software and services to boot. 87 00:04:19,965 --> 00:04:24,615 If your cloud contract is expiring within the next year, please reach out to us. 88 00:04:24,615 --> 00:04:28,755 If nothing else, you can get that sad dead look in your eyes that we all get 89 00:04:28,755 --> 00:04:32,685 when talking about contracts combined with cloud billing and commiserate. 90 00:04:33,100 --> 00:04:35,620 Odds are we might be able to help you out with things. 91 00:04:35,680 --> 00:04:39,280 Our product Skyway is doing amazing things around contract management. 92 00:04:39,490 --> 00:04:43,750 Please kick us out at duckbillhq.com. 93 00:04:44,290 --> 00:04:49,000 But as we were mentioning from the AI perspective, yeah, we have an entire 94 00:04:49,000 --> 00:04:54,220 policy internally around which systems we're allowed to put AI through for 95 00:04:54,250 --> 00:04:58,270 models, how we get those models, and we have a standing policy that customer data 96 00:04:58,300 --> 00:05:00,010 does not touch AI just for, you know. 97 00:05:00,660 --> 00:05:02,490 Customer comfort if nothing else. 98 00:05:02,640 --> 00:05:02,940 Chase Douglas: Right? 99 00:05:03,270 --> 00:05:07,650 So there, there's a bunch of stuff in this space about how do you handle 100 00:05:07,650 --> 00:05:10,680 this, whether you're dealing with public information, customer data, you're 101 00:05:10,680 --> 00:05:15,030 mixing things, whatever it is, how do you make sure you have confidence? 102 00:05:15,210 --> 00:05:18,690 There's one place where all of this flows through a single pipeline 103 00:05:18,870 --> 00:05:20,340 that is properly anonymized. 104 00:05:20,340 --> 00:05:23,730 If it needs to be for certain endpoint destinations, you 105 00:05:23,730 --> 00:05:25,560 know, who accesses those? 106 00:05:25,560 --> 00:05:28,920 You've got logging on like, who can see my Snowflake cluster? 107 00:05:28,920 --> 00:05:31,830 Or you know, wherever I. Got my audit logs. 108 00:05:31,890 --> 00:05:37,830 Once you've got that in place, then you can have confidence to turn on the gas 109 00:05:37,830 --> 00:05:43,950 and really enable your product teams to build, deliver, ship awesome new features. 110 00:05:44,250 --> 00:05:47,700 So that's what we are working with customers to do, is 111 00:05:47,700 --> 00:05:49,710 to make this turnkey easy. 112 00:05:50,160 --> 00:05:55,710 Uh, and that's where, you know, my, uh, co-founder myself ap Perva being, uh, 113 00:05:55,710 --> 00:05:59,850 a previously AWS principal engineers working in an enterprise environment. 114 00:06:00,094 --> 00:06:05,284 Working with legal teams to define and develop processes and 115 00:06:05,284 --> 00:06:09,724 implementations for compliance, considerations, security, and whatnot. 116 00:06:10,145 --> 00:06:17,284 As we worked on AI enabled solutions, if you're saying to yourself that. 117 00:06:17,895 --> 00:06:21,315 You are in this problem space, having to figure this out, and you wish 118 00:06:21,315 --> 00:06:26,295 you had two AWS principal engineers working on this and helping you solve 119 00:06:26,295 --> 00:06:30,405 this and figuring out how to talk with your lawyers internally or externally. 120 00:06:31,365 --> 00:06:32,594 Give us a, a ring. 121 00:06:32,715 --> 00:06:35,385 You can hit me up on on LinkedIn, chase Douglas. 122 00:06:35,385 --> 00:06:36,914 You can go to our site. 123 00:06:36,914 --> 00:06:38,875 It's uh, Archodex.com. 124 00:06:38,895 --> 00:06:43,664 That's A-R-C-H like architecture, A-R-C-H-O-D-E x.com. 125 00:06:43,784 --> 00:06:45,974 Quest a demo, hit our contact form. 126 00:06:46,245 --> 00:06:47,265 Uh, whatever you wanna do. 127 00:06:47,414 --> 00:06:48,255 We're here to help you. 128 00:06:48,315 --> 00:06:53,534 We're early days, so we're building things that meet what our customers 129 00:06:53,625 --> 00:06:55,664 are asking us to build for them. 130 00:06:55,965 --> 00:06:57,615 Uh, hit us up so that we can help you. 131 00:06:57,974 --> 00:07:00,405 Corey Quinn: I, I do want to highlight. 132 00:07:00,835 --> 00:07:02,545 That this has become a problem. 133 00:07:02,545 --> 00:07:03,415 What, what are the approaches? 134 00:07:03,415 --> 00:07:06,565 I do, 'cause I do a lot of random one-offs with ai. 135 00:07:06,565 --> 00:07:08,065 Go vibe code me a thing. 136 00:07:08,365 --> 00:07:12,205 I, I just redid my entire email newsletter production pipeline 137 00:07:12,205 --> 00:07:13,555 over the break for that reason. 138 00:07:13,555 --> 00:07:14,965 I've been sitting on that for six years. 139 00:07:15,115 --> 00:07:18,055 It was $50,000 of development effort to have a human do it. 140 00:07:18,265 --> 00:07:21,775 Or I can now finally get to a point where I can just poke Claude code for 141 00:07:21,775 --> 00:07:24,955 a few weeks and it winds up spitting out something that works for me. 142 00:07:25,375 --> 00:07:25,975 And. 143 00:07:26,425 --> 00:07:30,685 The way I do this is I have, I run Claude in its own EC2 instance. 144 00:07:31,105 --> 00:07:32,995 It has root on the box. 145 00:07:33,685 --> 00:07:35,995 I run this in its own AWS account. 146 00:07:36,025 --> 00:07:40,195 It has an administrator access IAM role for that entire account. 147 00:07:40,285 --> 00:07:44,755 And that account is called Superfund, as in the government program 148 00:07:44,755 --> 00:07:46,195 to clean up toxic waste sites. 149 00:07:46,195 --> 00:07:50,125 It's expensive and it's toxic and it does a whole lot of terrible 150 00:07:50,125 --> 00:07:52,255 things, but there's nothing persistent in that account. 151 00:07:52,255 --> 00:07:53,695 And there's no long-term data. 152 00:07:53,845 --> 00:07:55,225 There's no data that lives there at all. 153 00:07:55,305 --> 00:07:56,985 That is, has any sensitivity. 154 00:07:57,045 --> 00:08:00,585 So everything to get out of there has to go through a CI/CD pipeline that has 155 00:08:00,585 --> 00:08:04,155 more guardrails on it because I don't want the innovation to get slowed down. 156 00:08:04,155 --> 00:08:07,215 But I also, you know, don't want to accidentally the database. 157 00:08:07,245 --> 00:08:11,715 I can't run quad code in dangerous mode on my laptop that has stuff on 158 00:08:11,715 --> 00:08:13,395 there that cannot see the light of day. 159 00:08:13,605 --> 00:08:15,645 So I have to be a little bit more cautious. 160 00:08:15,765 --> 00:08:16,845 Chase Douglas: Yeah, and it's. 161 00:08:17,115 --> 00:08:22,065 Also a challenge when there's this thing that people started calling Shadow AI. 162 00:08:22,455 --> 00:08:24,225 It's this idea of, oh yeah. 163 00:08:24,405 --> 00:08:29,805 When we are, uh, rolling out our AI enabled products, we put it through 164 00:08:29,805 --> 00:08:32,980 Bedrock, so it's going through all the AWS guard rails and everything. 165 00:08:32,980 --> 00:08:33,299 That's great. 166 00:08:34,245 --> 00:08:37,845 But what are your developers on the teams doing when they're spiking their 167 00:08:37,845 --> 00:08:41,924 POCs and they're testing out, you know, some, some changes or whatever? 168 00:08:42,105 --> 00:08:46,125 I've talked to so many engineers from so many companies who say, yeah, we have 169 00:08:46,125 --> 00:08:48,074 Bedrock, but like, it's locked down. 170 00:08:48,135 --> 00:08:52,814 I can't use Bedrock to really figure out what model's gonna work best and, 171 00:08:52,814 --> 00:08:54,855 and, and how to prompt it the right way. 172 00:08:55,125 --> 00:08:59,385 So I reach for things like, uh, open router.ai, which now lets me hit any 173 00:08:59,385 --> 00:09:01,395 number of models and test them out. 174 00:09:01,770 --> 00:09:04,920 But, uh, ahead of security and compliance at your company. 175 00:09:04,920 --> 00:09:06,270 They're like, wait, you're doing what? 176 00:09:06,420 --> 00:09:09,360 We're, we don't have any relationship with Open Router. 177 00:09:09,360 --> 00:09:10,500 We don't know where that goes. 178 00:09:10,500 --> 00:09:12,300 We don't know how they log things. 179 00:09:12,300 --> 00:09:14,250 We don't know what data you're mixing into it. 180 00:09:14,340 --> 00:09:15,270 So that's the so-called shadow. 181 00:09:15,270 --> 00:09:18,210 Corey Quinn: Did you study your entire code base for this thing into Claude Code? 182 00:09:18,210 --> 00:09:19,530 Well, in my case, jokes on you. 183 00:09:19,530 --> 00:09:20,850 Claude Code wrote the whole thing. 184 00:09:20,850 --> 00:09:21,810 Anyway, so. 185 00:09:22,060 --> 00:09:23,530 Chase Douglas: Really being able to capture that and 186 00:09:23,530 --> 00:09:24,640 understand what's going on. 187 00:09:24,910 --> 00:09:27,645 Um, that's the other really kind of interesting, innovative part of, 188 00:09:27,650 --> 00:09:31,780 of what we built is, uh, with kinda like a zero instrumentation approach. 189 00:09:31,780 --> 00:09:34,870 You can slap on a little bit of instrumentation to your like compute 190 00:09:34,870 --> 00:09:37,120 cluster and we'll be able to capture. 191 00:09:37,305 --> 00:09:42,285 Every AI LLM interaction and tool interaction, uh, that your workloads 192 00:09:42,285 --> 00:09:45,194 do, whether they're in dev test, production, qa, whatever it is. 193 00:09:45,344 --> 00:09:48,555 And so that's kind of an exciting thing too, that not only are we helping 194 00:09:48,555 --> 00:09:52,395 make sure that people have these centralized guardrail pipelines on 195 00:09:52,395 --> 00:09:55,724 the backend, we're making sure that they're capturing everything on the 196 00:09:55,724 --> 00:09:58,094 front end too, even in dev and test. 197 00:09:58,645 --> 00:10:00,205 Corey Quinn: I wanna kick the tires on this myself. 198 00:10:00,235 --> 00:10:02,575 Uh, one of the things that even in the Superfund account, as you can 199 00:10:02,575 --> 00:10:06,535 imagine, it uses in some cases the clo, the Anthropic, API, in some cases 200 00:10:06,535 --> 00:10:08,755 bedrock in one Godforsaken instance. 201 00:10:08,755 --> 00:10:13,675 It's using a lot of Vertex AI over on Google Cloud for unrelated reasons. 202 00:10:13,975 --> 00:10:15,475 And one of the challenges heard. 203 00:10:16,130 --> 00:10:17,090 Well, it's, it's great. 204 00:10:17,090 --> 00:10:19,699 The challenge I've got is that every quarter when they come out with new 205 00:10:19,699 --> 00:10:23,060 models, I get to play Whack-a-Mole throughout a bunch of different micro 206 00:10:23,060 --> 00:10:27,170 repos and figure out what's calling what, and I've been able to do it so far by 207 00:10:27,170 --> 00:10:32,210 API key and guess in check, but I don't wanna be running sonnet two anymore. 208 00:10:32,240 --> 00:10:35,300 So seeing what's making what called to aware would be helpful. 209 00:10:35,420 --> 00:10:38,210 Chase Douglas: My co-founder, he was just playing around with our demo that 210 00:10:38,210 --> 00:10:40,579 we built yesterday, and he noticed that. 211 00:10:41,100 --> 00:10:44,670 Open router now one of their previously free models. 212 00:10:45,000 --> 00:10:50,100 It's just returning errors saying, uh, we no longer do this anymore. 213 00:10:50,100 --> 00:10:54,480 We're not doing free tier of this model, so you gotta go find a different model. 214 00:10:54,569 --> 00:10:56,760 You know, I don't know how many people are using open router in production 215 00:10:56,760 --> 00:11:00,030 per se, but like this is a kind of challenge, like it is, as you said, 216 00:11:00,030 --> 00:11:04,650 whack-a-mole of how do we keep this stuff operational when there's. 217 00:11:05,070 --> 00:11:09,840 Thousands of different models out there and how you access them, 218 00:11:09,840 --> 00:11:11,370 what their price points are. 219 00:11:11,460 --> 00:11:12,930 They're changing all the time. 220 00:11:12,930 --> 00:11:16,950 You even noted AWS is changing their price points on some of their, 221 00:11:17,070 --> 00:11:18,150 I don't know if it was models. 222 00:11:18,150 --> 00:11:19,680 It might have been, what was it again? 223 00:11:19,680 --> 00:11:21,840 They changed their pricing recently on. 224 00:11:22,145 --> 00:11:25,295 Corey Quinn: Capacity blocks for MLS raised, uh, by 15%. 225 00:11:25,714 --> 00:11:28,444 Uh, they do claim to adjust them every quarter and everyone's whining. 226 00:11:28,444 --> 00:11:30,454 It's like spot dynamic price. 227 00:11:30,964 --> 00:11:34,475 Changing things once every three months is not my definition of dynamic. 228 00:11:34,625 --> 00:11:36,185 Maybe that's enterprise dynamism. 229 00:11:37,204 --> 00:11:37,444 Chase Douglas: Yeah. 230 00:11:37,444 --> 00:11:41,045 I suppose maybe, maybe the enterprise CFO is totally fine with that. 231 00:11:41,405 --> 00:11:44,189 Corey Quinn: Speaking of enterprise dynamism. 232 00:11:44,755 --> 00:11:50,695 I am curious as far as what it is that you have seen about re:Invent this year. 233 00:11:50,845 --> 00:11:52,165 Chase Douglas: So I went to re:Invent. 234 00:11:52,225 --> 00:11:56,695 I, uh, went there to have conversations with people, a little bit to 235 00:11:56,695 --> 00:12:00,445 learn, but most of the time I don't go to like sessions or whatnot. 236 00:12:00,445 --> 00:12:05,185 It's about like networking, finding people in the hallways 237 00:12:05,185 --> 00:12:07,105 at the events, and both like. 238 00:12:07,535 --> 00:12:11,915 Expanding the people I know who I can then follow up with when I 239 00:12:11,915 --> 00:12:15,155 realize, oh, they know a thing that I don't know and I wanna know more. 240 00:12:15,185 --> 00:12:16,865 And I could do that after the fact. 241 00:12:16,895 --> 00:12:21,455 But also just like, you know, in real time learning, what's the, what's the 242 00:12:21,455 --> 00:12:23,135 state of play, what are people doing? 243 00:12:23,225 --> 00:12:24,635 And yeah, this year. 244 00:12:25,155 --> 00:12:26,415 This year was interesting. 245 00:12:26,625 --> 00:12:31,215 I, I didn't go last year, but I had been to previous to that eight 246 00:12:31,215 --> 00:12:36,405 re:Invents in a row, so this year things felt a little different. 247 00:12:36,405 --> 00:12:41,415 I don't know what you kind of sensed in it, but it was getting harder. 248 00:12:41,820 --> 00:12:48,780 To just have those networked conversations with people I was talking with like Ben 249 00:12:48,780 --> 00:12:53,160 Keho and we were struggling just to meet up, like where where are we gonna meet? 250 00:12:53,160 --> 00:12:58,260 Because if you've been to re:Invent in the past, you know that it's in this sort of 251 00:12:58,260 --> 00:13:01,470 like very large casino conference center. 252 00:13:01,730 --> 00:13:08,120 Place, and if you are walking around, there are places where 253 00:13:08,150 --> 00:13:12,050 there you're, it extends even into a mall and there's benches. 254 00:13:12,050 --> 00:13:17,210 There's this place called like St. Peter's Square with a ton of tables and chairs 255 00:13:17,210 --> 00:13:19,190 and you can sit and have a conversation. 256 00:13:19,550 --> 00:13:24,200 There are restaurants lining all of these venues that 10 years ago 257 00:13:24,200 --> 00:13:26,690 you could just go to any of these restaurants during the day and just sit 258 00:13:26,690 --> 00:13:28,430 down and have a coffee or something. 259 00:13:28,700 --> 00:13:30,230 Over the past few years. 260 00:13:30,245 --> 00:13:34,325 They would get sort of like locked down in the evening for private events. 261 00:13:34,685 --> 00:13:39,515 Uh, but this year, every single one of those places, including the ILI 262 00:13:39,545 --> 00:13:45,605 coffee shop, was locked out as a private event the whole day and night. 263 00:13:45,725 --> 00:13:48,845 Corey Quinn: Mike, my business partner, was saying that the catchphrase for 264 00:13:48,845 --> 00:13:50,375 re:Invent is closed for private events. 265 00:13:50,525 --> 00:13:54,125 God forbid you're making a terrible decision to vacation there that week. 266 00:13:54,275 --> 00:13:55,530 Where are you supposed to eat? 267 00:13:56,390 --> 00:13:58,760 Every restaurant is booked up and closed. 268 00:13:58,790 --> 00:13:59,780 It's awful. 269 00:14:00,140 --> 00:14:04,130 And he couldn't find a cup of coffee at one point, which if you know Mike, that's 270 00:14:04,130 --> 00:14:05,690 like three quarters of his personality. 271 00:14:05,900 --> 00:14:09,200 Chase Douglas: I just needed to get some easy, quick food. 272 00:14:09,560 --> 00:14:15,445 I ended up going like one block down, which Vegas one block is not a 273 00:14:15,450 --> 00:14:17,030 a block anywhere else in the world. 274 00:14:17,240 --> 00:14:21,140 One block down to get to the closest. 275 00:14:21,540 --> 00:14:27,420 Easily accessible fast foods like Subway or whatever, which is in like the very 276 00:14:27,420 --> 00:14:33,030 back of a low key like slots casino. 277 00:14:33,209 --> 00:14:34,020 And I was like, you know what? 278 00:14:34,020 --> 00:14:38,010 At, at least, at least if I went this far, it, it's semi quiett here. 279 00:14:38,010 --> 00:14:41,400 I got the bling going on of like the slots, but I, I don't 280 00:14:41,400 --> 00:14:43,410 know, like, it, it's hard. 281 00:14:43,439 --> 00:14:46,349 It's hard, you know, like I'm friends with, you know, many other 282 00:14:46,349 --> 00:14:47,520 people from lots of different. 283 00:14:47,970 --> 00:14:48,900 Companies. 284 00:14:49,020 --> 00:14:52,140 I was talking with some people from one company that has sponsored, 285 00:14:52,410 --> 00:14:56,460 uh, re:Invent in the past and had had booze, uh, at their expo and 286 00:14:56,460 --> 00:14:59,430 everything, and they're like, yeah, we're not even doing that this year. 287 00:14:59,460 --> 00:15:05,975 Like, it's so expensive as a vendor to show up and the leads are. 288 00:15:07,005 --> 00:15:10,545 It's hard to justify what you're getting out of it as the expensive had gone up 289 00:15:10,574 --> 00:15:13,155 and up and up, uh, for the conference. 290 00:15:13,245 --> 00:15:15,405 And so this is kind of happening on both sides. 291 00:15:15,405 --> 00:15:19,454 You got like a ticket to go to the conference is $2,100 this year. 292 00:15:19,545 --> 00:15:25,365 The sponsorship is like crazy entry price probably of like 50 293 00:15:25,365 --> 00:15:28,155 to 70 5K for a tiny little booth. 294 00:15:28,214 --> 00:15:33,375 At this point, I don't actually know and it reminds me of like Cory Eros. 295 00:15:33,680 --> 00:15:35,780 Idea of like in ification, 296 00:15:35,930 --> 00:15:37,190 Corey Quinn: it's, it's gotten terrible. 297 00:15:37,280 --> 00:15:41,330 I still, it, it was burned into my memory years ago that at re:Invent 298 00:15:41,330 --> 00:15:44,330 in the opening keynote Andy was talking about, what is this for? 299 00:15:44,330 --> 00:15:45,830 Is it a sales conference? 300 00:15:45,830 --> 00:15:46,700 No, he lied. 301 00:15:46,700 --> 00:15:48,740 Given how expensive everything has become. 302 00:15:49,160 --> 00:15:51,770 Is it a, is it a training event? 303 00:15:51,920 --> 00:15:52,340 What is it? 304 00:15:52,400 --> 00:15:53,810 He said it's about education. 305 00:15:53,840 --> 00:15:55,130 It's about learning things. 306 00:15:56,000 --> 00:15:56,450 Yeah. 307 00:15:56,450 --> 00:15:59,930 The problem I'm seeing is that the educational has a very, 308 00:16:00,170 --> 00:16:02,630 speaking from, uh, doctrine. 309 00:16:03,000 --> 00:16:06,900 As opposed to speaking from reality perspective, I couldn't find places 310 00:16:06,900 --> 00:16:10,440 to sit this year, and I don't know if that's because there were fewer 311 00:16:10,440 --> 00:16:13,590 places to sit or I just wasn't invited to some of the special places. 312 00:16:13,710 --> 00:16:16,890 I too got to sit with Ben Kehoe and chat, and the reason I was able to 313 00:16:16,890 --> 00:16:20,220 do that is because I snuck my way into a place I didn't belong because 314 00:16:20,220 --> 00:16:22,080 everyone thinks I'm an AWS hero. 315 00:16:22,710 --> 00:16:26,250 This episode is sponsored by my own company, duck Bill. 316 00:16:26,550 --> 00:16:30,090 Having trouble with your AWS bill, perhaps it's time to 317 00:16:30,090 --> 00:16:32,189 renegotiate a contract with them. 318 00:16:32,520 --> 00:16:37,890 Maybe you're just wondering how to predict what's going on in the wide world of AWS. 319 00:16:37,949 --> 00:16:40,590 Well, that's where Duck Bill comes in to help. 320 00:16:40,800 --> 00:16:43,530 Remember, you can't duck the duck bill. 321 00:16:43,530 --> 00:16:47,130 Bill, which I am reliably informed by my business partner 322 00:16:47,280 --> 00:16:49,740 is absolutely not our motto. 323 00:16:49,815 --> 00:16:52,995 To learn more, visit doc bill hq.com. 324 00:16:54,105 --> 00:16:58,815 Chase Douglas: There was some like ex AWS hero allowances that I availed 325 00:16:58,815 --> 00:17:00,555 myself of as well in the same way. 326 00:17:00,555 --> 00:17:06,765 But, but to your point, the thing I found hilarious was the Venetian entrance. 327 00:17:06,855 --> 00:17:10,755 Uh, so the Venetian is like the, the casino hotel that this is in, and it's 328 00:17:10,755 --> 00:17:14,025 got its own entrance, separate from the palazzo, separate from a bunch 329 00:17:14,025 --> 00:17:18,525 of other places you can get in The Venetian entrance had historically. 330 00:17:18,865 --> 00:17:21,714 This kind of like, you'd almost think it was a fountain, but there's 331 00:17:21,714 --> 00:17:23,095 actually no water fixture in there. 332 00:17:23,095 --> 00:17:26,845 It's just kind of like a planted area in this rotunda. 333 00:17:27,115 --> 00:17:32,215 And you could sit there on these low like areas. 334 00:17:32,365 --> 00:17:33,655 It was almost like a bench. 335 00:17:33,715 --> 00:17:38,665 And this year there were these like metal decorative grates 336 00:17:38,785 --> 00:17:41,185 that had like spiky tips. 337 00:17:41,774 --> 00:17:42,975 On those benches. 338 00:17:42,975 --> 00:17:44,835 You could not even sit there. 339 00:17:44,835 --> 00:17:49,245 I had to laugh when I was walking and I'm like, oh, I know I can go down there. 340 00:17:49,245 --> 00:17:52,784 And I walk a quarter mile down to that place and I get there and it's like, 341 00:17:53,115 --> 00:17:55,845 what's greeting me is metal spiky things. 342 00:17:55,845 --> 00:17:57,555 So I, I can't even sit there. 343 00:17:57,555 --> 00:18:01,485 So this is while I was doing a FaceTime with my family, uh, one of 344 00:18:01,485 --> 00:18:05,115 the evenings and I'm like, oh, turns out I'm just gonna be pacing around 345 00:18:05,115 --> 00:18:06,945 this rotunda as I talk with you. 346 00:18:06,975 --> 00:18:09,105 'cause I, I got nowhere to just sit. 347 00:18:09,254 --> 00:18:09,945 But yeah, like. 348 00:18:10,465 --> 00:18:13,675 Corey, Dr. O, the, like other Corey, 349 00:18:14,005 --> 00:18:15,955 Corey Quinn: he spells his name incorrectly, but I like him anyway. 350 00:18:16,405 --> 00:18:16,855 Chase Douglas: Right. 351 00:18:17,185 --> 00:18:21,475 Uh, is like phase one of gentrification is like, you know, making it, what was 352 00:18:21,475 --> 00:18:24,115 it like super awesome for the consumers? 353 00:18:24,115 --> 00:18:29,515 Like, you know, re:Invent 10, 15 years ago is all about education and then phase 354 00:18:29,515 --> 00:18:32,815 two is we're making up not so awesome, but the consumers we're making it super 355 00:18:32,815 --> 00:18:35,425 awesome for, for like vendors businesses. 356 00:18:35,425 --> 00:18:38,215 So like, okay, now you know, it's a great. 357 00:18:38,410 --> 00:18:41,440 Place to go and, and, and have a booth and vendor or whatever. 358 00:18:41,500 --> 00:18:44,800 And then like, phase three is we're now raking in so much money. 359 00:18:44,800 --> 00:18:45,850 This is our business. 360 00:18:45,850 --> 00:18:50,590 We're not making it any awesome for either the consumers or the businesses. 361 00:18:50,890 --> 00:18:53,800 And it's, it's just a money machine now. 362 00:18:53,800 --> 00:18:55,780 And I feel like that's what re:Invent is. 363 00:18:55,780 --> 00:19:01,000 It's like a WI, I don't actually know what the revenue profit is like 364 00:19:01,000 --> 00:19:04,750 for re:Invent itself, but you know, I, I just look at the pricing and 365 00:19:04,750 --> 00:19:05,920 everything and I'm like, well, who's. 366 00:19:05,990 --> 00:19:08,900 Who's actually making out well here and it's like, well I think 367 00:19:08,900 --> 00:19:13,610 it's AWS is primarily making out well here for their conference. 368 00:19:13,610 --> 00:19:15,080 So, yeah. 369 00:19:15,500 --> 00:19:17,990 Corey Quinn: So one thing I'm curious about is. 370 00:19:18,810 --> 00:19:22,500 I need to talk to more folks in various spaces, but re:Invent 371 00:19:22,500 --> 00:19:23,820 has lost all sense of itself. 372 00:19:23,970 --> 00:19:29,490 I gave a talk two talks this year, last year, now that this is January, and 373 00:19:29,970 --> 00:19:32,760 it was the only time for those talks where I left the Venetian and it was 374 00:19:32,760 --> 00:19:34,020 like a completely separate conference. 375 00:19:34,020 --> 00:19:34,860 There was entire world. 376 00:19:34,860 --> 00:19:36,780 I never got to see this. 377 00:19:36,930 --> 00:19:39,840 It could have been two separate conferences for all I noticed. 378 00:19:40,050 --> 00:19:45,075 So it was a. I don't fully understand the value or who people, what value 379 00:19:45,075 --> 00:19:46,605 people think they're getting from it. 380 00:19:46,845 --> 00:19:49,754 We're looking at budgeting for marketing for the year, and that's 381 00:19:49,754 --> 00:19:51,524 pretty expensive to sponsor that stuff. 382 00:19:51,524 --> 00:19:54,495 Like people like, uh oh, it'd be great if we had like a plate, like 383 00:19:54,495 --> 00:19:57,555 a desk we could do a demo at at our booth, rather just the kiosk. 384 00:19:57,555 --> 00:19:58,725 It's like, yeah, that, that's. 385 00:19:58,730 --> 00:20:01,400 That's gonna have a six figure build hatch to it if we want that. 386 00:20:01,430 --> 00:20:02,600 Like, is it worth that much? 387 00:20:02,600 --> 00:20:03,290 I don't know. 388 00:20:03,920 --> 00:20:05,810 It's a, it feels like a cash grab. 389 00:20:05,990 --> 00:20:09,740 Chase Douglas: I do have to say the best event I went to at 390 00:20:09,740 --> 00:20:14,780 re:Invent this year was your event at Atomic Liquors, because it was. 391 00:20:15,360 --> 00:20:18,180 Well off strip over in the Fremont area, 392 00:20:18,330 --> 00:20:19,350 Corey Quinn: it was free. 393 00:20:19,710 --> 00:20:20,850 The good people show off. 394 00:20:20,850 --> 00:20:22,890 Some people just show fly into town just for that event. 395 00:20:23,070 --> 00:20:26,610 Chase Douglas: And the people who end up there are there because 396 00:20:26,610 --> 00:20:29,010 they know it's going to be good. 397 00:20:29,010 --> 00:20:31,770 They're not just there like, where's my next free drink 398 00:20:31,770 --> 00:20:33,060 on the strip kind of thing. 399 00:20:33,570 --> 00:20:37,620 Uh, and I had such great conversations there that was like re:Invent of 400 00:20:37,620 --> 00:20:40,710 the old, but you have to find it in pockets here and there now. 401 00:20:41,284 --> 00:20:42,635 Corey Quinn: Uh, that is my favorite part of re:Invent. 402 00:20:42,635 --> 00:20:45,844 I would do that regardless, but the rest of it, ugh. 403 00:20:46,294 --> 00:20:51,635 I am at a point now where I don't go to sessions other than the ones I'm speaking 404 00:20:51,635 --> 00:20:54,935 at, because I don't like waiting in line forever to get turned away at the door. 405 00:20:55,415 --> 00:20:59,584 I a, a lot of the people I normally want to talk to just aren't there 406 00:20:59,584 --> 00:21:00,905 in the way that they used to be. 407 00:21:01,294 --> 00:21:04,175 Uh, I talked to like some of the folks that I love spending time 408 00:21:04,175 --> 00:21:05,915 with and re:Inventing the most of 'em just weren't showing up. 409 00:21:06,420 --> 00:21:09,690 It is a shadow of itself and I get nothing good can stay. 410 00:21:09,870 --> 00:21:13,020 That is, that is the inherent problem, like nostalgia. 411 00:21:13,050 --> 00:21:17,760 We always want things to be like they used to be, but I do find myself 412 00:21:17,760 --> 00:21:21,000 wondering who this conference is for, because increasingly it's not for me. 413 00:21:21,180 --> 00:21:22,170 And maybe that's okay. 414 00:21:22,470 --> 00:21:25,020 There's a lot more people out there who are not like me than 415 00:21:25,020 --> 00:21:26,250 people who are thank the world. 416 00:21:26,580 --> 00:21:29,340 And maybe that's the right answer. 417 00:21:29,730 --> 00:21:31,740 I'm sure AWS has a vision for this. 418 00:21:32,235 --> 00:21:34,125 I hope, I just don't know what it's, 419 00:21:34,455 --> 00:21:37,785 Chase Douglas: yeah, and the other interesting thing is I was curious when I 420 00:21:37,995 --> 00:21:42,435 was just seeing, like I follow the what's new feed to, to keep up with AWS and 421 00:21:42,435 --> 00:21:44,865 especially, uh, at re:Invent every day. 422 00:21:44,865 --> 00:21:47,355 It's like, okay, so what was actually released here? 423 00:21:47,685 --> 00:21:49,695 And uh, I remember. 424 00:21:50,085 --> 00:21:56,565 Five to 10 years ago, every day at re:Invent was like another hundred 425 00:21:56,625 --> 00:22:01,455 releases, and it would actually take me a couple hours just to sift through them. 426 00:22:01,485 --> 00:22:03,435 I'm not reading every one of 'em, I'm just sifting through the 427 00:22:03,435 --> 00:22:07,335 headlines and then poking in on the 10 that that are interesting. 428 00:22:07,695 --> 00:22:10,215 Now, that was also during the heyday of them just. 429 00:22:10,475 --> 00:22:14,195 Releasing anything for any niche that anyone said. 430 00:22:14,195 --> 00:22:14,975 Yeah, sure. 431 00:22:14,975 --> 00:22:19,024 That sounds like I would like that from AWS Yeah, go build that. 432 00:22:19,205 --> 00:22:21,425 And they would, it's not necessarily good either. 433 00:22:21,455 --> 00:22:26,344 But, uh, I went back, uh, I could only go back in their, uh, feed history to 434 00:22:26,344 --> 00:22:33,034 like 2022, but you could see like the amount of releases has, uh, dwindled too. 435 00:22:33,395 --> 00:22:36,274 So like 2022 was something like. 436 00:22:36,330 --> 00:22:40,350 101 releases, that's still coming outta COVID a little bit. 437 00:22:40,500 --> 00:22:42,449 Twenty twenty three, a hundred twenty seven releases. 438 00:22:42,899 --> 00:22:48,210 And then, uh, 20 24 66 releases, 20 25 69 releases. 439 00:22:48,480 --> 00:22:51,030 So it's really coming down and. 440 00:22:51,680 --> 00:22:56,270 You know, some of that is, I think like pivot to AI and having to re-figure out 441 00:22:56,270 --> 00:22:57,740 what are, what are they doing in AI. 442 00:22:58,070 --> 00:23:04,130 But I think some of it is also a little bit of like where is the continued 443 00:23:04,400 --> 00:23:06,380 innovation that they're known for. 444 00:23:06,530 --> 00:23:10,460 I'm gonna be very curious to see like just kind of how the future 445 00:23:10,460 --> 00:23:12,710 continues to unfold for AWS. 446 00:23:12,710 --> 00:23:14,180 'cause a lot of the people there. 447 00:23:14,270 --> 00:23:17,150 Who remain, who are super smart and doing amazing stuff. 448 00:23:17,240 --> 00:23:20,720 And you see that in some of the stuff that, that continues to get released. 449 00:23:20,840 --> 00:23:25,460 And you've also seen, you know, as I have, I'm sure, uh, a lot of people have left 450 00:23:25,550 --> 00:23:29,900 AWS for various reasons and I, and I kind of like, what, what is happening with 451 00:23:29,900 --> 00:23:35,510 the innovation And if that's part of what has also driven re:Invent over the years 452 00:23:35,510 --> 00:23:37,760 and what's been amazing about it, like. 453 00:23:38,435 --> 00:23:44,284 Where is that going to, that input that feeds into what makes re:Invent great. 454 00:23:44,764 --> 00:23:48,545 Corey Quinn: I am deeply curious to see how that's going to shape up. 455 00:23:48,545 --> 00:23:53,435 I, I just, at this point, it feels like it can't be getting results 456 00:23:53,585 --> 00:23:56,524 for folks like you take a look at some of these sponsor packages for 457 00:23:56,524 --> 00:23:58,340 tens of millions of dollars and. 458 00:23:59,265 --> 00:24:03,135 Are, are you generating that much business from re:Invent for that 459 00:24:03,135 --> 00:24:03,975 Chase Douglas: brand marketing? 460 00:24:03,975 --> 00:24:06,675 I think it's, I think it's brand marketing at this point. 461 00:24:06,705 --> 00:24:11,805 Uh, when you can go up and down the strip and see on every casino's, giants 462 00:24:11,805 --> 00:24:17,265 video board, about every two out of every three ads is for a tech company. 463 00:24:17,265 --> 00:24:22,605 Come see us at Expo booth number 5 92, and then hilariously the other 464 00:24:22,605 --> 00:24:25,725 third are for like Rodeo Cowboy. 465 00:24:25,815 --> 00:24:30,975 Gear because like as soon as re:Invent leaves, it's now like a, a like 466 00:24:30,975 --> 00:24:33,675 a PBR rodeo event or something. 467 00:24:33,975 --> 00:24:36,165 But yeah, it's, uh, it's, it's crazy. 468 00:24:36,225 --> 00:24:36,885 I, I don't know. 469 00:24:37,035 --> 00:24:40,185 Not a CMO at a, at a giant company, so 470 00:24:40,455 --> 00:24:41,295 Corey Quinn: it's a weird mess. 471 00:24:41,505 --> 00:24:44,985 If I really wanna thank you for taking the time to chat with me about this. 472 00:24:45,045 --> 00:24:46,995 Uh, what are you looking forward to this year? 473 00:24:47,055 --> 00:24:47,925 It's 2026. 474 00:24:48,615 --> 00:24:51,270 We are on the cusp of something interesting happening, I think. 475 00:24:52,290 --> 00:24:53,520 What are you excited about? 476 00:24:53,520 --> 00:24:56,010 Like there's a lot of pessimism is out there, but what do you 477 00:24:56,070 --> 00:24:59,065 actually, what makes you smile when you think about this cove year? 478 00:24:59,970 --> 00:25:01,110 Because if it's re:Invent, we're done. 479 00:25:01,500 --> 00:25:04,590 Chase Douglas: This is a little bit more like individual, a little 480 00:25:04,590 --> 00:25:06,810 bit more, more personal sized. 481 00:25:07,290 --> 00:25:09,870 I'll tell you what I have really enjoyed. 482 00:25:10,679 --> 00:25:16,980 I have had moments of, of skepticism about AI, and I also wanna say I'm not, like, I 483 00:25:16,980 --> 00:25:22,679 know there's environmental concerns about it too, and I want to hope that, uh, we 484 00:25:22,679 --> 00:25:27,120 will find ways of lessening its impact on all kinds of different ways, uh, and 485 00:25:27,120 --> 00:25:29,070 not just try to shove it under the rug. 486 00:25:29,135 --> 00:25:34,415 But that said, there's a lot of interesting stuff happening in AI 487 00:25:34,415 --> 00:25:38,345 where we're starting to realize on the coding front, like what does it mean? 488 00:25:39,000 --> 00:25:44,700 To code efficiently with AI when you get into spec driven, test driven development. 489 00:25:45,120 --> 00:25:47,820 And I have actually found it fun. 490 00:25:48,270 --> 00:25:51,060 Midpoint of last year, I was like throwing spaghetti at the 491 00:25:51,060 --> 00:25:52,740 wall, like AI, go do this for me. 492 00:25:52,740 --> 00:25:54,600 Try it, and getting terrible results. 493 00:25:54,690 --> 00:25:57,150 And then, you know, talking with people, it's like you need to do 494 00:25:57,150 --> 00:25:59,460 this spec driven kind of development. 495 00:25:59,580 --> 00:26:04,680 And all of a sudden it's like unlocked so much productivity and yet I still 496 00:26:04,680 --> 00:26:07,500 get enjoyment when I get to say to my. 497 00:26:07,730 --> 00:26:11,149 AI coding tool of choice being like, yeah, cool. 498 00:26:11,149 --> 00:26:16,550 I see what you did there, but my, my Rust program that you wrote, like you shouldn't 499 00:26:16,550 --> 00:26:18,590 be using threads for parallelism. 500 00:26:18,710 --> 00:26:26,064 You should be using Async Tokyo mechanisms and being able to take the raw. 501 00:26:27,120 --> 00:26:28,620 Computer engineering. 502 00:26:28,800 --> 00:26:33,390 You know, like I, I went to school for computer engineering using stuff I learned 503 00:26:33,390 --> 00:26:39,600 from the operating systems class back then still today, to mix that in, even 504 00:26:39,600 --> 00:26:45,450 though I'm not handwriting the code, I am still activating that part of my brain 505 00:26:45,660 --> 00:26:49,795 that's putting that to use and knowing that it's creating better software. 506 00:26:50,820 --> 00:26:54,930 Because I'm putting the right guardrails around the stuff. 507 00:26:55,080 --> 00:27:02,100 So I will say that coming out of 2025, I'm much more hopeful about just what it means 508 00:27:02,100 --> 00:27:04,590 to be a software engineer than I was. 509 00:27:05,115 --> 00:27:09,105 Coming into 2025 and hearing all this AI stuff, and yet 510 00:27:09,105 --> 00:27:11,295 feeling very unsure about it. 511 00:27:11,355 --> 00:27:13,095 I would say that I, I'd say that 512 00:27:13,274 --> 00:27:17,774 Corey Quinn: I'm very happy with a lot of what Claude Code is putting out and yeah. 513 00:27:17,774 --> 00:27:19,545 Does it, is it architecturally inconsistent? 514 00:27:19,545 --> 00:27:19,995 Yes. 515 00:27:19,995 --> 00:27:24,405 So is my own nonsense what of it you can iterate forward on it and 516 00:27:24,405 --> 00:27:27,254 efficiency of code versus does it work? 517 00:27:27,669 --> 00:27:28,480 Is important. 518 00:27:28,480 --> 00:27:31,090 So the question there right now is how do we review all of this 519 00:27:31,090 --> 00:27:32,590 slop code that's being generated? 520 00:27:33,280 --> 00:27:34,929 Well, how do we build systems that do that? 521 00:27:35,110 --> 00:27:38,169 Because on some level, like, well then I won't know what every 522 00:27:38,169 --> 00:27:39,550 line of code in production is. 523 00:27:39,669 --> 00:27:41,050 Yeah, you're right. 524 00:27:41,470 --> 00:27:42,100 You won't. 525 00:27:42,700 --> 00:27:46,330 Do you look at the compiled output in some languages? 526 00:27:46,389 --> 00:27:46,990 Of course you don't. 527 00:27:47,169 --> 00:27:47,770 No one does. 528 00:27:48,070 --> 00:27:51,340 So how do we wind up making sure it does the thing that we think 529 00:27:51,340 --> 00:27:54,730 it's doing and that leads to better test harnesses, uh, better. 530 00:27:55,334 --> 00:27:57,705 Controls around things that have criticality in them. 531 00:27:57,975 --> 00:28:01,605 But if I'm building something for fun or a demo for shit posting, 532 00:28:01,605 --> 00:28:03,105 I don't care as much as I do. 533 00:28:03,105 --> 00:28:04,125 If it's the payments code. 534 00:28:04,575 --> 00:28:07,635 There's a difference here, and that is, and there's a different 535 00:28:07,635 --> 00:28:08,865 level of rigor that I bring to it. 536 00:28:09,165 --> 00:28:14,024 Chase Douglas: And, and also I, I think it's useful to point out early career 537 00:28:14,054 --> 00:28:18,524 for myself when I reviewed code, I would review it line by line, like, 538 00:28:18,645 --> 00:28:23,415 is this actually doing the right thing Later career for myself to this point, 539 00:28:23,774 --> 00:28:29,565 as you start to get seniority and, and tenure and you're actually helping. 540 00:28:29,990 --> 00:28:35,210 Organizations of teams, build, ship, deliver, and, and manage 541 00:28:35,210 --> 00:28:37,160 their, their, uh, their code bases. 542 00:28:37,490 --> 00:28:41,780 You start to realize, like, it, you can't do line by line reviews. 543 00:28:41,810 --> 00:28:45,110 It's way more about what are the things I'm seeing here? 544 00:28:45,230 --> 00:28:46,040 What kinds of like. 545 00:28:46,125 --> 00:28:48,315 How did you think about the tests you chose to write? 546 00:28:48,525 --> 00:28:50,805 How did you think about the architecture you chose? 547 00:28:51,045 --> 00:28:53,355 How did you think about some of your implementation choices? 548 00:28:53,355 --> 00:28:54,495 The library choices? 549 00:28:54,495 --> 00:28:57,075 And I'm gonna poke in and just ask you this and ask you that. 550 00:28:57,345 --> 00:29:01,395 And if you're giving me good answers, then I'm gonna have way more confidence. 551 00:29:01,395 --> 00:29:02,835 This is doing what it should be doing. 552 00:29:03,135 --> 00:29:06,225 And if there are issues, it's probably just a bug that we need to address. 553 00:29:06,225 --> 00:29:08,775 But like every code has bugs you gotta address eventually. 554 00:29:09,255 --> 00:29:12,735 And if you're giving me not so good answers, then it's like, okay, now 555 00:29:12,735 --> 00:29:14,265 we need to dig a little deeper. 556 00:29:14,675 --> 00:29:21,185 That kind of skillset is still highly applicable to AI. 557 00:29:21,215 --> 00:29:24,485 I will have times where I'm like, I see what you're doing there, but 558 00:29:24,545 --> 00:29:28,025 I want you to change it a little bit because I know we needed to be 559 00:29:28,025 --> 00:29:29,495 doing it a slightly different way. 560 00:29:29,795 --> 00:29:32,915 There are other times I was dealing with some code yesterday where I'm. 561 00:29:33,885 --> 00:29:35,685 No, you're doing it wrong. 562 00:29:35,685 --> 00:29:38,235 And I'm telling you, I know you're doing it wrong. 563 00:29:38,595 --> 00:29:40,665 Uh, you need to be doing it this way. 564 00:29:40,665 --> 00:29:44,565 It was like it was even just trying to put some heavy dependencies into 565 00:29:44,715 --> 00:29:48,585 regular dependencies instead of just developer develop, uh, dependencies. 566 00:29:48,645 --> 00:29:51,645 And I had to like really say, no, you're doing this wrong. 567 00:29:51,645 --> 00:29:53,010 And eventually it's like, you know what? 568 00:29:53,015 --> 00:29:53,685 You're rights. 569 00:29:53,895 --> 00:29:56,475 That should be in the developer dependencies, so it doesn't 570 00:29:56,475 --> 00:29:57,555 end up in the container image. 571 00:29:57,555 --> 00:29:58,785 We're gonna be shipping to customers. 572 00:29:59,145 --> 00:30:07,575 That kind of skillset is still applicable in AI driven, uh, or, or augmented 573 00:30:07,845 --> 00:30:13,065 development today that, that has me happy about being a software engineer. 574 00:30:13,065 --> 00:30:13,395 Still, 575 00:30:13,605 --> 00:30:16,155 Corey Quinn: I find that, and this is gonna be controversial, I get it. 576 00:30:16,635 --> 00:30:18,735 People tend to view. 577 00:30:19,605 --> 00:30:21,975 Software engineering in different ways. 578 00:30:22,095 --> 00:30:25,995 For me, it's never been typing, it's never been wrestling with 579 00:30:26,205 --> 00:30:28,305 weird syntax or boilerplate. 580 00:30:28,635 --> 00:30:32,745 It's never been getting all of the weird things set up correctly. 581 00:30:32,745 --> 00:30:35,925 It's been more about architecture and designing the overall flow. 582 00:30:36,435 --> 00:30:38,055 Uh, maybe I'm the weird one. 583 00:30:38,055 --> 00:30:41,715 I've never been a professional software engineer, just enthusiastic, amateur. 584 00:30:42,165 --> 00:30:44,385 That's the part of it that I like and. 585 00:30:44,970 --> 00:30:47,910 I think that people are some, in some cases missing the forest for the trees. 586 00:30:47,970 --> 00:30:49,110 I'm sure that'll come back to bite me 587 00:30:49,320 --> 00:30:49,800 Chase Douglas: at least. 588 00:30:49,860 --> 00:30:53,730 There's always gonna be software that needs to be written, just like we are 589 00:30:53,730 --> 00:30:58,890 always gonna be engineering buildings to be constructed, and we're always 590 00:30:58,890 --> 00:31:04,890 going to be like engineering as just a skillset is always going to be needed. 591 00:31:05,220 --> 00:31:11,100 And so it's just a matter of figuring out like what are our best tools to use today 592 00:31:11,400 --> 00:31:13,945 and how do we figure out how to use them? 593 00:31:14,685 --> 00:31:19,185 Efficiently and hopefully enjoy using them in the process. 594 00:31:19,515 --> 00:31:20,535 Corey Quinn: We should all be so lucky. 595 00:31:20,535 --> 00:31:21,765 Maybe that's something to look forward to. 596 00:31:22,215 --> 00:31:24,915 Chase, I wanna thank you for taking the time to speak with me Again. 597 00:31:24,915 --> 00:31:28,245 People could visit you at archoex.com to learn more. 598 00:31:28,545 --> 00:31:32,475 Chase Douglas: Hit me up A, as I said, LinkedIn even it may be easier for 599 00:31:32,475 --> 00:31:34,155 you to just remember, chase Douglas. 600 00:31:34,545 --> 00:31:35,685 Send me a message. 601 00:31:35,715 --> 00:31:39,645 Connect like, there's lots of ways to reach me. 602 00:31:39,740 --> 00:31:44,060 To reach us and what we're doing with AI, governance, security 603 00:31:44,060 --> 00:31:45,440 compliance, that kind of stuff. 604 00:31:45,740 --> 00:31:49,340 Uh, for organizations, you know, we're, we're new so we don't have 605 00:31:49,340 --> 00:31:51,890 the gargantuan marketing reach yet. 606 00:31:51,950 --> 00:31:56,540 But yeah, if you, like I said, if you want Amazon Caliber Principle 607 00:31:56,540 --> 00:32:01,250 Engineers helping you figure out how to do compliance and governance and 608 00:32:01,250 --> 00:32:03,950 security for your AI workloads, find us. 609 00:32:03,980 --> 00:32:04,560 We'll help you out. 610 00:32:04,960 --> 00:32:07,600 Corey Quinn: And hopefully you don't want Amazon caliber UX 611 00:32:07,600 --> 00:32:09,490 designers implementing the same. 612 00:32:09,790 --> 00:32:11,650 Chase, thank you so much for your time. 613 00:32:11,650 --> 00:32:12,400 I appreciate it. 614 00:32:12,435 --> 00:32:13,180 Chase Douglas: Yeah, yeah, yeah. 615 00:32:13,270 --> 00:32:14,380 This has been so much fun. 616 00:32:14,500 --> 00:32:15,700 Anytime, anytime. 617 00:32:15,970 --> 00:32:18,170 Corey Quinn: Chase Douglas, CEO at Archodex. 618 00:32:18,250 --> 00:32:19,930 I'm cloud economist Cory Quinn. 619 00:32:20,170 --> 00:32:21,550 This is Screaming in the Cloud. 620 00:32:21,550 --> 00:32:24,580 My thanks to my own company, duck Bill HQ for sponsoring. 621 00:32:25,159 --> 00:32:28,100 And if you've enjoyed this podcast, please leave a five star review on 622 00:32:28,100 --> 00:32:29,659 your podcast platform of choice. 623 00:32:29,659 --> 00:32:32,929 Whereas if you've hated this podcast, please leave a five star review on 624 00:32:32,929 --> 00:32:36,379 your podcast platform of choice, along with an angry, insulting comment 625 00:32:36,439 --> 00:32:40,639 featuring an annotated copy of the re:Invent sponsorship prospectus.