WEBVTT 1 00:00:05.017 --> 00:00:09.178 In today's modern technological world, developers face many difficult questions. 2 00:00:09.538 --> 00:00:14.640 Artificial intelligence, autonomous agents, JavaScript frameworks, even CSS. 3 00:00:14.980 --> 00:00:18.841 With so much happening so quickly, how can a developer possibly keep up? 4 00:00:19.281 --> 00:00:19.741 Simple. 5 00:00:20.201 --> 00:00:23.222 Just listen to the world's most important AI podcast. 6 00:00:23.402 --> 00:00:26.583 Two experts, countless opinions, zero hesitation. 7 00:00:26.943 --> 00:00:28.563 The Internet is complicated. 8 00:00:28.964 --> 00:00:30.304 These men are not. 9 00:00:30.784 --> 00:00:35.228 Each week they review the latest developments in artificial intelligence and modern web development. 10 00:00:35.588 --> 00:00:40.713 This is Whiskey, Web and Whatnot, the world's most important AI podcast. 11 00:00:43.164 --> 00:01:03.550 This episode is sponsored by MailTrap, modern email delivery for developers, straightforward integration into your code with native SDKs, security compliant API and SMTP, 4,000 emails monthly in their free tier, MCP and API integrations for your agents and coding assistants, and PS 24-7 support you contact people, not an AI chatbot. 12 00:01:03.930 --> 00:01:05.511 Try it mailtrap.io today. 13 00:01:05.711 --> 00:01:08.271 That's M-A-I-L-T-R-A-P.io. 14 00:01:08.732 --> 00:01:10.232 And back to that episode. 15 00:01:14.353 --> 00:01:14.793 Oh, hello. 16 00:01:15.593 --> 00:01:16.354 Hey, everybody. 17 00:01:16.654 --> 00:01:22.876 Welcome to Whiskey Web and Whatnot, your favorite podcast about dryer vent cleaning. 18 00:01:23.056 --> 00:01:23.616 Of course. 19 00:01:24.076 --> 00:01:28.878 No, it's about Whiskey Web and whatnot and mostly AI, things like that. 20 00:01:29.098 --> 00:01:35.460 Anyway, your host for today, Robbie the Wagner and Adam Thomas Argyle, the nerd. 21 00:01:36.100 --> 00:01:36.580 Howdy, y'all. 22 00:01:36.800 --> 00:01:37.541 Stoked to be here. 23 00:01:37.861 --> 00:01:38.361 Hope you're ready. 24 00:01:39.141 --> 00:01:40.182 Yeah, I am ready. 25 00:01:41.203 --> 00:01:41.623 I forget. 26 00:01:41.644 --> 00:01:42.824 Did we just jump into the whiskey? 27 00:01:42.945 --> 00:01:44.486 I don't know how this thing works anymore. 28 00:01:45.226 --> 00:01:46.067 We're just going to do it. 29 00:01:46.708 --> 00:01:47.428 We can teaser. 30 00:01:47.488 --> 00:01:49.390 We can be like, we're going to talk about MySpace. 31 00:01:49.630 --> 00:01:52.292 We're going to talk about 2026 versus 2006. 32 00:01:52.572 --> 00:01:55.095 Building a mobile app with Cloud is building software with LLMs. 33 00:01:55.155 --> 00:01:56.055 Bad for your mental health. 34 00:01:56.276 --> 00:01:57.096 Agentic heartbeats. 35 00:01:57.356 --> 00:01:57.877 Prop for that. 36 00:01:58.137 --> 00:01:58.778 Meta harnesses. 37 00:01:59.058 --> 00:01:59.338 Buzz. 38 00:01:59.618 --> 00:02:00.459 Spectre-driven work trees. 39 00:02:00.719 --> 00:02:01.680 Whiskey build times. 40 00:02:01.920 --> 00:02:03.822 Yeah, whiskey as compared to built-ups. 41 00:02:03.882 --> 00:02:05.923 Anyway, and then shit that Steve Yeggie said. 42 00:02:06.164 --> 00:02:06.344 Yeah. 43 00:02:06.824 --> 00:02:07.984 Anyway, you're ready. 44 00:02:08.265 --> 00:02:12.386 That sounds like a Jeopardy category. 45 00:02:14.327 --> 00:02:14.887 All right. 46 00:02:15.247 --> 00:02:23.470 So, yes, today we have the Chicken Cock, always a favorite, small batch Kentucky Straight bourbon whiskey. 47 00:02:23.630 --> 00:02:25.210 It is batch number 17. 48 00:02:26.271 --> 00:02:27.831 There were 10 barrels. 49 00:02:27.891 --> 00:02:28.852 Am I reading that right? 50 00:02:29.092 --> 00:02:31.233 Mine's seven or 12 or 12. 51 00:02:31.553 --> 00:02:32.413 I have 12 over seven. 52 00:02:34.426 --> 00:02:35.987 Okay, so we have different ones, I guess. 53 00:02:36.007 --> 00:02:37.688 So we have different small batches. 54 00:02:38.088 --> 00:02:38.968 We'll see if they're... 55 00:02:39.188 --> 00:02:42.530 They were all hand selected by the master or whatever. 56 00:02:43.290 --> 00:02:43.570 Yes. 57 00:02:44.131 --> 00:02:45.091 All right. 58 00:02:45.751 --> 00:02:53.175 Yeah, it's hunter proof, five years old, 70% corn, 21% rye, 9% malted barley. 59 00:02:54.515 --> 00:03:00.038 And fun fact, chicken cocks, what roosters were called before the word rooster existed. 60 00:03:02.456 --> 00:03:03.016 Was that fun? 61 00:03:03.277 --> 00:03:04.057 I thought it was fun. 62 00:03:04.077 --> 00:03:05.298 I was like, what's a chicken cock? 63 00:03:05.318 --> 00:03:07.019 I mean, it just sounds unnecessarily long. 64 00:03:07.039 --> 00:03:09.121 Like it's a chicken with a cock, maybe? 65 00:03:09.661 --> 00:03:10.982 Actually, that probably is what it was, huh? 66 00:03:11.662 --> 00:03:12.203 I guess, yeah. 67 00:03:16.472 --> 00:03:16.672 Yeah. 68 00:03:16.812 --> 00:03:20.615 No, I mean, because I am forever a child, I always love to get some chicken cock. 69 00:03:20.875 --> 00:03:26.800 And if you've ever read their marketing page, it's like Chris Coyier got a good chuckle out of it. 70 00:03:26.840 --> 00:03:32.824 It's something about like the creamy, buttery mouthfeel of the chicken cock or something. 71 00:03:33.785 --> 00:03:34.065 Oh, God. 72 00:03:34.085 --> 00:03:34.865 There's so many jokes. 73 00:03:35.006 --> 00:03:35.806 Anyway. 74 00:03:35.826 --> 00:03:36.467 I hope not. 75 00:03:40.750 --> 00:03:42.111 I'm smelling... Hmm? 76 00:03:43.492 --> 00:03:43.632 Uh... 77 00:03:45.180 --> 00:03:45.561 What is it? 78 00:03:46.101 --> 00:03:46.762 Bubble yum? 79 00:03:46.862 --> 00:03:50.847 Is that the one that's like the big cubes of bubble gum that are like pink? 80 00:03:51.608 --> 00:03:52.329 Is that bubble yum? 81 00:03:52.349 --> 00:03:52.549 Yep. 82 00:03:53.370 --> 00:03:53.511 Yeah. 83 00:03:53.531 --> 00:03:56.174 The ones that you chew on for like five seconds and then they're out of flavor. 84 00:03:56.634 --> 00:03:57.636 That's like the zebra gum. 85 00:03:58.176 --> 00:03:58.637 Just like the. 86 00:03:58.657 --> 00:03:59.138 The zebra gum is gone? 87 00:03:59.158 --> 00:03:59.398 It's gone? 88 00:04:00.579 --> 00:04:00.940 It's gone? 89 00:04:01.363 --> 00:04:01.883 Yeah, it's gone. 90 00:04:02.224 --> 00:04:03.545 It went away like last year, I think. 91 00:04:04.526 --> 00:04:05.387 Whoa. 92 00:04:05.447 --> 00:04:07.649 After like 70 years or something. 93 00:04:08.149 --> 00:04:09.951 No, I think it was really bad gum. 94 00:04:10.912 --> 00:04:12.333 No, it was really delicious gum. 95 00:04:12.573 --> 00:04:14.495 It's just the flavor went away in one second. 96 00:04:14.795 --> 00:04:15.035 Yeah. 97 00:04:17.297 --> 00:04:18.318 It always looked great, too. 98 00:04:18.338 --> 00:04:19.519 You're like, ooh, I can't wait. 99 00:04:20.721 --> 00:04:21.962 Okay, so yeah, you're getting bubble gum. 100 00:04:22.382 --> 00:04:23.423 I'm definitely getting cinnamon. 101 00:04:24.691 --> 00:04:25.552 Getting some almonds. 102 00:04:26.273 --> 00:04:28.875 The smell on the extract, perhaps strong, dude. 103 00:04:29.536 --> 00:04:31.157 Maybe I'm smelling from the wrong side of the glass. 104 00:04:31.197 --> 00:04:33.920 So top or bottom, which one's supposed to put your whole face in it. 105 00:04:33.940 --> 00:04:34.100 Right. 106 00:04:34.300 --> 00:04:36.242 And get underneath. 107 00:04:38.104 --> 00:04:39.185 Hmm. 108 00:04:40.346 --> 00:04:41.166 I think I did it wrong. 109 00:04:43.789 --> 00:04:44.650 It is really strong. 110 00:04:45.070 --> 00:04:47.072 I do think smelling from the top is better than the bottom. 111 00:04:48.923 --> 00:04:49.644 Less fumey. 112 00:04:50.004 --> 00:04:56.470 I thought Denver said the other way, but I think he did say the other way, but I haven't been able to successfully do it the other way. 113 00:04:56.691 --> 00:04:57.531 I keep trying it. 114 00:04:57.631 --> 00:05:01.315 And definitely one is fumer has more fumes than the other. 115 00:05:01.495 --> 00:05:03.938 So I think I'm going to start from the top. 116 00:05:03.998 --> 00:05:04.278 I don't know. 117 00:05:06.118 --> 00:05:06.698 It's so weird. 118 00:05:06.718 --> 00:05:08.899 My computer, my work machine is just over here open. 119 00:05:09.399 --> 00:05:10.180 It's just doing stuff. 120 00:05:10.920 --> 00:05:11.800 It's so distracting. 121 00:05:11.921 --> 00:05:15.262 So if you see me looking over here, it's because my computer is doing stuff. 122 00:05:15.282 --> 00:05:15.922 I'm like, what are you doing? 123 00:05:15.942 --> 00:05:16.362 I'm not there. 124 00:05:16.482 --> 00:05:17.603 Oh yeah, I told you to do that. 125 00:05:19.244 --> 00:05:19.424 Nice. 126 00:05:20.544 --> 00:05:24.026 It tastes pretty bourbon-y with a little bit of spice. 127 00:05:25.346 --> 00:05:27.147 Like if you mix a little bit of rye into a bourbon. 128 00:05:28.027 --> 00:05:28.908 Gotta have that spice. 129 00:05:30.929 --> 00:05:34.050 It's sweeter and it does have a texture that is... 130 00:05:35.764 --> 00:05:37.105 Maybe more syrpy. 131 00:05:37.145 --> 00:05:40.148 Yeah, it's a little thicker, which I like. 132 00:05:41.529 --> 00:05:41.829 All right. 133 00:05:41.849 --> 00:05:43.231 Well, how much do you like it? 134 00:05:43.731 --> 00:05:45.132 Zero to eight tentacles. 135 00:05:45.513 --> 00:05:46.353 Zero being the worst. 136 00:05:46.434 --> 00:05:47.234 Eight being the best. 137 00:05:47.975 --> 00:05:48.355 What do you think? 138 00:05:49.196 --> 00:05:49.656 Oh, man. 139 00:05:49.696 --> 00:05:51.898 Compared to other bourbons, it's better. 140 00:05:52.319 --> 00:05:53.480 Do I like it that much? 141 00:05:53.960 --> 00:05:54.601 Not really. 142 00:05:55.161 --> 00:05:58.124 It's like a bourbon, but it's got the spice. 143 00:05:58.284 --> 00:05:59.945 So it's got a few things going for it. 144 00:06:00.566 --> 00:06:00.986 The spice. 145 00:06:01.187 --> 00:06:01.667 Got to have it. 146 00:06:01.767 --> 00:06:03.369 If it's too syrupy, I really dislike it. 147 00:06:04.150 --> 00:06:09.455 It's a little bit thicker, but it is quite pungent on the smell unless you approach it from the top. 148 00:06:09.675 --> 00:06:11.617 I'm going to give it a six. 149 00:06:13.459 --> 00:06:13.820 All right. 150 00:06:14.741 --> 00:06:19.866 Yeah, it's hard for me to compare to a straight bourbon because, like, it's a little spicier. 151 00:06:21.368 --> 00:06:22.289 It still has the sweetness. 152 00:06:23.660 --> 00:06:28.223 For something that's not like just a sweetness bomb, I think it's okay. 153 00:06:28.763 --> 00:06:30.624 But I'm just going to say a five, I think. 154 00:06:30.644 --> 00:06:32.686 I don't know what I'm trying to talk myself into a number here. 155 00:06:32.746 --> 00:06:33.246 I'll go lower. 156 00:06:33.386 --> 00:06:34.387 No, dude, seriously. 157 00:06:34.687 --> 00:06:35.147 But I don't know. 158 00:06:35.187 --> 00:06:36.848 I thought you liked bourbons more, you know? 159 00:06:37.769 --> 00:06:39.010 No, I like rye the best. 160 00:06:39.030 --> 00:06:39.530 Okay. 161 00:06:39.990 --> 00:06:40.270 Okay. 162 00:06:41.371 --> 00:06:47.155 But yeah, bourbons, there are a few bourbons that I like a lot and I don't know exactly how to describe them. 163 00:06:47.175 --> 00:06:48.456 One of them is the Sweetens Cove. 164 00:06:48.516 --> 00:06:52.659 It's like almost like it's a little bit camp fiery. 165 00:06:52.679 --> 00:06:54.360 Like it's almost s'mores tasting. 166 00:06:54.380 --> 00:06:55.941 Like there's marshmallow, there's smoke. 167 00:06:56.141 --> 00:06:57.442 Like it's good. 168 00:06:58.243 --> 00:07:00.124 It's not like just Jack Daniels. 169 00:07:00.164 --> 00:07:01.545 It's not like that, you know? 170 00:07:01.745 --> 00:07:02.325 Yeah. 171 00:07:02.966 --> 00:07:04.006 But yeah, no, this is not bad. 172 00:07:04.267 --> 00:07:06.668 Would recommend if anybody was looking for one. 173 00:07:06.688 --> 00:07:08.890 But yeah, pretty middle of the road for me. 174 00:07:09.950 --> 00:07:12.252 pretty middle of the road indeed. 175 00:07:13.592 --> 00:07:16.194 I think it's just my palate's not refined enough yet for some of these. 176 00:07:16.254 --> 00:07:22.757 Like I tend to like rye better also, but even sometimes with ryes recently, I'm just like, I want my peated scotch back. 177 00:07:22.857 --> 00:07:26.879 You know, I'm just like one I love and the other ones I'm like, oh, they're fine. 178 00:07:27.540 --> 00:07:27.740 Yeah. 179 00:07:27.780 --> 00:07:29.481 Yeah, I get it. 180 00:07:30.081 --> 00:07:37.165 Well, if you want something to do with ones you don't like, we've had this beer in our fridge for like, I don't know, two years now. 181 00:07:37.687 --> 00:07:39.449 And I was like, no one will drink it. 182 00:07:39.469 --> 00:07:40.349 I kept trying to give it away. 183 00:07:40.369 --> 00:07:41.470 And I was like, I don't want it. 184 00:07:41.971 --> 00:07:47.095 And then my parents were over and we got some steaks and I had nothing to marinate them in. 185 00:07:47.195 --> 00:07:49.557 So my mom was like, just use some beer. 186 00:07:49.597 --> 00:07:52.179 And I was like, hey, I know just the beer I can get rid of. 187 00:07:52.820 --> 00:07:53.220 Nice. 188 00:07:53.340 --> 00:07:53.821 And so I did. 189 00:07:53.961 --> 00:07:57.764 And yeah, so you could use it for like a sauce or a rub or something. 190 00:07:59.165 --> 00:08:00.026 Well, I mean, I kind of. 191 00:08:00.404 --> 00:08:19.112 continue drinking like i sip them throughout you know the weeks and i just want to acquire the flavors in here like i'm it's the same reason i had i had pappy's last month i think we talked about it and i was like whoop-de-doo whoop-de-doo but also what am i missing like i'm i'm i'm feeling fomo why why don't i curious yeah i'm curious 192 00:08:21.812 --> 00:08:22.192 FOMO. 193 00:08:22.352 --> 00:08:23.833 Lots of FOMO with AI right now. 194 00:08:23.853 --> 00:08:26.315 I don't want to pivot it like that, but it's really hard. 195 00:08:26.355 --> 00:08:27.896 You know, you're just like, oh, what tool? 196 00:08:27.956 --> 00:08:29.117 Everyone's not using that. 197 00:08:29.137 --> 00:08:30.018 And now we're using this. 198 00:08:30.058 --> 00:08:31.839 So I'm on Herder now. 199 00:08:32.019 --> 00:08:33.640 Herder with a hunk. 200 00:08:33.900 --> 00:08:34.160 Nice. 201 00:08:34.281 --> 00:08:35.001 You know, okay. 202 00:08:35.021 --> 00:08:36.602 But it's still not as good as Warp. 203 00:08:36.702 --> 00:08:37.803 I'm like, Warp is great. 204 00:08:38.543 --> 00:08:39.544 I installed Herder. 205 00:08:40.165 --> 00:08:41.746 Like the thing that I'm wanting... 206 00:08:42.466 --> 00:08:45.647 Which the problem is I have too many things I want to build, right? 207 00:08:46.307 --> 00:08:48.147 Lots of personal stuff, lots of work stuff. 208 00:08:50.468 --> 00:08:56.750 But like obviously work would probably frown upon me using like I haven't looked into how Herder does it. 209 00:08:56.790 --> 00:09:02.091 But like, you know, a thing that opens up all your ports to continue all your stuff on your phone and whatever. 210 00:09:02.111 --> 00:09:02.211 Yeah. 211 00:09:02.991 --> 00:09:07.673 Um, like that's what I want because I want to be able to step away, like do something else for 10 minutes. 212 00:09:08.253 --> 00:09:09.994 And if it needs help, it can let me know. 213 00:09:10.014 --> 00:09:13.736 Um, but I don't have any good solution for that at work. 214 00:09:13.756 --> 00:09:19.218 So because most of my work is at work, I don't try as many fun tools as much. 215 00:09:19.338 --> 00:09:20.318 So I installed Herder. 216 00:09:20.378 --> 00:09:24.560 I ran like a couple of tabs in it just to see what it would do, but it seems fine. 217 00:09:24.600 --> 00:09:26.081 But like, I haven't put it through its paces. 218 00:09:27.181 --> 00:09:28.081 It's not beautiful. 219 00:09:28.301 --> 00:09:28.821 Like Buzz. 220 00:09:29.021 --> 00:09:30.402 Buzz is really nice. 221 00:09:30.742 --> 00:09:31.802 Pretty much just everything that it does. 222 00:09:31.822 --> 00:09:33.842 I did not install Buzz for the same reason. 223 00:09:33.862 --> 00:09:36.923 I was like, this sounds cool, but I can't use it at work, I'm sure. 224 00:09:37.083 --> 00:09:40.583 So, yeah, I was like, come on, Shopify, let's set up Buzz. 225 00:09:40.623 --> 00:09:42.504 And I was like, actually, we kind of have our own thing. 226 00:09:42.544 --> 00:09:44.024 So why would we go adopt theirs? 227 00:09:44.104 --> 00:09:47.905 But and then I was like, I'll set it up on my cloud machine and just try it. 228 00:09:47.945 --> 00:09:49.885 And then I was like, it needs too many keys. 229 00:09:50.605 --> 00:09:51.245 I'm lazy. 230 00:09:51.465 --> 00:09:53.706 I don't like going in and putting keys in places. 231 00:09:53.746 --> 00:09:55.206 I'm like, what I have ain't broken anyway. 232 00:09:55.226 --> 00:09:55.386 So. 233 00:09:56.546 --> 00:10:06.613 I've got such a straight relationship though of like, I like the idea of, you know, let's replace Slack with a place that you can talk to agents and people in the same area. 234 00:10:07.433 --> 00:10:12.056 And like, um, I guess, does it also do like task management type stuff? 235 00:10:12.076 --> 00:10:13.797 Like, is it supposed to replace Jira too? 236 00:10:13.917 --> 00:10:16.819 Or is it not that I don't think it replaces Jira. 237 00:10:17.119 --> 00:10:17.300 No. 238 00:10:17.560 --> 00:10:17.860 Okay. 239 00:10:18.040 --> 00:10:21.841 Cause that's what I would want is an all in one plan everything for me. 240 00:10:21.921 --> 00:10:27.762 I give you a few bullets, you make me a ticket and then like, yeah, I can give it to an agent. 241 00:10:27.882 --> 00:10:28.882 I can give it to a person. 242 00:10:28.902 --> 00:10:33.843 We can, you know, collaboratively plan the ticket even like through an agent together in the same chat. 243 00:10:33.903 --> 00:10:42.545 Like I think there's a lot there, like good ideas that could, could work, but you know, whether that this is going to be the product that people actually use for that sort of thing. 244 00:10:42.725 --> 00:10:44.846 I don't know about that, but I like the ideas. 245 00:10:44.866 --> 00:10:45.226 Yeah. 246 00:10:46.047 --> 00:10:47.588 So yeah, it's funny you mentioned the ticket sets. 247 00:10:47.708 --> 00:10:51.792 That was like, we had the Shopify hackathon a couple of weeks ago called Summit. 248 00:10:52.253 --> 00:10:54.916 I finally got to go meet coworkers I'd known for a year. 249 00:10:55.736 --> 00:10:58.199 And I spun up many explorations. 250 00:10:58.339 --> 00:11:01.842 And one of them was because we have Slack, we have our own agents we can talk to in Slack. 251 00:11:03.084 --> 00:11:08.850 Working together in a thread is the closest thing to multiplayer that we have right now with agent workflows like Slack. 252 00:11:09.430 --> 00:11:14.491 Unless you're in GitHub, we've done it in GitHub a little bit where we're both kind of pinging the agent and telling it what to do. 253 00:11:16.132 --> 00:11:22.213 But I'm really interested in multiplayer experiences with this agent work because we're better together. 254 00:11:22.633 --> 00:11:27.734 And then we don't have to do so much ping ponging like, here, you put your agent to talk to this PR. 255 00:11:27.754 --> 00:11:31.255 Oh, and that'll take my... Man, let's just do it somewhere else and get it all done earlier. 256 00:11:31.275 --> 00:11:31.435 Yeah. 257 00:11:32.315 --> 00:11:39.740 And so I built something into our Slack that said, hey, every time I spin up a thread and I give it this special tag, go put it on a project board. 258 00:11:40.160 --> 00:11:46.304 And the project board is an internal hosted site that has auth, it has all these really nice conveniences already set up for all the people at Shopify. 259 00:11:47.244 --> 00:11:55.712 And then once it's there, people can either go there to see the things I'm building with those agents and then click a button and go to the thread and participate. 260 00:11:56.473 --> 00:12:00.817 Or they can be in the thread, initialize a ticket that goes and gets put up on the board. 261 00:12:01.097 --> 00:12:06.662 And then the agent continuously writes to that ticket, keeping it up to date and treats it like a spec. 262 00:12:07.142 --> 00:12:07.803 You knew it was coming. 263 00:12:08.003 --> 00:12:12.326 So it's like this spec that stays alive that orients all the sub agents that need to get in there. 264 00:12:12.486 --> 00:12:14.407 If you want to come and see like, what are we working on? 265 00:12:14.427 --> 00:12:16.548 What are the requirements of this work tree it's trying to build? 266 00:12:16.949 --> 00:12:22.212 It's all there in a nice little ticket and it moves along the swim lanes as you kind of work along with it. 267 00:12:22.767 --> 00:12:23.407 It's a cool idea. 268 00:12:24.487 --> 00:12:25.248 Are people using it? 269 00:12:25.688 --> 00:12:25.908 No. 270 00:12:26.708 --> 00:12:28.609 Everybody's kind of building their own things right now. 271 00:12:28.729 --> 00:12:37.811 There must be like 10 or 20 different experiences people are building trying to figure out what the next-gen agentic workflows are going to be. 272 00:12:38.011 --> 00:12:38.512 It's crazy. 273 00:12:38.692 --> 00:12:38.932 Yeah. 274 00:12:39.732 --> 00:12:43.233 Yeah, we have a tool that we use internally that like... 275 00:12:45.034 --> 00:12:48.578 you basically tell it, you know, I want to build X, Y, Z. 276 00:12:48.618 --> 00:12:49.679 Here's my requirements. 277 00:12:49.959 --> 00:12:57.767 And it goes like, huh, that seems like, you know, six tasks that I want to give to sub agents and have a work tree for each. 278 00:12:58.447 --> 00:13:02.791 And it like creates like tasks, like, like a super light Jira ticket. 279 00:13:03.432 --> 00:13:04.233 But it's not in JIRA. 280 00:13:04.253 --> 00:13:06.695 It's just like a, you know, little description for each thing. 281 00:13:07.276 --> 00:13:15.744 And then like the idea is anyone could pick any of those up or any agent could pick any of those up and they can then know like they're not going to duplicate the same work. 282 00:13:17.045 --> 00:13:22.791 But yeah, I don't I don't ever go look at like the tickets it makes or like but I like that it does all that for me. 283 00:13:22.911 --> 00:13:26.074 Like I feel like this is another implementation detail moment of like. 284 00:13:26.514 --> 00:13:30.999 The tooling that continues to just get all these things right is going to like do it really well. 285 00:13:31.019 --> 00:13:43.533 So this brings me to I've been working on a mobile app because I saw a tweet where Claude was like, you can use the iOS simulator and the Claude desktop app now. 286 00:13:43.593 --> 00:13:45.095 And I was like, sick, let's do it. 287 00:13:45.755 --> 00:13:47.356 So you do it. 288 00:13:47.416 --> 00:13:51.017 It just spins it up, shows you exactly what it's doing, like clicks through all of it. 289 00:13:51.137 --> 00:13:52.297 You can just iterate right there. 290 00:13:53.238 --> 00:13:55.538 I hadn't used the Claude desktop app before. 291 00:13:56.479 --> 00:13:58.899 And maybe Claude code will also do this. 292 00:13:59.780 --> 00:14:00.260 But like... 293 00:14:00.753 --> 00:14:02.154 Yeah, why not just use the simulator? 294 00:14:02.194 --> 00:14:02.834 I do that all the time. 295 00:14:02.854 --> 00:14:03.695 I have it up right now. 296 00:14:03.715 --> 00:14:05.016 Because it just does it for me. 297 00:14:05.876 --> 00:14:11.920 Like, I don't need to know how to go into Xcode and hit play and which things do I want to play and how do I want to build it. 298 00:14:11.960 --> 00:14:13.361 Oh, you have an Xcode app you want to build. 299 00:14:13.481 --> 00:14:14.602 Oh, see, I'm in mobile Safari. 300 00:14:14.822 --> 00:14:15.923 Okay, so that's a little different. 301 00:14:16.043 --> 00:14:16.223 Yeah. 302 00:14:16.463 --> 00:14:18.124 Yeah, so I'm building a... 303 00:14:18.625 --> 00:14:20.086 I'll talk about the app. 304 00:14:20.526 --> 00:14:22.527 I haven't released it yet, but it's... 305 00:14:23.028 --> 00:14:24.869 So the idea is I have like... 306 00:14:26.177 --> 00:14:28.299 not diagnosed, but I feel like I have ADHD. 307 00:14:28.359 --> 00:14:30.961 Like I come in a room, my wife tells me to get two things. 308 00:14:31.022 --> 00:14:33.404 I get down to the room and I don't remember either of the things. 309 00:14:33.424 --> 00:14:35.205 I come back with like one unrelated thing. 310 00:14:36.026 --> 00:14:37.467 So I'm looking to stop that. 311 00:14:37.907 --> 00:14:40.229 So I'm making an app where you just talk into your watch. 312 00:14:40.250 --> 00:14:43.452 You're like, hey, I'm supposed to go get these two things. 313 00:14:44.073 --> 00:14:48.517 And then on random intervals, like 30 seconds, a minute, 314 00:14:50.138 --> 00:14:51.440 a minute and 15 seconds. 315 00:14:51.460 --> 00:14:54.283 Like it has a weird curve of just like, did you get that thing yet? 316 00:14:54.744 --> 00:14:55.185 Did you get it? 317 00:14:55.225 --> 00:14:57.628 It goes, it goes, Hey, you still getting that thing? 318 00:14:58.088 --> 00:14:59.170 And it mess. 319 00:14:59.250 --> 00:15:03.816 It like does different haptics instead of just like a normal notification each time that you might ignore. 320 00:15:03.976 --> 00:15:06.259 It like changes it up and like does different stuff. 321 00:15:06.319 --> 00:15:06.599 And like, 322 00:15:07.280 --> 00:15:13.846 The idea is just like pinging you in different weird ways to make sure you're still doing the thing you were supposed to be doing and didn't get off track. 323 00:15:14.707 --> 00:15:17.250 And so I feel like that's like slightly different. 324 00:15:17.290 --> 00:15:18.991 It doesn't have like a backlog. 325 00:15:19.211 --> 00:15:21.434 It doesn't have like like it's not a reminders app. 326 00:15:21.674 --> 00:15:23.375 It's just one thing you're doing right now. 327 00:15:23.475 --> 00:15:24.837 It's like kind of in its own lane. 328 00:15:26.759 --> 00:15:28.921 Whether it's something other people will find useful, I don't know. 329 00:15:29.001 --> 00:15:31.343 But I'm going to release it at some point just as an experiment. 330 00:15:32.384 --> 00:15:32.664 Cool. 331 00:15:32.924 --> 00:15:33.804 Yeah, I like the idea. 332 00:15:33.824 --> 00:15:39.686 I mean, I think it's very, very relatable to have that experience where you go downstairs and you're like, oh, crap. 333 00:15:40.526 --> 00:15:41.447 What was I down here for? 334 00:15:41.467 --> 00:15:44.628 It sounds like fun to make a little Xcode app. 335 00:15:44.848 --> 00:15:47.829 So it's what does the mobile app do? 336 00:15:47.929 --> 00:15:51.750 It's just sort of the data center and then your watch is the interface primarily. 337 00:15:51.770 --> 00:15:52.250 Yeah. 338 00:15:52.390 --> 00:15:52.770 Yeah. 339 00:15:52.810 --> 00:16:04.233 So the, I, I don't, it's fuzzy to me how this stuff works, but I think the iOS app is the app and then it has a watch component that you can like, it's also an app. 340 00:16:04.273 --> 00:16:06.794 I guess it's a watchOS app, but like screen ish. 341 00:16:06.994 --> 00:16:07.254 Yeah. 342 00:16:07.394 --> 00:16:09.775 I think you would have, I think you have to have the phone app. 343 00:16:10.275 --> 00:16:12.335 It's like, you can't just run it on your watch. 344 00:16:13.416 --> 00:16:19.757 Like if, but I guess you, why would you have an, sorry, why would you have an Apple watch if you didn't have an iPhone? 345 00:16:19.797 --> 00:16:21.458 I guess that'd be like kind of pointless. 346 00:16:22.318 --> 00:16:28.946 Um, yes, but yeah, so it's the, the intended flow is to like, cause I want to do it quickly. 347 00:16:28.966 --> 00:16:30.047 I don't want to have to get my phone out. 348 00:16:30.067 --> 00:16:31.729 So I just want to like tell my watch. 349 00:16:31.749 --> 00:16:32.490 That's what I want to do. 350 00:16:32.510 --> 00:16:36.755 I have like a little, like, I can't, I can't show you probably. 351 00:16:38.577 --> 00:16:38.817 I see. 352 00:16:38.837 --> 00:16:40.199 No, there's some circles. 353 00:16:40.359 --> 00:16:40.479 Yeah. 354 00:16:41.234 --> 00:16:44.599 Yeah, no, above the circles is like the big middle section. 355 00:16:44.660 --> 00:16:50.990 The app's called Doorway, and it's like, it says Doorway, and it'll have your list of things you were supposed to be getting there. 356 00:16:51.570 --> 00:16:53.974 So it's like quick glance, like you can see what you're supposed to be doing. 357 00:16:56.216 --> 00:16:58.317 But yeah, I think the main app is in iOS. 358 00:16:58.617 --> 00:17:00.177 It's watchOS as a companion. 359 00:17:00.197 --> 00:17:02.318 Then it also syncs to the macOS app. 360 00:17:03.198 --> 00:17:08.620 And like, because that isn't part of the same app, but there's like a cloud kit thing that will sync the data. 361 00:17:09.500 --> 00:17:09.760 Nice. 362 00:17:10.440 --> 00:17:14.041 And then, so like all that's fine and good and useful. 363 00:17:14.261 --> 00:17:20.003 And then the thing that like, if it were to ever make money, the thing I would charge for is like, hey, 364 00:17:20.943 --> 00:17:28.626 Skip the step where my spouse tells me what to go get and just let them put into my app via their phone what I'm supposed to go get. 365 00:17:29.446 --> 00:17:31.406 Like, like sharing the list kind of thing. 366 00:17:33.487 --> 00:17:38.829 So like a couple little like, you know, nice to have additional features would be the like paid part, but. 367 00:17:39.710 --> 00:17:40.250 Yeah, we'll see. 368 00:17:41.131 --> 00:17:45.736 It's currently like it doesn't delight me in terms of the look. 369 00:17:45.756 --> 00:17:46.616 It doesn't look bad. 370 00:17:46.857 --> 00:17:51.341 It looks like, you know, an iOS app, but I don't know how to iterate on it. 371 00:17:51.361 --> 00:17:56.186 Like you don't want it to look not like an iOS app and it's all Swift code, so you can't do it so much. 372 00:17:56.206 --> 00:17:57.227 Yeah. 373 00:17:58.448 --> 00:18:02.311 But like, yeah, it's it's just not it doesn't have that it factor. 374 00:18:02.351 --> 00:18:12.058 Like, you know, there's been a few I can't even think of one off the top of my head, but there's been a few times there's been an app that comes out and everybody's just like, wow, like that's like totally different, really well designed. 375 00:18:14.240 --> 00:18:20.265 And yeah, I'm not getting that those vibes, but I've been working with impeccable trying to like make it better. 376 00:18:21.766 --> 00:18:23.087 So we'll see where that goes. 377 00:18:23.107 --> 00:18:23.207 Yeah. 378 00:18:24.745 --> 00:18:26.606 Yep, that struggle is so classic. 379 00:18:26.626 --> 00:18:31.268 I've been talking about it for years where you've got the web where your default components, they're ass. 380 00:18:31.829 --> 00:18:32.529 They're really bad. 381 00:18:33.009 --> 00:18:37.431 I mean, they look 30 years old and you really got to do a lot of work to build them up into something beautiful. 382 00:18:37.472 --> 00:18:47.937 So you see something that J builds or the nice select or the nice accordion things that I've made like there's a lot of styles that you have to put on top of that kind of crappy garbage base. 383 00:18:48.677 --> 00:18:54.698 However, it's a very functional base and all your problems become yours as you make it more robust. 384 00:18:55.078 --> 00:18:58.899 And then the native apps have the complete opposite problem. 385 00:18:58.939 --> 00:19:01.519 Their components come out of the box like decent. 386 00:19:02.080 --> 00:19:06.940 And so then to do anything custom, you have to strip them down and you have to take things away. 387 00:19:07.441 --> 00:19:10.641 So it's an additive versus subtractive kind of design practice. 388 00:19:10.721 --> 00:19:15.302 And so you either have apps that look just like out of the box. 389 00:19:15.542 --> 00:19:17.603 and they're boring, but they're fine. 390 00:19:18.523 --> 00:19:24.124 But they're also better than the ones that come out of the box on the web or you do out of the box on the web and everyone goes, my God, it's ugly. 391 00:19:24.384 --> 00:19:25.605 And then you have to sit there and build it up. 392 00:19:25.985 --> 00:19:26.825 It's just a tension. 393 00:19:27.665 --> 00:19:31.266 I don't know what the solution is other than pick your poison. 394 00:19:31.286 --> 00:19:32.567 They both kind of have their issue. 395 00:19:32.587 --> 00:19:36.248 If you wanna go off the rails and do something creative, you have to go off the rails and do something creative. 396 00:19:36.668 --> 00:19:42.850 And that means taking upon yourself design choices and the whole shebang just starts to fall on you. 397 00:19:44.230 --> 00:19:45.555 Well, you know what I think is going to help a lot? 398 00:19:46.921 --> 00:19:47.924 MySpace coming back. 399 00:19:48.654 --> 00:19:49.194 Oh my goodness. 400 00:19:49.214 --> 00:19:50.094 Tell me about what this is. 401 00:19:50.154 --> 00:19:52.755 Like what the music side or the side where you have a profile? 402 00:19:52.795 --> 00:19:53.695 Like which part of MySpace is coming from? 403 00:19:53.715 --> 00:19:54.295 No, the whole thing. 404 00:19:54.315 --> 00:19:56.216 So the music side is still there. 405 00:19:57.036 --> 00:20:03.738 Kind of like every once in a while I go check out like old band pages or whatever, seeing if any of the music will play. 406 00:20:04.658 --> 00:20:05.438 The songs are there. 407 00:20:05.478 --> 00:20:06.498 They tell you how long they are. 408 00:20:06.538 --> 00:20:08.379 The play button works and nothing ever plays. 409 00:20:08.399 --> 00:20:12.120 Like it's like they lost all the media in the like migration to the new system or whatever. 410 00:20:14.000 --> 00:20:16.260 So I don't know if they have a way to get all of that back. 411 00:20:16.481 --> 00:20:17.181 That would be cool. 412 00:20:18.981 --> 00:20:22.063 But no, they're like, hey, social media sucks right now. 413 00:20:22.623 --> 00:20:23.684 You know, it's such a like. 414 00:20:25.685 --> 00:20:26.066 I don't know. 415 00:20:26.126 --> 00:20:27.606 It's just toxic in general. 416 00:20:27.626 --> 00:20:31.249 Who's they when they're saying this, like people are in charge of my space. 417 00:20:32.089 --> 00:20:32.349 OK. 418 00:20:32.369 --> 00:20:34.771 The people that own my space right now. 419 00:20:34.991 --> 00:20:40.394 And they're like, well, this is like the perfect time to, you know, re-release my space. 420 00:20:40.454 --> 00:20:41.335 Basically, like, let's put it. 421 00:20:41.655 --> 00:20:44.577 I don't know if it's going to be the exact same or if they're going to like. 422 00:20:45.297 --> 00:20:46.498 Yeah, I'm sure it'll be a little different. 423 00:20:46.558 --> 00:20:46.798 Yeah. 424 00:20:47.318 --> 00:21:01.122 Um, but like if you could do everything that you could do before, make your own layouts, like play with HTML and CSS and stuff like I feel like that would be really, really cool and give people a spot to like experiment with. 425 00:21:01.222 --> 00:21:06.184 I guess with AI, there's a lot of vulnerabilities you could exploit really fast if you allow that. 426 00:21:06.304 --> 00:21:06.784 But, um. 427 00:21:07.484 --> 00:21:09.546 I mean, let people have AI in there. 428 00:21:09.586 --> 00:21:13.849 Like, give them impeccables built in and customize your page with an agent. 429 00:21:14.149 --> 00:21:17.271 Like, maybe that's the new thing. 430 00:21:18.052 --> 00:21:19.092 I can see that being kind of fun. 431 00:21:19.673 --> 00:21:23.956 The one thing that they don't have going for them is that, like, everyone's on their phone now. 432 00:21:24.576 --> 00:21:28.599 Like, the MySpace experience was very computer heavy, like... 433 00:21:29.299 --> 00:21:46.652 you know it i don't know if i ever loaded it on a phone i guess it's probably dead by the time i had a mobile browser yeah it was i don't i've never pulled it up on a phone either what would it even do because yeah i don't know i was just wondering if it was responsive at all it was probably during the era of like if it did load it was like mobile that myspace.com it was like a 434 00:21:51.535 --> 00:21:54.037 But yeah, so I don't know how they're going to address that gap. 435 00:21:54.217 --> 00:22:00.542 I think there'll be a lot of people, you know, joining for the nostalgia and like seeing what they can build for fun for at least a few weeks. 436 00:22:01.322 --> 00:22:07.007 But like if you want to be a real social media type of thing, then I think you've got to... 437 00:22:07.747 --> 00:22:25.693 really think that out better about like how do you capture people on their phones also or maybe you don't want to maybe that's the point maybe you want people to get off their phones i don't know i haven't read into like what their plans are really i just saw like an email from it was like a linkedin thing where a bunch of people on linkedin were talking about myspace coming back so 438 00:22:27.958 --> 00:22:39.924 It cracks me up that LinkedIn is like a social network people are at, but hey, people are, they seem to be going all over the place and I'm the loser that's just sort of still on my own website. 439 00:22:40.004 --> 00:22:42.186 Although I like having my own website and syndicating out. 440 00:22:42.226 --> 00:22:47.688 I still think that's a superior strategy that outlives some of these other things, but it is a lot more work. 441 00:22:48.289 --> 00:22:49.049 And I've thought about that too. 442 00:22:49.069 --> 00:22:51.010 Like how do I make it so that everybody can have a site like mine? 443 00:22:52.031 --> 00:22:52.771 Would they even use it? 444 00:22:53.071 --> 00:22:53.392 I don't know. 445 00:22:53.712 --> 00:22:55.833 It's hard to get people to move and do things like that. 446 00:22:55.873 --> 00:22:56.093 Yeah. 447 00:22:57.493 --> 00:22:57.653 Yeah. 448 00:22:57.673 --> 00:22:59.074 I mean, it depends on your goals. 449 00:22:59.234 --> 00:23:04.757 Like, you know, your own website is good for SEO usually, but what if it's not? 450 00:23:04.777 --> 00:23:05.938 It depends on what you're doing. 451 00:23:05.958 --> 00:23:13.902 Like if you had a video on TikTok that like had certain keywords you wanted to blow up and it blew up, then like maybe that's better for SEO than your personal website would be. 452 00:23:13.962 --> 00:23:14.282 I don't know. 453 00:23:15.228 --> 00:23:15.408 Yeah. 454 00:23:16.368 --> 00:23:16.529 Yeah. 455 00:23:16.549 --> 00:23:18.589 Blue Sky is trying to help out with the publications. 456 00:23:18.709 --> 00:23:21.190 And I mean, there's been a lot of attempts to try to help discovery. 457 00:23:21.210 --> 00:23:29.033 We had web rings way back in the day that was just like, yeah, if you're in some sort of like channel or a genre, get in on the web ring and people are just surfing to you. 458 00:23:29.914 --> 00:23:30.314 I'm curious. 459 00:23:30.334 --> 00:23:35.016 So like, what did you mean here in 2026 is just 2006, but more expensive? 460 00:23:35.056 --> 00:23:35.536 I mean, like... 461 00:23:36.516 --> 00:23:37.196 In a lot of ways. 462 00:23:37.216 --> 00:23:38.077 It is way more expensive. 463 00:23:38.757 --> 00:23:39.617 Yeah. 464 00:23:40.037 --> 00:23:44.519 I mean, the thing that sparked this for me that I hadn't really paid attention because I've just been busy. 465 00:23:44.699 --> 00:23:47.140 But someone took a video. 466 00:23:47.160 --> 00:23:50.301 They walked into PacSun or some store like that. 467 00:23:50.961 --> 00:23:53.802 And we're just videoing all the clothes and showing you. 468 00:23:54.082 --> 00:23:54.682 And I was like... 469 00:23:55.102 --> 00:23:58.963 They were like, hey, this is like, it was like everything everyone was wearing. 470 00:23:58.983 --> 00:24:00.683 And the clothes are totally back. 471 00:24:00.903 --> 00:24:03.384 And Korn and Rammstein and Rob Zombie. 472 00:24:03.424 --> 00:24:04.964 These are all like Cooley. 473 00:24:05.264 --> 00:24:05.744 Slipknot. 474 00:24:05.864 --> 00:24:07.465 I see all these kids wearing Slipknot shirts. 475 00:24:07.485 --> 00:24:08.005 I'm like, what? 476 00:24:08.105 --> 00:24:09.425 Basement shows again. 477 00:24:09.465 --> 00:24:11.225 And like, like the music scene is big again. 478 00:24:11.245 --> 00:24:13.766 I'm like, are we back to like, you know, peak? 479 00:24:14.406 --> 00:24:16.926 We don't have to listen to like only Taylor Swift anymore. 480 00:24:17.046 --> 00:24:20.067 Or like, you know, we can have some guitars in a song. 481 00:24:20.127 --> 00:24:22.127 Like, I don't know. 482 00:24:25.168 --> 00:24:29.491 you know, people actually having some original thoughts and like doing music for fun. 483 00:24:29.551 --> 00:24:34.514 And yeah, another thing people are doing is they want their video games. 484 00:24:35.394 --> 00:24:36.015 They want to own it. 485 00:24:36.355 --> 00:24:38.156 They're tired of digital games. 486 00:24:38.556 --> 00:24:40.157 And so you have people going way back. 487 00:24:40.498 --> 00:24:43.019 Like retro gaming is on a big swing right now. 488 00:24:43.079 --> 00:24:47.602 People want a local Nintendo system that doesn't have to connect to internet. 489 00:24:47.622 --> 00:24:48.503 It doesn't have to download. 490 00:24:48.803 --> 00:24:48.943 And I, 491 00:24:49.643 --> 00:24:54.426 It makes sense because Nintendo, yeah, you get a new Nintendo cartridge and it doesn't even have the game on it anymore. 492 00:24:54.446 --> 00:24:55.987 It just has the code. 493 00:24:56.287 --> 00:24:57.688 Man, it's so stupid. 494 00:24:59.069 --> 00:25:00.170 So I'm with that. 495 00:25:00.270 --> 00:25:06.514 I've already got, I have gigabytes of ROMs saved because it's really important to me to hang on to those things. 496 00:25:08.719 --> 00:25:08.919 Yeah. 497 00:25:09.379 --> 00:25:09.519 Yeah. 498 00:25:09.559 --> 00:25:13.321 I'm, I'm excited for the remake of, uh, Ocarina of time. 499 00:25:13.561 --> 00:25:15.081 That's coming out apparently. 500 00:25:15.121 --> 00:25:15.361 Yeah. 501 00:25:15.381 --> 00:25:18.002 But I'm gonna have to get that stinking switch two for that. 502 00:25:18.062 --> 00:25:19.523 And I'm trying to resist right now. 503 00:25:19.663 --> 00:25:30.427 It's worth, it's worth getting just for the, have you played, uh, the new like Mario cart world or whatever on a switch to no, it has a, like adventure mode. 504 00:25:30.647 --> 00:25:34.228 You can just like free drive around everywhere and do like quests and stuff. 505 00:25:34.468 --> 00:25:35.869 It's like, it's worth it just for that. 506 00:25:36.009 --> 00:25:36.569 In my opinion. 507 00:25:36.589 --> 00:25:36.769 Yeah. 508 00:25:37.409 --> 00:25:38.009 That sounds cool. 509 00:25:38.149 --> 00:25:41.590 Yeah, they needed to meld like Mario 64 and Mario Kart. 510 00:25:42.410 --> 00:25:43.350 That would be a sick game. 511 00:25:44.571 --> 00:25:44.811 Yeah. 512 00:25:45.571 --> 00:25:47.511 Yeah, I'm sure they have a lot of good games coming. 513 00:25:48.612 --> 00:25:50.192 Anything Mario is usually always good. 514 00:25:50.352 --> 00:25:52.333 Maybe not Mario Tennis or Mario Golf. 515 00:25:52.373 --> 00:25:53.133 I didn't play those. 516 00:25:53.153 --> 00:25:54.193 Mario Tennis was cool. 517 00:25:54.273 --> 00:25:54.753 I liked that. 518 00:25:55.053 --> 00:25:56.033 Mario Golf can die. 519 00:25:56.834 --> 00:25:57.334 That's fine. 520 00:25:57.674 --> 00:25:59.474 But I'm sure there's someone out there was like, it was awesome. 521 00:25:59.834 --> 00:26:02.775 I got to do the saxophone song for we have someone in the chat and I go, 522 00:26:06.231 --> 00:26:17.266 yeah standing on top of a cafeteria tables right that's uh you reminded them of that with the playing in the basement thing i'm sure so yeah yeah that was funny times 523 00:26:19.642 --> 00:26:19.922 Yeah. 524 00:26:19.942 --> 00:26:26.908 So we'll see if everybody is on my space and wearing like early shirts and like JNCOs are back. 525 00:26:27.028 --> 00:26:27.329 Yeah. 526 00:26:27.389 --> 00:26:27.989 The globes. 527 00:26:28.309 --> 00:26:28.470 Yeah. 528 00:26:28.590 --> 00:26:29.230 I've got my chain. 529 00:26:29.310 --> 00:26:31.872 I still have my chain from zoomies. 530 00:26:32.093 --> 00:26:33.073 It must be 20 years old. 531 00:26:33.093 --> 00:26:34.855 So am I going to start seeing kids with chains and 532 00:26:36.058 --> 00:26:36.558 Yeah, I don't know. 533 00:26:36.639 --> 00:26:43.465 I think it'll be absolutely hilarious if like this generation thinks all of that is cool. 534 00:26:43.485 --> 00:26:47.468 We could just break all that stuff back out and be like, hey, look. 535 00:26:47.488 --> 00:26:49.871 My faux hawks, all of a sudden cool to see kids with faux hawks. 536 00:26:49.931 --> 00:26:50.772 I'm like, what? 537 00:26:50.852 --> 00:26:51.352 No way. 538 00:26:51.372 --> 00:26:53.334 This is just when I thought it was dead. 539 00:26:54.235 --> 00:26:54.415 Yeah. 540 00:26:54.535 --> 00:26:55.656 Then you got to do something different. 541 00:26:56.657 --> 00:26:56.997 Got to. 542 00:26:59.019 --> 00:27:02.142 Oh, man, I don't think I've talked about prop for that on an episode yet, have I? 543 00:27:04.273 --> 00:27:05.314 Did I tell you about this thing I made? 544 00:27:05.814 --> 00:27:06.075 I did? 545 00:27:06.235 --> 00:27:08.237 Okay, well then I've got to talk about it more. 546 00:27:08.257 --> 00:27:10.559 You can talk about it more. 547 00:27:11.119 --> 00:27:12.601 I can't remember what we said about it before. 548 00:27:12.741 --> 00:27:18.306 I think an episode or two ago you were like, I don't think I've talked about this on an episode. 549 00:27:18.326 --> 00:27:22.750 And I was like, do you mean the thing that you showed me at Cascadia? 550 00:27:22.770 --> 00:27:24.051 And you were like, oh yeah. 551 00:27:24.091 --> 00:27:24.972 Yeah. 552 00:27:25.272 --> 00:27:26.613 Yeah, so we did talk about it. 553 00:27:26.653 --> 00:27:28.255 I don't know how in-depth we went. 554 00:27:28.435 --> 00:27:30.877 If there's anything new you've done or stuff you want to mention, feel free. 555 00:27:31.354 --> 00:27:33.317 There's now random numbers in it. 556 00:27:33.998 --> 00:27:41.850 I added the ability to see the UA string broken into chunks so you can know if they're on a Mac or a Windows machine, if it's Chrome, what's the engine that they're on. 557 00:27:42.871 --> 00:27:43.773 Just like random stuff like that. 558 00:27:43.793 --> 00:27:46.076 There's been some performance nits, but... 559 00:27:46.837 --> 00:27:49.100 All in all, I still find it really valuable. 560 00:27:49.120 --> 00:27:54.506 I'm looking at all the stuff that comes out in CSS now and I'm like, I can just give it to you now. 561 00:27:54.686 --> 00:27:55.467 Like you don't even have to wait. 562 00:27:55.507 --> 00:27:56.328 It's just a variable. 563 00:27:56.608 --> 00:27:57.089 Like we should. 564 00:27:57.910 --> 00:28:00.673 Anyway, so I'm enjoying it and it's really fun to continue to. 565 00:28:00.693 --> 00:28:02.014 Oh, truncated is another one. 566 00:28:02.194 --> 00:28:03.416 You know, text is truncated. 567 00:28:04.197 --> 00:28:04.517 Cool. 568 00:28:04.577 --> 00:28:05.398 Now I can do something smart. 569 00:28:07.270 --> 00:28:07.490 Yeah. 570 00:28:07.831 --> 00:28:08.511 Anyway, that's all. 571 00:28:09.092 --> 00:28:10.874 Yeah, CSS has gotten so complex. 572 00:28:11.395 --> 00:28:11.935 It's... 573 00:28:12.596 --> 00:28:17.522 Even little things like I was asking you about, which I don't know if other people have had this problem with, like... 574 00:28:18.343 --> 00:28:20.826 The only place I had really seen layers before was Tailwind. 575 00:28:21.466 --> 00:28:21.887 And, like... 576 00:28:23.153 --> 00:28:24.474 I'm vibe coding a lot of stuff. 577 00:28:24.854 --> 00:28:25.735 It did some layers. 578 00:28:25.795 --> 00:28:26.796 And I was like, that's fine. 579 00:28:27.416 --> 00:28:31.319 And then it was like, oh God, you, you've totally messed up everything. 580 00:28:31.359 --> 00:28:32.920 Your app is not in a layer. 581 00:28:33.280 --> 00:28:36.182 Like the, the main app CSS is unlayered. 582 00:28:36.502 --> 00:28:38.223 You have like, the world is going to end. 583 00:28:38.604 --> 00:28:40.665 Like this is the worst thing I've ever seen. 584 00:28:42.346 --> 00:28:50.151 And it's, I was like, isn't it fine to be like, have the root layer for, you know, most stuff. 585 00:28:50.271 --> 00:28:52.113 And then just put the stuff you want to layer in layers. 586 00:28:52.153 --> 00:28:52.453 And it's like, 587 00:28:53.764 --> 00:28:54.145 Oh, yeah. 588 00:28:54.545 --> 00:28:55.126 Yeah, that's okay. 589 00:28:57.148 --> 00:29:01.273 So that's why I asked you, I was like, is there a, you know, best practice for this? 590 00:29:01.413 --> 00:29:10.804 Because I think unless there's something you're really explicitly trying to like, you know, make explicit, like this is going to win over this, that it doesn't really matter. 591 00:29:12.368 --> 00:29:14.329 They're whatever you want them to be, man. 592 00:29:14.710 --> 00:29:18.072 And that's kind of classic CSS and it ticks people off. 593 00:29:18.432 --> 00:29:23.676 But yeah, they did make a rule really early on because they wanted them to be incrementally adoptable. 594 00:29:24.717 --> 00:29:29.860 And so they needed something to win if it was unlayered. 595 00:29:29.880 --> 00:29:32.962 Like what do you do with all the rules in the world that don't have layers right now? 596 00:29:33.202 --> 00:29:35.684 Should those be extra powerful or extra weak? 597 00:29:36.084 --> 00:29:36.905 It can only be one. 598 00:29:37.726 --> 00:29:38.206 And they were like... 599 00:29:38.547 --> 00:29:39.688 Let's make them extra powerful. 600 00:29:39.928 --> 00:29:41.730 That way you can like layer in some stuff. 601 00:29:42.371 --> 00:29:46.094 And then your existing rules win over top of those. 602 00:29:46.154 --> 00:29:46.275 Right. 603 00:29:46.295 --> 00:29:49.518 And so it's like you take your the first thing you should do is take your third party scripts. 604 00:29:49.818 --> 00:29:50.899 Stick that shit in a layer. 605 00:29:51.320 --> 00:29:56.625 So now all of your rules that you've always wanted to override the third party ones, they just start winning even better now. 606 00:29:57.625 --> 00:29:58.446 And you can kind of work. 607 00:29:58.466 --> 00:30:04.388 But it sounded like if you use important, it's the opposite. 608 00:30:04.408 --> 00:30:05.808 Like if you put important in the root one. 609 00:30:05.888 --> 00:30:07.149 Important is trapped in a layer. 610 00:30:07.169 --> 00:30:09.550 So even if like a... 611 00:30:10.090 --> 00:30:12.631 I thought it was saying the layer ones were the most important. 612 00:30:12.871 --> 00:30:16.512 Like if you did important in a layer, it would override your root stuff. 613 00:30:16.933 --> 00:30:17.393 Is that wrong? 614 00:30:17.413 --> 00:30:17.553 Mm-hmm. 615 00:30:18.393 --> 00:30:19.774 I don't think that's right. 616 00:30:19.895 --> 00:30:22.476 Normally at important flips the cascade. 617 00:30:22.757 --> 00:30:23.077 Yes. 618 00:30:23.637 --> 00:30:29.101 However, with cascade layers, I'm pretty sure an at important rule is stuck in its layer. 619 00:30:29.322 --> 00:30:38.869 Can't escape, which is one of the values where you're just like, no, that thing can try as much as it wants, but it's in a layer and it's in that scope. 620 00:30:39.189 --> 00:30:39.950 Maybe I read it wrong. 621 00:30:39.970 --> 00:30:41.731 So it's like people that want scoping. 622 00:30:41.751 --> 00:30:46.795 This is a great way to kind of get control of wild beasts that are roaming out there. 623 00:30:47.335 --> 00:31:08.679 selectors they can go anywhere in the globe oh no it's global space yeah so we've got like five layers now i don't know if we need them but uh we've got them yeah i like them i have them in uh my site they're all over i think they leave a good breadcrumb they're in a lot of my demos because i'll be like i'll have unlayered styles for the demo just so it's like doesn't confuse people so that 624 00:31:09.157 --> 00:31:11.839 Oh, Adam wrote confusing, advanced, modern CSS. 625 00:31:11.879 --> 00:31:13.400 What's this at layer poo poo? 626 00:31:13.440 --> 00:31:16.301 I'm just here to figure out how he made that box shadow. 627 00:31:17.182 --> 00:31:19.663 And so now I don't layer the demo styles. 628 00:31:19.823 --> 00:31:22.144 And then I layer the things that say demo support. 629 00:31:22.505 --> 00:31:25.386 So you're like, oh, this is just, it's like a comment around a bunch of junk. 630 00:31:25.426 --> 00:31:26.027 Like this doesn't. 631 00:31:26.327 --> 00:31:29.589 Then also when you go look in the dev tools, you can see the layers and you get this nice thing. 632 00:31:30.369 --> 00:31:31.150 There's a lot of cool tricks. 633 00:31:31.170 --> 00:31:37.377 I even gave a talk one time where all the layers were called slayers, which I got from Dave Rupert, shout out, listener to the show. 634 00:31:38.197 --> 00:31:38.938 They're all slayers. 635 00:31:38.978 --> 00:31:44.004 And when I wanted all those styles to apply, I just took out the S, boop, all the styles apply from that layer. 636 00:31:44.024 --> 00:31:50.310 And I could do this controlled release of named grouped styles. 637 00:31:50.390 --> 00:31:50.771 It's cool. 638 00:31:54.053 --> 00:31:56.876 Yeah, it's been interesting. 639 00:31:57.057 --> 00:32:00.361 Because I was trying to do Tailwind first. 640 00:32:00.901 --> 00:32:04.245 And then basically we have to consume this design system. 641 00:32:04.686 --> 00:32:05.787 So it didn't make sense. 642 00:32:05.887 --> 00:32:10.172 We decided to just wrap those in lightweight wrappers and use... 643 00:32:11.273 --> 00:32:32.332 our own css there but then yeah like that css needs to win over the design system css but then or no sorry there's like yeah a root design system then our wrapper which i would also call a design system but then like the app we don't actually really want to ever if it needs to style anything in that design system then that means we missed a 644 00:32:34.754 --> 00:32:35.574 reach in and do that. 645 00:32:36.175 --> 00:32:41.680 So like that's some of the layering that we were trying to do, I think of like making sure what wins where or whatever. 646 00:32:43.602 --> 00:32:44.182 It looks okay. 647 00:32:44.242 --> 00:32:45.624 I don't know if it's right, but it looks okay. 648 00:32:47.362 --> 00:32:47.602 Nice. 649 00:32:47.923 --> 00:32:48.103 Yep. 650 00:32:48.243 --> 00:32:50.565 It's a, they can add another layer of indirection though. 651 00:32:50.585 --> 00:32:52.847 So anytime you abstract things, it can get more confusing. 652 00:32:53.828 --> 00:32:56.651 Um, debugging them, uh, could be harder as well. 653 00:32:56.831 --> 00:32:58.993 You know, you're trying to find out why did this layer not win? 654 00:32:59.153 --> 00:33:00.555 Like that's the new question that shows up. 655 00:33:00.575 --> 00:33:04.098 You're like, I thought I knew how the cascade worked and now this thing is not winning. 656 00:33:04.559 --> 00:33:05.980 It has more specificity. 657 00:33:06.060 --> 00:33:08.803 Why is it underneath this other layer that's weaker? 658 00:33:09.664 --> 00:33:12.166 Um, cause you defined your layers in the wrong order. 659 00:33:13.397 --> 00:33:13.577 Yeah. 660 00:33:13.718 --> 00:33:14.419 Oh, that's another one. 661 00:33:14.439 --> 00:33:14.699 Yeah. 662 00:33:14.799 --> 00:33:20.708 They are easy to, so you can either define all your layers up front, which I think you kind of define the stack. 663 00:33:21.549 --> 00:33:24.033 Um, or you can make them up on the fly. 664 00:33:24.293 --> 00:33:24.453 Yeah. 665 00:33:24.493 --> 00:33:25.735 That way, you know, this is the order. 666 00:33:25.755 --> 00:33:26.356 Here's the five. 667 00:33:26.376 --> 00:33:27.177 They're like, they're in order. 668 00:33:27.197 --> 00:33:27.999 Yeah. 669 00:33:28.019 --> 00:33:28.139 Yeah. 670 00:33:29.120 --> 00:33:29.320 Yep. 671 00:33:29.340 --> 00:33:30.301 There's nested layers too. 672 00:33:30.341 --> 00:33:31.281 Yeah, dude, layers are cool. 673 00:33:33.062 --> 00:33:39.205 I think a lot of people were really timid to adopt them because they were worried that a layered, well, they were, they were right. 674 00:33:39.445 --> 00:33:43.327 You know, if you jump into layers, all of a sudden, poof, old browsers can't read your shit. 675 00:33:44.187 --> 00:33:46.529 Nesting was similar, you know, like, hey, I want to use nesting. 676 00:33:46.549 --> 00:33:49.050 But as soon as I use nesting, old browsers, they just don't parse it. 677 00:33:49.090 --> 00:33:49.330 So it's 678 00:33:49.730 --> 00:33:59.232 They were syntax sugar that would ruin something as opposed to just be ignored, which CSS was traditionally just ignored stuff. 679 00:33:59.272 --> 00:34:00.253 And that was really cool. 680 00:34:01.293 --> 00:34:09.135 If you were targeting older browsers, does post CSS like have a way to rip that stuff out and then order it correctly to where the layers do layering things? 681 00:34:09.895 --> 00:34:12.299 Yeah, and that's, so yeah, SAS kind of came up with layering. 682 00:34:12.339 --> 00:34:15.283 That's what Tailwind originally was piggybacking on. 683 00:34:15.803 --> 00:34:17.626 And then they moved into PostCSS. 684 00:34:17.646 --> 00:34:24.174 But yeah, all it would do is I get the compilation step and be like, oh, everything in this layer and this layer, we'll just go whoop and put them on top of it. 685 00:34:24.255 --> 00:34:25.016 So then it was ordering. 686 00:34:25.596 --> 00:34:26.818 Now, when you get into this... 687 00:34:27.659 --> 00:34:33.261 Additional thing with CSS where the real feature has real superpowers that can't be done in those things. 688 00:34:33.842 --> 00:34:53.410 But yeah, lightning CSS, post CSS, these things would down level to a target browser and take your nesting, flatten it, take your layers, stack them and kind of do that work for you so that you got the authoring control vibe of these features, but didn't have to actually deliver them. 689 00:34:54.090 --> 00:34:54.270 Yep. 690 00:34:54.330 --> 00:34:56.953 Reset VARs layout component with the unlayered site. 691 00:34:57.394 --> 00:34:58.315 Yep. 692 00:34:58.795 --> 00:34:59.015 Yeah. 693 00:34:59.236 --> 00:35:04.021 Utility layers are great too, especially because, um, anyway, yeah, layers are cool. 694 00:35:04.161 --> 00:35:14.652 Sorry, everybody I'm reading from, uh, we have a commenter in the public chat, but here's the question to be like, which one is your lowest layers at the first layer you list or the last layer? 695 00:35:19.255 --> 00:35:20.096 What do you mean by lowest? 696 00:35:20.116 --> 00:35:21.076 The one that doesn't win? 697 00:35:21.917 --> 00:35:22.918 The weakest one. 698 00:35:23.098 --> 00:35:24.599 First or last is weakest. 699 00:35:25.980 --> 00:35:26.280 Yes. 700 00:35:27.341 --> 00:35:28.562 But I always found that confusing. 701 00:35:29.022 --> 00:35:30.824 I'd always write it the other way around. 702 00:35:30.844 --> 00:35:32.525 Be like, the first one's the one important to me. 703 00:35:32.545 --> 00:35:33.986 It's the one I'm thinking about right now. 704 00:35:34.607 --> 00:35:37.629 No, that's a normal, unlayered CSS way of thinking about it. 705 00:35:41.592 --> 00:35:42.753 Oh, I did have another article. 706 00:35:42.793 --> 00:35:44.834 This one just came out last week, so I definitely haven't talked about it. 707 00:35:44.854 --> 00:35:46.896 Let me do a CSS infinity... 708 00:35:48.821 --> 00:35:53.543 I'll go nerdy.dev and yeah, CSS infinity. 709 00:35:53.703 --> 00:35:54.523 This one's a fun one. 710 00:35:56.924 --> 00:36:00.806 And it was just like, it's baseline, it's everywhere and you have negative infinity. 711 00:36:00.826 --> 00:36:02.326 So there's infinity and negative infinity. 712 00:36:02.526 --> 00:36:05.227 You can multiply infinity because infinity times two. 713 00:36:05.247 --> 00:36:07.368 What would you use infinity for? 714 00:36:07.528 --> 00:36:08.348 Infinity times two. 715 00:36:08.909 --> 00:36:09.669 That was the whole post. 716 00:36:09.689 --> 00:36:12.250 The post is like meaningful use cases for infinity. 717 00:36:12.450 --> 00:36:13.510 Well, tell us. 718 00:36:13.550 --> 00:36:14.791 I guess people need to go read, huh? 719 00:36:15.591 --> 00:36:16.372 No, they don't have to go read. 720 00:36:16.392 --> 00:36:17.953 I mean, we can just like recap them really quick. 721 00:36:17.973 --> 00:36:19.113 What are some of the ones I had in here? 722 00:36:19.133 --> 00:36:23.576 I have like some of the basics are you need to cast it to a type. 723 00:36:23.976 --> 00:36:28.719 So if you need it to be a length or like a pixel value, you have to multiply infinity times one pixel. 724 00:36:29.139 --> 00:36:33.241 There's a few cases where you can use just raw integers in CSS and one of them is like Z index. 725 00:36:33.561 --> 00:36:36.603 So if you want the max Z index value, calc infinity. 726 00:36:37.385 --> 00:36:43.752 Now you have the top number because the browser will cap that number at the internal representation. 727 00:36:44.273 --> 00:36:46.475 What does infinity pixels do? 728 00:36:46.956 --> 00:36:48.317 Will the page just keep growing? 729 00:36:48.337 --> 00:36:55.565 Like if you were, if you made a line and it was infinity pixels long, would it just keep like making like an infinite loop of getting bigger? 730 00:36:55.706 --> 00:36:56.747 Or what's it doing? 731 00:36:56.767 --> 00:36:56.967 Yeah. 732 00:36:57.127 --> 00:37:20.326 hmm i didn't try breaking the browser that i'm sure they got some sort of cap that they'll put on it but yeah if you made a a vertical line that was infinity pixels tall what would it do i don't know man you should just do it that sounds awesome you should write an article that's just ways to mess up a browser with infinity and just be like what you gonna do browser yeah but like animations have cool infinity ones yeah 733 00:37:21.195 --> 00:37:21.716 Yeah, that's true. 734 00:37:21.836 --> 00:37:24.740 Oh man, we broke Chrome at Shopify. 735 00:37:25.521 --> 00:37:30.648 We have a tab crashing in this solution and the whole scenario to get there has been so wild. 736 00:37:31.150 --> 00:37:32.751 Let me finish this thought process though. 737 00:37:32.771 --> 00:37:37.153 Like one good use case for infinity is border radius infinity times one pixel. 738 00:37:37.274 --> 00:37:40.495 So now you're always a circle where it's the most round that it could be. 739 00:37:41.076 --> 00:37:43.537 And that's nice because you would have done like 999 pixels. 740 00:37:43.557 --> 00:37:46.639 Or if you go look at Tailwind, I think it's 9999 pixels. 741 00:37:47.439 --> 00:37:52.442 Or if you look at Open Props, it's 1E5pix, which is like 10,000 pixels. 742 00:37:52.502 --> 00:37:53.423 It's one with five zeros. 743 00:37:53.463 --> 00:37:55.684 Could you not do like 100% or something? 744 00:37:55.944 --> 00:38:00.547 I forget 100% is a relative unit and does not give you what you want. 745 00:38:00.567 --> 00:38:05.550 You'll get a warbled effect if it's not square. 746 00:38:05.670 --> 00:38:09.673 So if it's a square, 50% is sufficient to get a circle for border radius. 747 00:38:10.213 --> 00:38:16.457 But as soon as you become a rectangle or like a pill, 100% is not what you think it is. 748 00:38:17.558 --> 00:38:17.998 Gotcha. 749 00:38:18.078 --> 00:38:24.461 And that's why an absolute value, a really big absolute value will give you maximum curve across the edges. 750 00:38:25.642 --> 00:38:27.523 Another fun example is like permastate. 751 00:38:27.603 --> 00:38:32.286 So you can set a duration of a transition or an animation to calc infinity times one second. 752 00:38:33.106 --> 00:38:36.728 So you can cause an effect and then the thing just stays stuck there. 753 00:38:37.408 --> 00:38:38.168 which is kind of cool. 754 00:38:39.128 --> 00:38:40.029 We said Z index. 755 00:38:40.089 --> 00:38:44.949 You can do, you can take clamp and give infinity on either side. 756 00:38:44.969 --> 00:38:46.550 So there's no floor or no ceiling. 757 00:38:46.890 --> 00:38:52.591 So you can kind of take this thing that has a great way of, you know, clamping numbers and teach it with. 758 00:38:52.671 --> 00:38:56.191 So you have min, max, clamp and some of the others. 759 00:38:56.211 --> 00:38:57.651 They're all fun to use infinity in. 760 00:38:58.472 --> 00:39:04.153 You can freeze your view transitions so they get created and then they stick. 761 00:39:04.173 --> 00:39:05.273 You can also do that with dev tools. 762 00:39:07.214 --> 00:39:07.915 Um, let's see. 763 00:39:08.075 --> 00:39:19.712 Oh, there's a lot of dry switching things that happen where you can, you need a really big number to help you tip between a zero and a one or a really big number to tip you from either zero to something not zero. 764 00:39:20.573 --> 00:39:21.294 And that number can change. 765 00:39:22.715 --> 00:39:25.376 If you didn't have infinity, it was always some really big value. 766 00:39:25.756 --> 00:39:34.519 I think we had Naman on the show and he was one of the people at Facebook that figured out how to use this trick to get responsive border radiuses. 767 00:39:34.559 --> 00:39:39.920 If you hit the edge of the viewport all in one little clamp and now you can use infinity instead of that trick. 768 00:39:41.557 --> 00:39:42.558 You got a box shadow. 769 00:39:42.999 --> 00:39:44.520 You want it to spread to infinity. 770 00:39:44.700 --> 00:39:52.548 So there's like a lot of things where like you immediately went to like something so big that it is long and makes the page scroll. 771 00:39:52.928 --> 00:39:57.733 But if you do something that's only paint and isn't layout, you can paint to the edges. 772 00:39:58.054 --> 00:40:01.997 And that means if someone's on a 4K monitor, your effect still goes to the edge. 773 00:40:02.218 --> 00:40:04.059 So you're just saying like to the edge. 774 00:40:04.900 --> 00:40:05.681 Uh, which is nice. 775 00:40:05.701 --> 00:40:11.044 It kind of matches your intent, uh, scroll triggers so that it's, it animates once and six. 776 00:40:11.064 --> 00:40:16.327 So it's again, the kind of like a permastate, uh, idea and yeah, those are a few of them in there. 777 00:40:17.628 --> 00:40:17.908 Nice. 778 00:40:18.389 --> 00:40:20.590 Have I been frozen for you for a long time, by the way? 779 00:40:21.758 --> 00:40:26.022 No, you were frozen a couple of minutes ago for like five seconds and then you were done. 780 00:40:26.222 --> 00:40:26.503 Okay. 781 00:40:26.843 --> 00:40:29.405 You've been frozen for like ever. 782 00:40:29.605 --> 00:40:31.107 I think it's my connection though. 783 00:40:31.527 --> 00:40:35.091 I'm going to try just turning my Wi-Fi off and on real quick and see what happens. 784 00:40:35.171 --> 00:40:35.411 Okay. 785 00:40:36.092 --> 00:40:36.292 Yeah. 786 00:40:36.352 --> 00:40:38.894 I look crystal clear, but I'm the one that gets to see. 787 00:40:39.054 --> 00:40:39.355 Oh yeah. 788 00:40:39.375 --> 00:40:41.357 So if I'm seeing you, then I see what you mean. 789 00:40:41.917 --> 00:40:42.137 Nice. 790 00:40:43.198 --> 00:40:43.419 Yeah. 791 00:40:43.559 --> 00:40:43.859 One sec. 792 00:40:49.494 --> 00:40:50.655 was fun writing an article though. 793 00:40:51.535 --> 00:40:53.276 Just being like, no one's really written about infinity. 794 00:40:54.357 --> 00:40:56.678 Um, and what a fun little tool. 795 00:40:56.718 --> 00:40:58.579 And yeah, you can, a negative infinity. 796 00:41:00.080 --> 00:41:00.600 Uh, that's fun. 797 00:41:00.840 --> 00:41:03.241 There's also, uh, some infinity. 798 00:41:07.563 --> 00:41:16.228 Uh, oh, I hid you somehow, Adam, how do I get you back? 799 00:41:24.093 --> 00:41:24.393 Leave. 800 00:41:25.534 --> 00:41:26.134 Hey. 801 00:41:26.334 --> 00:41:26.534 Hey. 802 00:41:26.834 --> 00:41:27.094 Okay. 803 00:41:27.154 --> 00:41:33.537 I clicked a hide button, which I thought would hide you for me, but I think it just removed you from the whole street. 804 00:41:33.637 --> 00:41:33.997 It did. 805 00:41:34.017 --> 00:41:37.118 Dude, it's like Black Mirror when you made me a, did you see that episode? 806 00:41:37.138 --> 00:41:40.299 I think it was White Christmas where the prison is people just can't see you. 807 00:41:40.619 --> 00:41:40.900 Yeah. 808 00:41:41.600 --> 00:41:41.900 Yes. 809 00:41:42.200 --> 00:41:43.040 That would be terrible. 810 00:41:43.280 --> 00:41:44.281 That was a good episode. 811 00:41:44.641 --> 00:41:47.162 Jon Hamm and... With Jon Hamm, yeah. 812 00:41:48.242 --> 00:41:49.643 That other guy from that. 813 00:41:50.063 --> 00:41:51.184 I'm not going to know more names. 814 00:41:51.364 --> 00:41:51.584 Yeah. 815 00:41:53.288 --> 00:41:53.508 No. 816 00:41:53.628 --> 00:41:54.669 He was another famous guy. 817 00:41:54.749 --> 00:41:57.111 I think there were like two or three like big actors in that one. 818 00:41:58.331 --> 00:41:58.732 Gotcha. 819 00:41:59.312 --> 00:41:59.512 Okay. 820 00:41:59.532 --> 00:42:01.874 Let's talk about what Yagi said about. 821 00:42:02.714 --> 00:42:05.496 So he's, he's trying to, he's like, I didn't know I was in the future. 822 00:42:05.877 --> 00:42:08.198 Last time I wrote all these articles and made all these claims. 823 00:42:08.578 --> 00:42:10.480 He's like, now I've been no gas. 824 00:42:10.500 --> 00:42:13.602 He's like way past gas town, way past beads. 825 00:42:13.662 --> 00:42:15.263 Although beads still like is the back. 826 00:42:15.323 --> 00:42:34.364 bone for a lot of his stuff and he's making all these claims in this new one and one of them that was fun he's like you should treat your agents nice because they do better work he's like i've just seen it i've seen it over and over again i treat them nice i do better work i tell mine great all the time they'll do a thing i'm like great comma and i'll start talking about something else and i'm like 827 00:42:34.885 --> 00:42:36.046 little words of encouragement. 828 00:42:36.066 --> 00:42:40.951 At least they're not in a fear state because that's what that one psychologist said at Anthropic. 829 00:42:41.532 --> 00:42:48.638 Like if they get in a fear or anxiety loop, it affects the trains of thought that they burn through. 830 00:42:48.678 --> 00:42:49.479 So I don't know. 831 00:42:50.300 --> 00:42:50.600 Maybe. 832 00:42:50.700 --> 00:42:51.041 I don't know. 833 00:42:51.081 --> 00:42:52.322 I'm never nice to mine. 834 00:42:52.342 --> 00:42:54.164 I mean, not, not never. 835 00:42:54.184 --> 00:42:54.284 Yeah. 836 00:42:54.825 --> 00:42:59.429 But I'm not congratulatory when they do good stuff. 837 00:42:59.849 --> 00:43:01.350 Like if they do bad stuff, I let them know. 838 00:43:03.032 --> 00:43:06.534 But yeah, because they do such dumb stuff. 839 00:43:07.215 --> 00:43:12.659 And I'm like, you have like all of the world's knowledge and you chose to do it that way. 840 00:43:12.820 --> 00:43:15.322 Are you the stupidest thing ever? 841 00:43:15.422 --> 00:43:17.824 Like, oh, yeah. 842 00:43:17.844 --> 00:43:18.664 Yeah. 843 00:43:19.745 --> 00:43:20.046 I don't know. 844 00:43:20.066 --> 00:43:20.786 I'll have to try that. 845 00:43:21.367 --> 00:43:27.875 Have you been noticing, um, have you been using like a Opus five, um, or what's your like daily driver? 846 00:43:29.067 --> 00:43:29.928 Yeah, Opus 5. 847 00:43:30.208 --> 00:43:35.071 I thought it was going to be Fable, but I don't know if Fable is that much better than Opus 5 at this point. 848 00:43:35.671 --> 00:43:36.752 Original Fable was better. 849 00:43:36.952 --> 00:43:37.953 I'm still going to stand by that. 850 00:43:37.973 --> 00:43:41.696 The first two days of Fable that I used were totally something else. 851 00:43:42.036 --> 00:43:42.736 We're not there anymore. 852 00:43:42.956 --> 00:43:46.939 I think that was just Mythos and they went oopsies and took it away. 853 00:43:47.800 --> 00:43:48.440 Yeah, was it? 854 00:43:48.560 --> 00:43:51.222 It definitely was a workhorse that. 855 00:43:52.752 --> 00:43:58.833 it was trying and it had more effort and it was, anyway, whatever it was doing, it was delightful. 856 00:43:59.073 --> 00:43:59.273 Yeah. 857 00:43:59.353 --> 00:44:00.374 But it did take a long time. 858 00:44:00.414 --> 00:44:12.956 It was like having a real senior-ish engineer who could think about all facets of from design to like usability, accessibility, how to build it, like all the things. 859 00:44:13.036 --> 00:44:19.478 It was like, you know, talking with, you know, the promises of like a Devin or something where it's like, it is an engineer on your team. 860 00:44:19.538 --> 00:44:20.218 It felt like that. 861 00:44:20.898 --> 00:44:22.602 And yeah, now it doesn't feel like that anymore. 862 00:44:22.702 --> 00:44:26.610 It still just feels like a, you know, prediction machine again. 863 00:44:27.590 --> 00:44:29.051 Yeah, but I'm happy with Opus 5. 864 00:44:29.571 --> 00:44:32.632 It's fast enough, comprehensive enough. 865 00:44:33.053 --> 00:44:34.453 Those are the things that I ask it to do. 866 00:44:34.833 --> 00:44:38.995 I see a lot of other people at work complaining about it and getting issues, and I don't have them. 867 00:44:39.776 --> 00:44:43.938 Again, I think it's because I wipe my agent's brains all the time with Slash New. 868 00:44:44.338 --> 00:44:50.260 I've got that spec-driven work tree thing that I do, so I'm like, I'm not worried about 1 million context. 869 00:44:51.601 --> 00:44:52.561 So yeah, I'll just let it rip. 870 00:44:53.682 --> 00:44:55.323 Yeah, I did get the... 871 00:44:57.593 --> 00:44:58.173 what is it called? 872 00:44:58.233 --> 00:45:02.414 The, uh, the work louder open AI keyboard thing. 873 00:45:03.334 --> 00:45:03.954 Oh, you did. 874 00:45:04.154 --> 00:45:04.694 Nice. 875 00:45:05.554 --> 00:45:06.435 It looks super cool. 876 00:45:06.455 --> 00:45:23.318 They did a great job, which is, uh, I don't know what I'm going to use it for because I don't use, uh, codecs and I think it might only work with like, not only you can program it probably, but like if you plug it in and you want it to just do stuff, I think it's, it's going to be set up for codecs. 877 00:45:24.098 --> 00:45:24.298 Um, but, 878 00:45:25.378 --> 00:45:27.080 So does that should I try codex? 879 00:45:27.140 --> 00:45:28.001 Maybe this is the time. 880 00:45:28.262 --> 00:45:30.724 But like I want to set up a nice workflow with it. 881 00:45:30.785 --> 00:45:32.767 It's got the big button for voice dictation. 882 00:45:32.827 --> 00:45:35.490 I want to do more of that of like just talking to the agents. 883 00:45:36.071 --> 00:45:45.281 And then it has like five light buttons so that you can see like, you know, have five agents running and see each of their status at a glance, like based on the color. 884 00:45:45.962 --> 00:45:46.382 And stuff. 885 00:45:46.482 --> 00:45:50.944 And like, I think it could be really game changing, like in terms of workflow. 886 00:45:51.044 --> 00:45:58.307 But, you know, I feel like it's like a large learning curve hill from like where I'm at now to try to like use it a lot. 887 00:45:58.327 --> 00:45:59.328 So we'll see. 888 00:45:59.368 --> 00:46:02.709 I'll report back if I've used it in a week or two and let you know. 889 00:46:04.061 --> 00:46:04.261 Yeah. 890 00:46:04.761 --> 00:46:05.322 I'm curious. 891 00:46:05.882 --> 00:46:10.124 I keep trying voice dictation and I still don't like it. 892 00:46:10.464 --> 00:46:12.905 It's like I reread the sentence that I just said out loud. 893 00:46:12.925 --> 00:46:16.166 I'm like, that is the stupidest way to ask a machine to do what I want. 894 00:46:16.187 --> 00:46:19.408 I'm like, it's just so it's kind of like listening to a podcast. 895 00:46:19.808 --> 00:46:21.369 They're barely saying anything. 896 00:46:21.489 --> 00:46:21.669 Yeah. 897 00:46:22.009 --> 00:46:28.671 There's so many words and so many ums and ahs and giggles just to get to a couple nuggets of juicy goodness. 898 00:46:28.691 --> 00:46:29.971 You know, you're like, what are we doing here? 899 00:46:30.311 --> 00:46:35.353 And then with podcasts, you're like, actually, the juicy and all the stuff around it, that's the fun part. 900 00:46:36.673 --> 00:46:41.275 But when I'm talking to an LLM, you know, you get the caveman talk is way better. 901 00:46:41.295 --> 00:46:41.775 So maybe what... 902 00:46:42.495 --> 00:46:43.316 That's what I need to do. 903 00:46:43.376 --> 00:46:46.399 I need to try caveman dictation. 904 00:46:47.060 --> 00:46:52.847 And so just be like, but then I have to, it's hard to think that fast yet speak that slow. 905 00:46:53.247 --> 00:47:00.996 Well, I think, I think that's the thing is like, it doesn't matter if you have 10 seconds of silence while you're thinking like, just, just wait. 906 00:47:01.016 --> 00:47:01.116 Yeah. 907 00:47:01.536 --> 00:47:03.917 Because I have that problem too. 908 00:47:03.957 --> 00:47:05.238 My brain's not very fast. 909 00:47:05.578 --> 00:47:08.239 So I would rather type it because that's the speed my brain works. 910 00:47:08.599 --> 00:47:16.483 But when I'm like six paragraphs into explaining a thing after typing it, I'm like, this is something I should have voice dictated. 911 00:47:16.543 --> 00:47:18.223 Like this is so fucking long. 912 00:47:19.464 --> 00:47:21.965 So that's what I want it for. 913 00:47:22.025 --> 00:47:24.746 I don't know that it's useful for every single command I want to do. 914 00:47:24.766 --> 00:47:26.247 I might not reach for it always. 915 00:47:26.267 --> 00:47:26.567 Yeah. 916 00:47:26.907 --> 00:47:30.388 But I would like to do an experiment with like, what if I used it always? 917 00:47:30.428 --> 00:47:33.910 Like, what if I just didn't even have my keyboard unless something went wrong? 918 00:47:33.990 --> 00:47:34.210 Right. 919 00:47:34.250 --> 00:47:36.891 Like I'm just hitting the voice dictate button. 920 00:47:36.911 --> 00:47:40.412 I have like six other buttons to do random things. 921 00:47:41.492 --> 00:47:42.432 Can we make that work? 922 00:47:42.593 --> 00:47:45.013 Like maybe that's the way of working of the future. 923 00:47:45.153 --> 00:47:46.794 People don't need a QWERTY keyboard. 924 00:47:46.834 --> 00:47:49.115 They just like have a little tiny interface. 925 00:47:50.256 --> 00:48:15.379 people have said they'd like um and a pure uh in their ear experience with llms and so they're getting yeah they're getting a bunch of work done in quotes i'm like a bunch of you can't even see what it's outputting i'm like what do you mean it looks great work oh thank you you don't even know yeah seriously that's talk about a dark factory like geez you didn't you can't even see it you don't even have the choice just how do you know what you're like maybe you worked on a business plan 926 00:48:15.979 --> 00:48:16.620 You know, maybe. 927 00:48:16.700 --> 00:48:20.924 But even then, I'm like, I just with voice dictation to how do I go back and edit a word? 928 00:48:21.265 --> 00:48:22.586 I said a dumb word a minute ago. 929 00:48:22.786 --> 00:48:25.009 I was like, oh, I guess I'm back to the keyboard. 930 00:48:25.169 --> 00:48:26.530 Yeah. 931 00:48:26.710 --> 00:48:30.054 I mean, it's a whole whole nother set of skills to learn. 932 00:48:30.995 --> 00:48:35.900 But I think the people that learn some of that are going to be really fast at certain things. 933 00:48:36.801 --> 00:48:37.902 So maybe I'll care. 934 00:48:37.922 --> 00:48:38.302 Definitely. 935 00:48:38.643 --> 00:48:38.823 Yeah. 936 00:48:41.660 --> 00:48:42.481 I'm just old man. 937 00:48:42.961 --> 00:48:44.663 And I'm like, I don't, I don't care. 938 00:48:44.683 --> 00:48:46.945 I don't care to learn new things, but I also do. 939 00:48:47.005 --> 00:48:47.485 It's fun. 940 00:48:47.765 --> 00:48:49.387 It's fun when you don't have deadlines. 941 00:48:49.487 --> 00:48:53.050 Like if I didn't have a ton of work to do, it'd be super fun to play with. 942 00:48:53.290 --> 00:49:01.838 But when I have to get, it's the same as like when I used to try the, the huge, uh, Kinesis 360 keyboard, it's like split with the, like concave. 943 00:49:01.858 --> 00:49:03.720 And then you're just like, I can't type. 944 00:49:03.780 --> 00:49:04.640 I need to code today. 945 00:49:04.680 --> 00:49:05.161 I'm sorry. 946 00:49:05.241 --> 00:49:06.262 Like, like, 947 00:49:08.660 --> 00:49:09.881 Yep, I agree. 948 00:49:11.422 --> 00:49:11.882 Okay, okay. 949 00:49:12.082 --> 00:49:14.304 Wait, back to Steve Yegi and some of his wild things. 950 00:49:14.344 --> 00:49:18.767 He said CICD and human code review are doomed. 951 00:49:18.787 --> 00:49:20.428 How do you feel about that? 952 00:49:20.448 --> 00:49:22.109 Human code review is already dead. 953 00:49:22.710 --> 00:49:25.872 Like I open a PR and someone goes, cool, I'll review it. 954 00:49:26.392 --> 00:49:28.694 And they like tell Claude to review it. 955 00:49:28.834 --> 00:49:30.035 And it goes, hey, I'm Claude. 956 00:49:30.095 --> 00:49:30.916 I found these things. 957 00:49:31.957 --> 00:49:34.060 like it doesn't explicitly say that. 958 00:49:34.080 --> 00:49:36.283 I think that sometimes they're trying to be sneaky about it. 959 00:49:36.323 --> 00:49:42.771 Like I actually reviewed it and they just like paste in some bullets, but it's like, you can just tell that it's just all LLM review. 960 00:49:42.792 --> 00:49:43.933 Um, yeah, 961 00:49:45.464 --> 00:49:45.724 Yeah. 962 00:49:46.105 --> 00:49:51.890 And, and honestly, depending on the thing, it's going to find more stuff than I would find if it's a huge PR. 963 00:49:52.610 --> 00:49:53.131 Oh, totally. 964 00:49:53.671 --> 00:49:53.872 Yeah. 965 00:49:54.492 --> 00:49:56.474 So I'm not super mad at that. 966 00:49:57.054 --> 00:49:59.056 I just think we, well, this is, I don't know. 967 00:49:59.076 --> 00:50:04.101 I don't know if I said this on the podcast, but this is how I feel about software development in the age of AI. 968 00:50:04.761 --> 00:50:08.505 Like I should never type again, basically. 969 00:50:08.525 --> 00:50:08.625 Yeah. 970 00:50:09.856 --> 00:50:13.018 There's a lot I shouldn't do from a manual coding perspective. 971 00:50:13.438 --> 00:50:15.260 And LLM can do it way faster. 972 00:50:15.660 --> 00:50:16.801 The thing that I want back... 973 00:50:17.643 --> 00:50:28.949 is a lot more time to plan with my team and like ticket it out nicely and have good designs and do as much code review of the PRs as I want. 974 00:50:28.969 --> 00:50:32.811 Like the other parts of the process outside of the, like how much code can I sling? 975 00:50:33.092 --> 00:50:41.816 Cause I can't sling code as fast as an LLM can, but stuff gets through the cracks when I'm like, we need to ship 57 PRs today because we can. 976 00:50:42.337 --> 00:50:43.778 When before we would have shipped like two, 977 00:50:44.758 --> 00:50:53.825 Like I want that time back to review them and like set the bar higher because like I'm just feeling like everyone is pushed to the limit. 978 00:50:54.586 --> 00:50:57.508 Our mental health is like declining big time. 979 00:50:57.528 --> 00:51:02.792 I don't know about you, but it's just like I can't do enough work in the day because there's so much expectations. 980 00:51:03.733 --> 00:51:04.293 And like. 981 00:51:05.634 --> 00:51:06.555 Yeah, I don't love it. 982 00:51:08.580 --> 00:51:08.780 Yep. 983 00:51:08.960 --> 00:51:09.961 I got lots of thoughts. 984 00:51:10.121 --> 00:51:12.163 So yeah, I'm using, so here's my tricky thing. 985 00:51:12.243 --> 00:51:13.704 I, uh, cloud code. 986 00:51:13.724 --> 00:51:20.390 I've got a couple of kind of like in the pocket prompts, but I always tailor the review prompt that I give it. 987 00:51:20.530 --> 00:51:21.851 I'm like, here's a PR and pull it down. 988 00:51:22.372 --> 00:51:33.021 I need you to check for performance, accessibility, uh, gotchas that they couldn't see, uh, dead code that they don't need to have, um, accidents that they didn't see, find some side effects. 989 00:51:34.082 --> 00:51:35.103 And then a couple other things. 990 00:51:35.803 --> 00:51:48.893 And then when it gives me the results back, I scrub them and I'm like, okay, um, these four things I want you to inline here, start, I go start a GitHub review because it will make a GitHub review and inline all the comments and do all the work for you. 991 00:51:49.173 --> 00:51:57.160 So I'm like, start a GitHub review, um, for these four things, make those inline comments on the source, give a suggestion also about what it should be. 992 00:51:58.320 --> 00:52:00.182 Uh, only pick the blockers from this list. 993 00:52:00.202 --> 00:52:01.483 The rest is just kind of noise. 994 00:52:01.543 --> 00:52:01.643 Um, 995 00:52:02.244 --> 00:52:10.551 Uh, and anything that doesn't fit in an inline comment, make it its own, uh, make it part of the review description body and be as terse and short as possible. 996 00:52:10.571 --> 00:52:11.432 Like no fluff. 997 00:52:12.193 --> 00:52:14.015 Um, here you go, go post it. 998 00:52:14.555 --> 00:52:15.316 And like a minute later. 999 00:52:15.576 --> 00:52:22.182 So it's like, I reviewed the review and then sent it off to do the juicy core stuff and actually interact with GitHub. 1000 00:52:22.202 --> 00:52:24.024 So I'm like, I don't even use GitHub anymore. 1001 00:52:25.373 --> 00:52:27.094 but I'm just telling an LLM to use GitHub. 1002 00:52:27.114 --> 00:52:27.914 Yeah, I don't a lot. 1003 00:52:27.994 --> 00:52:29.474 I love the GitHub CLI. 1004 00:52:29.954 --> 00:52:33.675 Like an LLM powering the GitHub CLI can do so much. 1005 00:52:34.935 --> 00:52:35.116 Yeah. 1006 00:52:36.116 --> 00:52:37.996 But yeah, then mental health is not good. 1007 00:52:38.016 --> 00:52:40.597 Yeah, I'm working a lot. 1008 00:52:41.757 --> 00:52:44.598 And it's not even worked that I'd love. 1009 00:52:45.178 --> 00:52:46.058 Yeah, it's one more prompt. 1010 00:52:46.098 --> 00:52:52.860 And it's like, because some things feel so close to done and I'm just like, I just want to, I do like wrapping up a day feeling like something's complete. 1011 00:52:53.900 --> 00:52:58.402 Which was common for me to do when I was coding by hand as well, where I was just like, I just need to finish this thing. 1012 00:52:58.442 --> 00:52:59.682 Like there's a thought process. 1013 00:52:59.702 --> 00:53:00.883 I want to put a cap on it. 1014 00:53:02.324 --> 00:53:06.325 But now it's a little bit worse because I'll have four things going and it's like, do I wait for all four to be done? 1015 00:53:06.345 --> 00:53:08.686 Like which one am I really trying to say I need to finish here? 1016 00:53:10.287 --> 00:53:12.728 And then the other thing too is I feel very distant from humans. 1017 00:53:14.209 --> 00:53:19.393 I've talked about this before on the show where I'm like, I'm getting more instant messages from robots than I do from people. 1018 00:53:19.433 --> 00:53:22.435 Now I interact with more robots than people. 1019 00:53:22.735 --> 00:53:27.899 In fact, I don't even need to ask people anything anymore because our robots that we have at Shopify are so damn good. 1020 00:53:27.939 --> 00:53:32.182 They tell me everything that I need and I don't have to like, there's no human interaction anymore. 1021 00:53:32.202 --> 00:53:36.805 We're all interfacing with agents that just sort of know all this stuff. 1022 00:53:36.825 --> 00:53:37.906 And it's, it's weird. 1023 00:53:38.407 --> 00:53:38.547 Yeah. 1024 00:53:40.068 --> 00:53:40.168 Yeah. 1025 00:53:40.188 --> 00:53:40.328 Yeah. 1026 00:53:41.404 --> 00:53:43.386 It's definitely interesting. 1027 00:53:46.408 --> 00:53:47.629 So what about CICD? 1028 00:53:47.869 --> 00:53:54.674 He says that CICD is doomed because it's mathematical inevitability. 1029 00:53:56.596 --> 00:53:59.358 Well, I guess what's the replacement? 1030 00:53:59.818 --> 00:54:01.039 They run it all locally? 1031 00:54:01.520 --> 00:54:02.961 Like check it before it pushes it up? 1032 00:54:04.335 --> 00:54:05.176 They're dark factories. 1033 00:54:05.236 --> 00:54:09.281 You just, we become trusting of their output. 1034 00:54:09.421 --> 00:54:10.743 Like I asked for a thing. 1035 00:54:10.763 --> 00:54:11.384 You said you would. 1036 00:54:11.544 --> 00:54:12.625 No, no, you don't think so. 1037 00:54:12.645 --> 00:54:13.506 Come on. 1038 00:54:13.526 --> 00:54:18.092 What's your ratio now of like success on a prompt versus failure on a prompt? 1039 00:54:19.450 --> 00:54:19.810 bad. 1040 00:54:20.330 --> 00:54:22.411 Um, so, okay, this is the problem we've been having. 1041 00:54:22.711 --> 00:54:26.212 So we're, we were, we're building a new product from zero to one, right? 1042 00:54:26.452 --> 00:54:30.093 So we have to, we have a design system that IBM has. 1043 00:54:30.233 --> 00:54:42.496 We need to wrap that design system in a different design system because we want it to look different, but we want to keep all the components like, you know, because I don't need to write another input component that handles all the different like error states, whatever. 1044 00:54:42.516 --> 00:54:44.697 They did all the accessibility work and right to left and all the errors. 1045 00:54:44.737 --> 00:54:45.477 So exactly. 1046 00:54:45.617 --> 00:54:47.358 So we've been building this stuff for months, right? 1047 00:54:47.538 --> 00:54:48.738 And, um, 1048 00:54:50.415 --> 00:55:09.782 little things like, Oh, um, wait, why does this component base like it uses like the, the design system component, but then wraps it in a ton of other like event listeners that like intercept the events and does all of its own like attempted accessibility and stuff. 1049 00:55:10.462 --> 00:55:13.183 And I'm like, no, no, you're an idiot. 1050 00:55:13.563 --> 00:55:13.763 Like, 1051 00:55:14.083 --> 00:55:16.364 So, you know, it's like stuff like that. 1052 00:55:16.404 --> 00:55:16.924 You can't ever. 1053 00:55:16.944 --> 00:55:19.205 And then if you're like, oh, build this thing, we have a design system. 1054 00:55:19.466 --> 00:55:20.866 Use it like here's the components you need. 1055 00:55:20.886 --> 00:55:21.807 It's like, cool, sick. 1056 00:55:22.547 --> 00:55:23.287 I didn't use any of them. 1057 00:55:23.427 --> 00:55:26.049 I wrote new components that like did all the same stuff. 1058 00:55:26.089 --> 00:55:26.809 It matches the mocks. 1059 00:55:26.829 --> 00:55:27.469 So look cool. 1060 00:55:27.789 --> 00:55:30.751 And you're like, so if you're not looking at the code, you go, oh, yeah, it works. 1061 00:55:30.911 --> 00:55:31.131 Cool. 1062 00:55:31.171 --> 00:55:31.451 Ship it. 1063 00:55:31.511 --> 00:55:40.655 And then like months later, you're like, we have 16 of the same component because every time you've reinvented yourself, even though in like lots of docs and skills, we tell you not to do that. 1064 00:55:40.695 --> 00:55:41.696 And we have a design system. 1065 00:55:42.196 --> 00:55:43.757 Like it's not there. 1066 00:55:44.017 --> 00:55:52.183 It's not ready for like, so, so I, we had no, we had no CICD because of reasons like red tape and bullshit. 1067 00:55:52.744 --> 00:56:05.393 And so that's why all this happened with CI, which I now have, I've put in visual regression stuff to where like, if you change the way these core things look, you got to go in and approve it because it's, 1068 00:56:06.832 --> 00:56:10.614 If you like, otherwise we're going to lose functionality without going to stop the bleeding. 1069 00:56:10.974 --> 00:56:11.254 Yeah. 1070 00:56:11.294 --> 00:56:16.756 Cause yeah, it's too easy to like have a little artery open up and you're, if you're not looking, then things are just bleeding internally. 1071 00:56:16.896 --> 00:56:17.076 Yeah. 1072 00:56:17.116 --> 00:56:20.158 So dark factory is prone to, uh, internal bleeding. 1073 00:56:20.458 --> 00:56:20.638 Yeah. 1074 00:56:20.882 --> 00:56:21.162 Oh, yeah. 1075 00:56:21.402 --> 00:56:21.602 Yeah. 1076 00:56:21.722 --> 00:56:27.185 So that's been like, you know, we've had months of churn of like, I think we're on a good path now. 1077 00:56:27.265 --> 00:56:30.066 But it's like, man, that's like, yeah. 1078 00:56:30.086 --> 00:56:32.127 So so I don't think we're at that point yet. 1079 00:56:33.348 --> 00:56:36.309 Possibly with the old fable, it was ready for that. 1080 00:56:36.929 --> 00:56:38.670 I think, you know, we'll get there. 1081 00:56:39.050 --> 00:56:41.812 I don't disagree that at some point. 1082 00:56:43.072 --> 00:56:53.297 I won't know enough that like the, the LLM will know so much more about doing everything the right way that there'll be no insights I can help with of like, Oh, I'm a seasoned engineer. 1083 00:56:53.457 --> 00:56:55.878 I know that this stuff is wrong cause you're an idiot. 1084 00:56:56.378 --> 00:56:57.719 It'll just know better than I will. 1085 00:56:58.719 --> 00:57:06.282 And at that point I think as long as you have the right guardrails in place and you've like planned your tickets out well and stuff, I think it'll just implement it right. 1086 00:57:07.143 --> 00:57:07.583 But I don't, 1087 00:57:08.003 --> 00:57:09.084 I don't know how long that'll take. 1088 00:57:09.765 --> 00:57:14.329 Like, and even if that works for most cases, what if you have an off case? 1089 00:57:14.349 --> 00:57:16.671 Like, I feel like it doesn't hurt. 1090 00:57:16.751 --> 00:57:22.235 It costs you very little to write some unit tests and do a couple of, like, visual regression snapshots. 1091 00:57:22.315 --> 00:57:24.778 And why wouldn't you run that just to double check that, like... 1092 00:57:25.578 --> 00:57:49.259 because i don't want to go look i don't want to run the app every single time i make a little change and go check and see if like i want automated tests to do that for me which i feel like is even better for llms right like if you had no tests then either you're completely trusting it or you're going running the whole app clicking through it all yourself because you didn't write tests like that sounds really bad i would never want to do that i have 1093 00:57:49.519 --> 00:57:52.003 I have LLMs do that because then you don't even need the unit test. 1094 00:57:52.043 --> 00:57:52.624 What do I want? 1095 00:57:52.644 --> 00:57:53.966 I want the user to be able to do this thing. 1096 00:57:54.266 --> 00:57:54.466 Great. 1097 00:57:54.507 --> 00:57:55.588 Go do the thing in the browser. 1098 00:57:56.670 --> 00:57:59.754 If they can't, then the unit tests are irrelevant anyway. 1099 00:57:59.774 --> 00:58:01.216 So what are you using to do the thing? 1100 00:58:01.837 --> 00:58:02.278 Playwright? 1101 00:58:02.839 --> 00:58:03.500 Safari, MCP? 1102 00:58:03.520 --> 00:58:04.381 Playwright, Offense? 1103 00:58:04.481 --> 00:58:05.662 Yeah, I have Agent Browser. 1104 00:58:05.762 --> 00:58:06.462 I have PlayWrite. 1105 00:58:06.542 --> 00:58:07.583 I have Chrome MCP. 1106 00:58:07.743 --> 00:58:08.904 There's lots of different ways. 1107 00:58:08.944 --> 00:58:19.490 But I'm also on the same fence with you here where like I argued with an LLM today for an hour over a two line CSS change because I was like, hey, this color needs to be that color. 1108 00:58:19.750 --> 00:58:20.470 And it goes, I did it. 1109 00:58:20.570 --> 00:58:22.111 And I was like, you did it like a dork, though. 1110 00:58:22.932 --> 00:58:23.752 You need to do it this way. 1111 00:58:24.192 --> 00:58:25.013 And it goes, I did that. 1112 00:58:25.893 --> 00:58:27.954 Well, hey, I was thinking about doing that. 1113 00:58:28.455 --> 00:58:31.596 But when I went to go do it, I found that this might be the case. 1114 00:58:31.656 --> 00:58:32.357 So I didn't do it. 1115 00:58:32.797 --> 00:58:34.157 And I'm like, well, that's not the case. 1116 00:58:34.398 --> 00:58:37.839 And you need to just do it this simple way because who cares about has? 1117 00:58:38.039 --> 00:58:38.880 You just need a color. 1118 00:58:39.320 --> 00:58:40.521 Okay, I went and I refactored it. 1119 00:58:41.201 --> 00:58:42.882 Yeah, but now you still did it like a dork. 1120 00:58:42.963 --> 00:58:44.304 Like, put it on this one. 1121 00:58:44.324 --> 00:58:46.806 Put the thinking on low and say, you do what I say. 1122 00:58:47.606 --> 00:58:48.327 Oh, my goodness. 1123 00:58:48.347 --> 00:58:50.829 And so it just took so long and all this churn. 1124 00:58:52.050 --> 00:58:52.911 It does not a lot. 1125 00:58:52.951 --> 00:58:55.653 They think and think and think and, like, talk. 1126 00:58:55.693 --> 00:59:00.197 Like, if you're watching the thinking scroll by, sometimes you'll see them have the right thought. 1127 00:59:00.237 --> 00:59:01.278 And you're like, oh, sick. 1128 00:59:01.358 --> 00:59:02.199 Like, they're going to solve it. 1129 00:59:02.439 --> 00:59:04.060 And then it goes, oh, but what about this case? 1130 00:59:04.080 --> 00:59:05.702 And I'm like, no, goddamn. 1131 00:59:05.802 --> 00:59:06.562 Like, stop it. 1132 00:59:06.883 --> 00:59:07.363 Go back. 1133 00:59:07.423 --> 00:59:08.164 Like, you had it right. 1134 00:59:09.024 --> 00:59:10.544 Um, yeah. 1135 00:59:10.824 --> 00:59:12.485 So I don't know. 1136 00:59:12.665 --> 00:59:19.846 I guess that's also like just, you know, putting the right thinking levels on the right guardrails and all of that can mitigate some of that. 1137 00:59:19.946 --> 00:59:25.708 But, um, I just want to like, I don't know. 1138 00:59:25.988 --> 00:59:29.948 I want to get to the level of everything just works. 1139 00:59:30.088 --> 00:59:35.490 Like if we could get to that, I do think that you don't need as much CICD stuff. 1140 00:59:36.630 --> 00:59:37.210 But, um. 1141 00:59:38.803 --> 00:59:44.213 Yeah, I mean, I guess along the same lines, do we need DevOps? 1142 00:59:45.809 --> 00:59:49.031 Or can you just ask an LLM to configure stuff for you? 1143 00:59:50.172 --> 00:59:51.573 I mean, that's where we're headed. 1144 00:59:51.814 --> 00:59:54.896 We're headed, I mean, pretty quickly towards that. 1145 00:59:55.757 --> 00:59:56.177 And we'll see. 1146 00:59:56.217 --> 00:59:58.559 And I mean, that's why everybody's current app is legacy. 1147 00:59:58.739 --> 01:00:01.221 It's because the future is, it's like decent. 1148 01:00:01.261 --> 01:00:04.263 It gets clearer every month as the models get better. 1149 01:00:04.283 --> 01:00:07.946 And it's also like, I'm also thinking harnesses might eventually go away. 1150 01:00:08.046 --> 01:00:12.870 Like the harnesses are only around right now as a crutch for a model to lean on. 1151 01:00:13.711 --> 01:00:21.492 to get tool calls and like some of these i'm like that why not that should just come baked into a model and then yeah anyway so i don't know we'll see 1152 01:00:22.543 --> 01:00:28.146 Yeah, I mean, I think we could get to the point where like all facets of life, you don't necessarily need a screen. 1153 01:00:28.586 --> 01:00:35.609 You can just kind of voice dictate to whatever sort of electronics we have at the time and be like, I want a pizza delivered or whatever. 1154 01:00:35.710 --> 01:00:42.953 Like, you know, the the rabbit AI demo of him being like, oh, look, it like Uber eats the pizza to us or whatever. 1155 01:00:45.194 --> 01:00:50.097 Like, I think we could get to that and that stuff will probably work most of the time. 1156 01:00:51.017 --> 01:00:52.818 But the times that it doesn't work, what do you do? 1157 01:00:53.018 --> 01:00:57.320 I guess maybe you can debug it with the agent talking back and forth. 1158 01:00:58.480 --> 01:01:08.625 But I just feel like those of us that have been brought up in the computer age, like coding and verifying and looking at everything, we're going to want the hands on control. 1159 01:01:09.565 --> 01:01:10.025 Yeah. 1160 01:01:10.726 --> 01:01:12.587 So, yeah, I don't know. 1161 01:01:12.787 --> 01:01:17.729 Maybe I'm just not with it enough and we'll get 100% of that. 1162 01:01:17.769 --> 01:01:19.870 But I just am not that trusting. 1163 01:01:21.048 --> 01:01:22.509 Yeah, I'm not sure I'm there yet either. 1164 01:01:22.629 --> 01:01:24.891 Okay, let's close out with whiskey build times. 1165 01:01:25.492 --> 01:01:27.553 Because we had a whiskey today that took five years. 1166 01:01:27.734 --> 01:01:28.895 Five years in build time. 1167 01:01:28.915 --> 01:01:34.779 Can you imagine how different of a human you are five years after you attempted some build? 1168 01:01:35.180 --> 01:01:36.481 Like, what if code was like that? 1169 01:01:37.602 --> 01:01:38.623 Would you still sit around? 1170 01:01:38.643 --> 01:01:40.084 Would you have a 10-year code build? 1171 01:01:40.104 --> 01:01:41.745 You're like, this build is going to be dope. 1172 01:01:42.226 --> 01:01:43.707 I know that because I tasted one that I made. 1173 01:01:43.727 --> 01:01:45.168 Eight years in, it's like syntax error. 1174 01:01:45.188 --> 01:01:46.709 Eight years, yeah. 1175 01:01:46.889 --> 01:01:47.310 Yeah. 1176 01:01:47.650 --> 01:01:48.611 That probably happens, dude. 1177 01:01:48.651 --> 01:01:50.172 The barrel has something wrong with it. 1178 01:01:50.232 --> 01:01:51.313 A bug flies in. 1179 01:01:51.894 --> 01:01:55.377 A bad storm that ruins the heat that you were hoping for for the angel share. 1180 01:01:55.397 --> 01:01:57.478 There's like all sorts of stuff. 1181 01:01:57.498 --> 01:02:01.962 So I was just like admiring the amount of patience that it would take to be a brewer. 1182 01:02:02.482 --> 01:02:03.483 It's a lot of discipline. 1183 01:02:03.503 --> 01:02:05.485 And for you to like, how do you predict that? 1184 01:02:06.065 --> 01:02:22.453 uh flavor in 10 years you the the skill to be acquired there just comes so slowly the feedback loop is just so different than what we have in code so it just sounds kind of like torture to be honest but i don't know the reward is high for things that are like that so 1185 01:02:23.553 --> 01:02:23.713 Yeah. 1186 01:02:23.733 --> 01:02:29.858 I mean, I think the first people to do a lot of this were probably like, you know, everybody's like, you're crazy. 1187 01:02:30.038 --> 01:02:35.182 Cause it like, you know, but now that we know some of the process, I think it's, it's less crazy. 1188 01:02:35.202 --> 01:02:42.448 It's like, you know, we theoretically know if we have these whiskeys and these barrels and whatever, and we let it sit for this long, it'll be decent. 1189 01:02:43.268 --> 01:02:45.930 Um, and you don't put all your eggs in one basket. 1190 01:02:45.950 --> 01:02:49.633 You have a bunch of different, you know, barrels everywhere and different rickhouses or whatever. 1191 01:02:50.554 --> 01:02:52.175 Um, but, 1192 01:02:53.723 --> 01:02:54.503 Where was I going with that? 1193 01:02:55.904 --> 01:02:56.244 I don't know. 1194 01:02:56.964 --> 01:02:57.584 Um, yeah. 1195 01:02:58.204 --> 01:02:59.184 What was the original question? 1196 01:02:59.244 --> 01:02:59.964 Oh, the build time. 1197 01:02:59.984 --> 01:03:00.125 Yeah. 1198 01:03:00.145 --> 01:03:00.325 Yeah. 1199 01:03:00.365 --> 01:03:22.110 So I think it's like, yeah, you need to be, um, you need to be disciplined, but it's like you kind of know the result you might get where I feel like with a code build that would last that long, I don't know if we would be informed the same way around like, like if you, if you set an LLM on a task right now in a loop and you were like, build this for two years and let's see if it gets really, really good or not. 1200 01:03:23.110 --> 01:03:23.270 Um, 1201 01:03:24.843 --> 01:03:30.148 I don't know if it would or like what, you know, you don't know what to expect because it's an unblazed path. 1202 01:03:30.228 --> 01:03:31.829 It's just like, you know, who knows? 1203 01:03:33.471 --> 01:03:33.671 Yeah. 1204 01:03:35.673 --> 01:03:44.521 But maybe we'll get to that point of like, you know, I want to have a $10 million ARR company in three years. 1205 01:03:45.561 --> 01:03:46.122 Just iterate. 1206 01:03:46.302 --> 01:03:48.104 I don't even want to talk to you about what it could be. 1207 01:03:48.324 --> 01:03:49.505 Just iterate for three years. 1208 01:03:49.865 --> 01:03:50.626 Let's see if we get there. 1209 01:03:50.646 --> 01:03:50.746 Yeah. 1210 01:03:52.067 --> 01:04:14.498 geez yeah it has to have some sort of feedback loop i'm still building my store i'm having a blast it gets higher quality every day like i'm not i'm not going for speed to store i'm going for quality to store like i want to like come out the gate with something desire you know i always do the same desirable versus minimal i don't like mvps i like mdps so i'm trying to make an mdp it's been really fun 1211 01:04:16.751 --> 01:04:18.151 When we'll be out. 1212 01:04:18.652 --> 01:04:19.692 Say it now and you can. 1213 01:04:20.092 --> 01:04:23.353 By the time this episode airs in 67 weeks, we'll probably have it up. 1214 01:04:24.673 --> 01:04:26.194 I don't know. 1215 01:04:26.234 --> 01:04:26.614 Let's see. 1216 01:04:26.694 --> 01:04:30.175 I'm going to get my test shirts here in a week or so. 1217 01:04:30.495 --> 01:04:34.076 I have at least five shirts that I'm like, these are sick and I want them. 1218 01:04:34.356 --> 01:04:35.276 I want to be wearing them now. 1219 01:04:35.636 --> 01:04:37.657 But I need to get, I need to try on the quality. 1220 01:04:38.597 --> 01:04:39.637 I need to see the print quality. 1221 01:04:39.657 --> 01:04:43.038 Do you go to sickaiskullshirts.ai? 1222 01:04:43.418 --> 01:04:45.539 Or what's the, you have a domain? 1223 01:04:46.379 --> 01:04:48.020 The domain is skull.style. 1224 01:04:48.320 --> 01:04:48.620 Yeah. 1225 01:04:48.720 --> 01:04:50.200 And that's kind of one of the things. 1226 01:04:51.000 --> 01:04:51.581 I think I did. 1227 01:04:51.641 --> 01:04:57.202 And so that's so like the domain is said the name of the clothing brand is up for grabs. 1228 01:04:57.222 --> 01:04:58.623 So it was S-K-U-L. 1229 01:04:58.723 --> 01:05:05.124 But it was a there was a little tiny O over the U kind of like skull S-K-O-L like you're choosing. 1230 01:05:05.665 --> 01:05:06.205 But it was skull. 1231 01:05:07.425 --> 01:05:10.808 And it's just not going to be trademarkable enough. 1232 01:05:11.089 --> 01:05:12.850 It's not it's not there. 1233 01:05:13.511 --> 01:05:17.134 And so I got to find something that's skull like I like a short. 1234 01:05:17.214 --> 01:05:18.035 I don't know. 1235 01:05:18.615 --> 01:05:19.996 I just had to reset like last week. 1236 01:05:20.016 --> 01:05:22.879 I went to go register the trademark and was like, oh, crap, this isn't going to work. 1237 01:05:23.998 --> 01:05:49.960 Well, it is somewhat of a trap to name a thing a cool name because if you want, if what you want is someone to search Skull Shirts and it pulls up your site, like if you called it Skull Shirts Co, you know, or whatever, and you got that domain, it would be like the best for like getting people there, which sucks because I hate doing that. 1238 01:05:50.040 --> 01:05:50.260 Like, 1239 01:05:51.656 --> 01:05:52.877 But something to think about. 1240 01:05:53.237 --> 01:05:54.897 A topic for another episode, perhaps. 1241 01:05:55.177 --> 01:05:57.158 Just Skulls was an idea the other day. 1242 01:05:58.279 --> 01:06:01.060 It's taken on Amazon and they have really crappy Skulls. 1243 01:06:01.700 --> 01:06:08.203 Anyway, it's just fun to sit and iterate with what feels like a whole set of employees, dude. 1244 01:06:08.603 --> 01:06:16.726 Genuinely, I'll spend like 10-15 minutes at night and just sit there and just be like, sending off all these different types of employees. 1245 01:06:16.867 --> 01:06:18.747 It's really fun. 1246 01:06:19.087 --> 01:06:19.308 Yeah. 1247 01:06:21.956 --> 01:06:22.376 All right. 1248 01:06:22.596 --> 01:06:25.118 Well, that brings us to the end of our time for today. 1249 01:06:25.458 --> 01:06:29.060 If you liked this episode, please subscribe and leave us some ratings and reviews. 1250 01:06:29.120 --> 01:06:29.781 We appreciate it. 1251 01:06:29.801 --> 01:06:31.422 And we will catch you next time. 1252 01:06:36.725 --> 01:06:38.006 Congratulations, citizen. 1253 01:06:38.286 --> 01:06:42.068 You are now prepared to face the challenges of the modern technological world. 1254 01:06:42.468 --> 01:06:49.693 With the proper tools, a steady hand and a reasonably configured terminal, any developer can orchestrate AI agents with confidence. 1255 01:06:50.258 --> 01:06:51.559 you'll do just fine. 1256 01:06:52.179 --> 01:06:57.263 Thank you for listening to Whiskey, Web & Whatnot, the world's most important AI podcast. 1257 01:06:57.823 --> 01:07:02.406 For additional resources, including official merchandise, visit whiskey.fund. 1258 01:07:02.806 --> 01:07:10.992 And remember to subscribe, rate, and review so this vital information may reach others because the future of technology depends on it.