1 00:00:00,440 --> 00:00:03,820 -Hey. I'm Michael Dyrynda. -And I am Jake Bennett. 2 00:00:03,820 --> 00:00:04,290 [claps] 3 00:00:05,360 --> 00:00:10,799 And welcome to episode 186 of the North Meet South Web Podcast. [upbeat music] 4 00:00:10,800 --> 00:00:16,100 We are back on the Spindrift kick, folks. It is back and better than ever. Raspberry 5 00:00:16,100 --> 00:00:17,600 -lime. -Raspberry lime. 6 00:00:17,600 --> 00:00:18,700 -So good. -Sounds good. 7 00:00:18,700 --> 00:00:20,220 -Yeah. -There was- 8 00:00:20,220 --> 00:00:22,200 -It's a tasty one -... there's a, there's a brand of, I think 9 00:00:22,200 --> 00:00:26,419 it's like Swedish or something cider, Rekorderlig, and they've got like a- 10 00:00:26,420 --> 00:00:28,620 -Oh, okay -... strawberry and lime flavored cider- 11 00:00:28,620 --> 00:00:30,480 -Ooh! -... which is quite... It's those things, 12 00:00:30,480 --> 00:00:33,360 when I was younger, they were great, but they're so full of sugar, I can't drink 13 00:00:33,360 --> 00:00:36,420 -them anymore. Too old for that stuff. -Yeah, I wonder, I wonder how much sugar 14 00:00:36,420 --> 00:00:40,540 this one has. I don't, I don't think it actually has... It's 1 gram. 15 00:00:40,540 --> 00:00:43,400 -Yeah. -1 gram of sugar, and it's like, it's like 16 00:00:43,400 --> 00:00:47,519 supposed to be zero grams of sugar added. So it's, the ingredients are literally 17 00:00:47,520 --> 00:00:50,760 carbonated water, raspberry puree, and lime juice. That's it. 18 00:00:50,760 --> 00:00:51,020 -Nice. -So- 19 00:00:51,020 --> 00:00:53,360 -It's just fruit. -It should be, you know... Yeah, not too 20 00:00:53,360 --> 00:00:54,350 bad, not too bad. 21 00:00:55,440 --> 00:00:56,920 -[inhales] Oh, my word! -None of that corn syrup that you guys 22 00:00:56,920 --> 00:00:59,600 -like. -Okay. Yeah. Yeah, none of that, none of 23 00:00:59,600 --> 00:01:05,420 that corn syrup stuff. Just, just straight up raspberry puree and, you know, lime 24 00:01:05,420 --> 00:01:06,500 -juice, whatever. -Amazing. 25 00:01:06,500 --> 00:01:09,660 -Um, yes, it is- -I've got 40 minutes before my, uh, my 26 00:01:09,660 --> 00:01:12,1000 -Claude Pro, [chuckles] my to- my, uh- -Before your- 27 00:01:13,000 --> 00:01:13,950 -... my limit refreshes- -Before your- 28 00:01:13,950 --> 00:01:16,780 So, uh, we've got 40 minutes of- [chuckles] 29 00:01:16,780 --> 00:01:19,520 So tell me real quick, you said Claude Pro is on sale for, like, 100 bucks or 30 00:01:19,520 --> 00:01:22,340 -something? -Yeah, I think the price has changed. It's- 31 00:01:22,340 --> 00:01:22,680 Okay 32 00:01:22,680 --> 00:01:27,360 ... uh, Claud- I think the Claude Max subscription is now $100 a month, and- 33 00:01:27,360 --> 00:01:28,720 -Hmm -... I think, 34 00:01:29,340 --> 00:01:29,900 -just looking at- -Rather than $200? 35 00:01:29,900 --> 00:01:33,200 I only looked at it very quick. Yeah, I, because there's two options now. So 36 00:01:33,200 --> 00:01:36,200 there's, there's, like, the free one, which gives you whatever the limit is. 37 00:01:36,200 --> 00:01:39,090 -Yeah. -Then there's Claude Pro, which is 20 bucks 38 00:01:39,090 --> 00:01:40,670 -a month, and it gives you- -Okay 39 00:01:40,670 --> 00:01:44,660 ... five times the limit of the free one. And then there's Claude Max, 40 00:01:45,420 --> 00:01:51,360 which gives you the option now of either five times or 20 times the limit of- 41 00:01:51,360 --> 00:01:52,900 -Yeah -... whatever. So. 42 00:01:52,900 --> 00:01:57,900 Yeah, I think that I've gotten... Let me see here. Adjust usage. 'Cause I have, 43 00:01:57,900 --> 00:02:04,880 right now, I have, hmm, I have the Pro plan, but I have the 5X more 44 00:02:04,880 --> 00:02:05,960 -usage of the Pro plan. -Yeah. 45 00:02:05,960 --> 00:02:06,870 -So I'm paying 100 bucks- -Right 46 00:02:07,780 --> 00:02:12,020 ... for the 5X Pro, and then there's Max, which is, like, another... Yeah, so it's 47 00:02:12,020 --> 00:02:16,340 like 50% off, it says. It's usually, it's usually 400, now it's 200, I think is what 48 00:02:16,340 --> 00:02:17,710 -it is. -Mm. 49 00:02:17,710 --> 00:02:18,260 -Maybe? -Yep. 50 00:02:18,260 --> 00:02:19,850 -I don't know. Yeah. -Yeah, I saw that. 51 00:02:19,850 --> 00:02:21,340 -Yeah. -I just went... I don't remember why I went 52 00:02:21,340 --> 00:02:24,480 to look, but I went to look the other day and saw that it was 53 00:02:25,340 --> 00:02:29,480 only $200 a year, so I signed up finally for, for Claude. So, 54 00:02:30,300 --> 00:02:33,280 -but I've, I've been using- -$200 a month, not a year, right? Not a 55 00:02:33,280 --> 00:02:33,820 -year. Not a- -No, no 56 00:02:33,820 --> 00:02:36,480 -... not a year, a month. -Claude, Claude, Claude Pro is $200 a 57 00:02:36,480 --> 00:02:38,780 -month. Claude Max- -A month. Yeah, yeah. 58 00:02:40,260 --> 00:02:45,220 -Claude Pro, $200 a year. Claude- -A year? 59 00:02:45,220 --> 00:02:47,420 -... Max, yes. -Oh, yeah, yeah, yeah. 60 00:02:47,420 --> 00:02:48,290 -So Claude- -Sorry, sorry. Yes. 61 00:02:48,290 --> 00:02:48,300 -Yeah. -Yes. 62 00:02:48,300 --> 00:02:49,170 -So Claude Pro- -Sorry. Thank you 63 00:02:49,170 --> 00:02:52,049 -... is like $20 a month or $17 a month- -Yeah, yes 64 00:02:52,049 --> 00:02:52,500 -... if you pay for a year. -Thank you. 65 00:02:52,500 --> 00:02:53,489 -Yeah. -Yeah, yeah. 66 00:02:53,489 --> 00:02:56,840 And then Claude Max is $100 a month, or I guess $200- 67 00:02:56,840 --> 00:02:58,640 -Got it -... is still an option there. Yeah, yeah. 68 00:02:58,640 --> 00:02:59,870 -So I'm not, I'm not, uh- -Mm 69 00:02:59,870 --> 00:03:03,310 -... I'm not down to pay $100 a month for- -Okay 70 00:03:03,310 --> 00:03:03,980 -... Claude Max. -Yeah. 71 00:03:03,980 --> 00:03:08,850 But I, I ponied up for the, the Pro annual. We'll see how that goes. 72 00:03:08,850 --> 00:03:08,860 -Yeah. -So. 73 00:03:08,860 --> 00:03:13,920 I, I went, I went with the $100 plan, so I'm on the $100 a month plan, and, 74 00:03:13,920 --> 00:03:17,460 [inhales] uh, it's great. I don't really have to care so much about my token usage. 75 00:03:17,460 --> 00:03:22,440 Just us- use Opus for everything and then let it roll, let it roll. Um, I think I 76 00:03:22,440 --> 00:03:24,720 -have hit token limits once, maybe. -Mm-hmm. 77 00:03:24,720 --> 00:03:27,540 But other than that, I just let it roll. So- 78 00:03:27,540 --> 00:03:29,340 -So, I mean- -I've, I've only switched to Sonnet, like, 79 00:03:29,340 --> 00:03:33,260 -once, but- -Mm-hmm. Yeah, so I... Because... I mean, 80 00:03:33,260 --> 00:03:38,820 for work, we've got team plans, and, and those are fine, and so I, I still tend to 81 00:03:38,820 --> 00:03:40,340 prefer 82 00:03:41,680 --> 00:03:43,120 Opus for planning. 83 00:03:43,960 --> 00:03:45,240 -Mm-hmm. -'Cause, you know, it's got the deep 84 00:03:45,240 --> 00:03:48,320 reasoning, it thinks about things, it goes and figures stuff out. And then I use 85 00:03:48,320 --> 00:03:49,400 Sonnet for 86 00:03:50,220 --> 00:03:53,220 the actual implementation, 'cause it doesn't need to think. It just takes 87 00:03:53,220 --> 00:03:54,360 whatever the plan is and runs with it, so- 88 00:03:54,360 --> 00:03:56,650 Take the plan and go do it. Yeah, yeah, exactly. 89 00:03:56,650 --> 00:03:56,740 -Yeah. -Yep. 90 00:03:56,740 --> 00:03:59,010 -Um, but I had been using- -That's a good, I think a good call 91 00:03:59,010 --> 00:04:01,450 -... GLM, because that was- -Okay 92 00:04:01,450 --> 00:04:04,440 ... $200 a year, and that seemed to work fairly well. And then I think- 93 00:04:04,440 --> 00:04:09,500 -Is that Zed AI? -Yeah, Zed AI. Yep. So that one seemed to 94 00:04:09,500 --> 00:04:12,520 have been working okay, but I thought I would just give it a spin because I was 95 00:04:12,520 --> 00:04:16,279 looking... Minimax is quite good. I, I played with that on the open code, Zen, 96 00:04:16,279 --> 00:04:20,839 when they had that for free. I played with Kimi as well, which, which seems really 97 00:04:20,839 --> 00:04:26,320 good. Um, [chuckles] GLM seems to hallucinate into Chinese. Like, it- 98 00:04:26,320 --> 00:04:27,880 -Okay. Oh, yeah -... it shows you the thinking, and then it 99 00:04:27,880 --> 00:04:29,860 just starts spitting out, um, 100 00:04:30,660 --> 00:04:34,260 Chinese characters or Mandarin or whatever, whatever it is. So that, you 101 00:04:34,260 --> 00:04:38,310 know, it's fine. It's just funny to see that every now and then. But, uh- 102 00:04:38,310 --> 00:04:40,180 -Yeah -... I've, I've been, like, multi-wielding. 103 00:04:40,180 --> 00:04:44,740 I've been, I've been doing, um, Laracon things, I've been doing work things. I've 104 00:04:44,740 --> 00:04:47,920 been working on Arbor, which, uh, is in a, like- 105 00:04:47,920 --> 00:04:48,510 -Yeah -... very close- 106 00:04:48,510 --> 00:04:52,440 -We should talk about Arbor -... to, to usage state. I've been, I've 107 00:04:52,440 --> 00:04:53,040 been using 108 00:04:53,740 --> 00:04:57,500 Claude, I've been using... Well, I've been using Opus and Sonnet. I've been using 109 00:04:57,500 --> 00:05:01,640 GLM, I've been using Kimi, I've been using AMP, all of these things. I'm, I'm right 110 00:05:01,640 --> 00:05:06,880 into it now. Um, still haven't installed OpenClau or MulkBot or Claude Bot, 111 00:05:06,880 --> 00:05:07,600 -whatever- -Yeah, I, I want 112 00:05:07,600 --> 00:05:09,240 -... it is today. -Okay, I want to talk about OpenClau as 113 00:05:09,240 --> 00:05:10,090 -well. So like, okay- -Yeah 114 00:05:10,090 --> 00:05:12,430 ... so I want to talk about Arbor, I want to talk about OpenClau. Um- 115 00:05:12,430 --> 00:05:14,040 -We have to be careful about turning- -And then I want to talk about- 116 00:05:14,040 --> 00:05:15,550 ... this into an AI show as well. [chuckles] 117 00:05:15,550 --> 00:05:17,979 Well, I mean, so here's the deal. It's like, that's just kind of like what's 118 00:05:17,980 --> 00:05:18,680 -happening right now. -Mm-hmm. 119 00:05:18,680 --> 00:05:20,800 It doesn't have to be an A... It's not an AI show. It's just like- 120 00:05:20,800 --> 00:05:22,980 -Mm -... it's, it's kind of like the most 121 00:05:22,980 --> 00:05:25,860 interesting thing that is going on right now. 122 00:05:25,860 --> 00:05:28,200 -Mm-hmm. -And so I want to talk... Uh, we can 123 00:05:28,200 --> 00:05:30,520 briefly touch on those things, and then I can talk about a couple other things if 124 00:05:30,520 --> 00:05:36,120 you want to. Um, but then I also want to talk about horseless carriages. Have we 125 00:05:36,120 --> 00:05:37,810 -talked about this before? -Horseless carriages. 126 00:05:37,810 --> 00:05:39,600 -Horseless carriages. -We have not. 127 00:05:39,600 --> 00:05:41,180 -Okay, I want to talk about that too. -That's a tease- 128 00:05:41,180 --> 00:05:43,040 -Okay, so let's- -... for later in the show. [chuckles] 129 00:05:43,040 --> 00:05:48,430 It is. It is indeed. Okay, so couple things here, um, OpenClau. 130 00:05:48,430 --> 00:05:49,840 -Mm. -So do you want to talk about OpenClau or 131 00:05:49,840 --> 00:05:53,520 -Arbor first? -Let's talk about OpenClau. Let's not talk 132 00:05:53,520 --> 00:05:56,600 -about me. -Okay. Yeah, so I saw Taylor post the other 133 00:05:56,600 --> 00:06:00,760 day that like: "Hey, Forge, you can now provision an OpenClau thing," whatever. 134 00:06:00,760 --> 00:06:02,660 -Mm-hmm. -So what is that? I don't even know what 135 00:06:02,660 --> 00:06:06,200 that is. I, I, I was gonna go look it up, and then I was like, "Eh," I didn't have 136 00:06:06,200 --> 00:06:09,308 time. I was, I was working.... working that day. 137 00:06:09,308 --> 00:06:11,688 -So- -So anyway, tell me about OpenClaw. 138 00:06:11,688 --> 00:06:16,207 OpenClaw was previously, um, 139 00:06:17,888 --> 00:06:21,248 previously MaltBot, which was previously Claude Bot. 140 00:06:23,328 --> 00:06:26,068 -Okay. -So it's the... I think the idea behind 141 00:06:26,068 --> 00:06:30,308 this is that it's, it uses whatever model 142 00:06:31,588 --> 00:06:33,878 -under the covers. So you, you set it up- -Right 143 00:06:33,878 --> 00:06:37,207 ... as like a tele- it uses Telegram as a chat interface, or you can connect it to- 144 00:06:37,207 --> 00:06:37,928 -Oh, interesting -... Slack, WhatsApp, whatever. 145 00:06:37,928 --> 00:06:39,798 -Telegram, okay. -And so it's just- 146 00:06:39,798 --> 00:06:43,707 -Sure -... there is a bot, and then it uses, um, 147 00:06:43,707 --> 00:06:48,308 Claude, uh, you know, Opus or Sonnet or GLM or Kiwi. You can connect it to any 148 00:06:48,308 --> 00:06:49,348 -provider- -Sure. Okay, okay 149 00:06:49,348 --> 00:06:52,508 ... or model under the hood that it supports, and then it, and it uses that. 150 00:06:52,508 --> 00:06:55,587 -Bring your own key? -Bring your own key kind of thing. Um, and 151 00:06:55,587 --> 00:07:00,328 so you host it on a VPS, in a Docker container on your Mac, if you're brave. 152 00:07:00,328 --> 00:07:03,308 There was, like, a, a run a couple of weeks ago, like, everyone went out and 153 00:07:03,308 --> 00:07:09,368 bought Mac Minis to, to host their, um, Claude Bot instances. Um, and, and the 154 00:07:09,368 --> 00:07:13,408 idea is that it's, it's like a, a personal assistant that- 155 00:07:13,408 --> 00:07:15,428 -Okay -... you ask it to do things. You get it to 156 00:07:15,428 --> 00:07:19,568 schedule tasks, you get it to do whatever. So it's like ChatGPT on 157 00:07:19,568 --> 00:07:21,658 -steroids, but I think the, the main- -In Telegram. 158 00:07:22,548 --> 00:07:23,587 -Yeah. -Like, so that's your interface. 159 00:07:23,587 --> 00:07:24,638 -Yeah. -You choose an interface, right? 160 00:07:24,638 --> 00:07:25,768 -Correct. -And so I guess Telegram is a common 161 00:07:25,768 --> 00:07:27,768 -interface. -Yeah. Telegram is probably the easiest 162 00:07:27,768 --> 00:07:32,128 one, I think, from what I've seen, to, to get up and running. Um, 163 00:07:32,188 --> 00:07:32,648 and 164 00:07:33,368 --> 00:07:37,848 it, it's basically, yes, Siri on steroids. Like, if Siri actually worked. Sorry to 165 00:07:37,848 --> 00:07:38,488 everyone whose 166 00:07:39,288 --> 00:07:40,398 -home assistants- -No, it's true 167 00:07:40,398 --> 00:07:43,988 -... have just triggered there, so. Um- -It's freaking true. Yeah. It's ridiculous. 168 00:07:43,988 --> 00:07:45,468 Uh, yeah, so you can get it to do 169 00:07:45,468 --> 00:07:47,908 whatever. You can... You know, people have been hooking it up to their email 170 00:07:47,908 --> 00:07:50,938 accounts, their calendars, reminders, whatever. Like, all of this stuff that 171 00:07:51,828 --> 00:07:55,908 -feels kind of scary, and I think- -Like, using MCPs? 172 00:07:55,908 --> 00:07:59,048 No, just, like, "Here are, here are my username and password. Here are my, you 173 00:07:59,048 --> 00:08:03,448 know, keys," whatever, just giving it access to these, these different tools. So 174 00:08:04,207 --> 00:08:07,368 it's, uh, interesting. It- a- and people go, "You know, it's self-hosted, it's all 175 00:08:07,368 --> 00:08:11,988 local." I think, well, the, the bot software itself, you know, the wrapper 176 00:08:11,988 --> 00:08:17,207 around it, is self-hosted, but it's still shelling out, you know, using the remote 177 00:08:17,207 --> 00:08:19,847 -models. So it's not- -Totally, yeah. 178 00:08:19,847 --> 00:08:23,228 It's not as self-hosted as you think. There's, there's the ability to use- 179 00:08:23,228 --> 00:08:26,447 -You're not running Ollama. -Yeah, there is the ability to, to host 180 00:08:26,448 --> 00:08:27,058 -your own models- -I mean- 181 00:08:27,058 --> 00:08:29,548 ... but I don't think most people are doing that. 182 00:08:29,548 --> 00:08:32,417 -Right. Right. Okay. -But there's, uh- 183 00:08:32,417 --> 00:08:32,688 -Interesting. -Yeah, there's, like, a whole- 184 00:08:32,688 --> 00:08:35,428 I, I would like to see... I would be interested to know kind of what people are 185 00:08:35,428 --> 00:08:36,388 -doing with that, yeah. -Yeah. 186 00:08:36,388 --> 00:08:38,627 -Um- -Jeffrey, like, Jeffrey's got some videos 187 00:08:38,627 --> 00:08:42,507 up on Laracasts about it. He's been talking about it. Um, Dreis has been 188 00:08:42,508 --> 00:08:47,288 talking about it. Like, a whole bunch of the Laravel team have been using it a lot, 189 00:08:47,288 --> 00:08:50,828 and, and so now, you, you know, we've got one-click deployment of it on, uh, on 190 00:08:50,828 --> 00:08:55,348 -Laravel VPS, so yeah. -So instead of po- instead of hosting it 191 00:08:55,348 --> 00:08:57,127 locally on a little, on a little MacBook Mini- 192 00:08:57,127 --> 00:08:58,398 -Yeah, instead of- -... or, yeah, sorry, a Mac Mini 193 00:08:58,398 --> 00:09:01,947 ... hosting it on your own device or, or spinning up a VPS manually and doing it, 194 00:09:01,948 --> 00:09:03,938 -you just, you know, Laravel VPS- -Just go deploy it 195 00:09:03,938 --> 00:09:08,088 ... comes up in five minutes, then, you know, OpenClaw comes up in five minutes 196 00:09:08,088 --> 00:09:09,768 -on, on your- -Hmm 197 00:09:09,768 --> 00:09:12,627 -... um, environment there, so. -I mean, it makes sense. I mean, it's like, 198 00:09:13,648 --> 00:09:16,688 if all you're doing is hooking it up to another location... Like, again, if you're 199 00:09:16,688 --> 00:09:21,018 not using Ollama, you don't really need it, the bunch of, of hardware. 200 00:09:21,018 --> 00:09:23,188 -No. -It's like if you're just literally taking 201 00:09:23,188 --> 00:09:26,308 it and using it as a place to, like, say, "Okay, my chat interface is over in 202 00:09:26,308 --> 00:09:26,798 -Telegram." -Mm-hmm. 203 00:09:26,798 --> 00:09:30,488 "I want you to use my key for this particular thing, and then I'm letting you 204 00:09:30,488 --> 00:09:35,428 do particular other pieces on my behalf," and just kind of keeping it all there. 205 00:09:35,428 --> 00:09:37,028 -Yeah, okay. Interesting. -Mm. 206 00:09:37,028 --> 00:09:38,537 -I can see that. -Yeah. 207 00:09:38,537 --> 00:09:39,668 -Hmm, cool. -I, uh- 208 00:09:39,668 --> 00:09:41,627 -Yeah -... I still haven't dipped in yet. I- 209 00:09:41,627 --> 00:09:43,488 I'll just say I'll be a late, I'll be a late adopter to that. 210 00:09:43,488 --> 00:09:43,918 -I'll be- -I'll figure it out someplace 211 00:09:43,918 --> 00:09:46,448 ... I'll be a late adopter once all the security concerns and things are figured 212 00:09:46,448 --> 00:09:48,078 -out. So it's, uh- -Mm, mm 213 00:09:48,078 --> 00:09:49,648 ... it's definitely interesting, but, I mean, 214 00:09:50,508 --> 00:09:54,867 my, my biggest thing with it is that, you know, when Siri ca- and I, and I don't 215 00:09:54,867 --> 00:10:01,148 know, I'm gonna prefix this with, I don't know if I don't use she who shall not be 216 00:10:01,148 --> 00:10:05,697 -named, because they're not very good- -Yeah 217 00:10:05,697 --> 00:10:09,048 ... or because, like, I just have no use for it. You know, the, the only thing I 218 00:10:09,048 --> 00:10:11,828 really use it for is setting a timer, and even that can sometimes be hidden. 219 00:10:11,828 --> 00:10:12,638 -Timers- -Like, I- 220 00:10:12,638 --> 00:10:13,617 -... reminders- -Yeah 221 00:10:13,617 --> 00:10:14,748 -... calendars- -Yeah 222 00:10:14,748 --> 00:10:17,908 ... sending text messages, making phone calls, right? 223 00:10:17,908 --> 00:10:19,838 -Yeah. -Um, use it for all those things. 224 00:10:19,838 --> 00:10:22,148 Like, stuff that it does well enough that it's, you know... But 225 00:10:23,367 --> 00:10:28,568 g- getting it to do all this other stuff, I just... I'm interested to see how it 226 00:10:28,568 --> 00:10:32,867 shakes out, but I'm not interested enough to actually use it just yet. 227 00:10:32,867 --> 00:10:38,406 Okay, so, so l- let's, let's play the game here real quick, okay? Hypothetically, I 228 00:10:38,408 --> 00:10:38,896 -wanna make sure- -Hmm 229 00:10:38,896 --> 00:10:41,768 ... I'm understanding this correctly, all right? The reason why I asked about the 230 00:10:41,808 --> 00:10:46,188 MCP thing is 'cause that seems like the natural play here. Like, isn't an MCP 231 00:10:46,188 --> 00:10:52,528 basically a new... L- like, so if I have a, um, I don't know, a Gmail MCP, right? 232 00:10:52,588 --> 00:10:53,918 -Mm-hmm. -That's installed. 233 00:10:53,918 --> 00:10:55,887 -Yeah. -And then I authenticate that with my 234 00:10:55,887 --> 00:10:57,087 -Gmail, right? -Mm-hmm. 235 00:10:57,088 --> 00:11:01,087 It now has access probably to my calendar, and to my emails, and to my chats- 236 00:11:01,088 --> 00:11:01,588 -Yeah -... whatever. 237 00:11:01,588 --> 00:11:02,028 -Sure. -Right? 238 00:11:02,028 --> 00:11:05,068 -Mm-hmm. -And it is exposing to my LLM, "Here are 239 00:11:05,068 --> 00:11:08,488 the tools that are available to you through this MCP." Yes? 240 00:11:08,488 --> 00:11:09,617 Yeah. That would be the same way of doing it. 241 00:11:09,617 --> 00:11:12,778 Why would I... That's, that's how I would do it, right? I mean, like, what are we 242 00:11:12,778 --> 00:11:14,968 talking about? They're putting their username and password into, like, 243 00:11:14,968 --> 00:11:17,668 something else. It's gotta be an MCP, right? For OpenClaw. 244 00:11:17,668 --> 00:11:21,188 Well, does Gmail, does, does Gmail have an MCP? 245 00:11:21,188 --> 00:11:21,938 -I don't know. -Like- 246 00:11:21,938 --> 00:11:23,748 I don't know. I would assume that they would. 247 00:11:23,748 --> 00:11:24,728 -Oh, yeah, it does. -Google? 248 00:11:24,728 --> 00:11:25,248 -End of- -Yeah 249 00:11:25,248 --> 00:11:27,568 ... December, they, they, they enabled it, like- 250 00:11:27,568 --> 00:11:28,127 -Okay. -But again- 251 00:11:28,127 --> 00:11:31,268 -So, so, yes -... using an MCP only gives you whatever 252 00:11:31,268 --> 00:11:35,248 -the provider gives you. -So what is this doing? Like, this, what 253 00:11:35,248 --> 00:11:38,208 are they... It's OpenClaw, like, is going on the browser for you and doing UI 254 00:11:38,208 --> 00:11:38,968 -interactions? -Yeah. 255 00:11:38,968 --> 00:11:40,988 -Is the idea? -People... I've seen people do, do all 256 00:11:40,988 --> 00:11:42,928 -kinds of things, um- -Okay 257 00:11:42,928 --> 00:11:46,748 ... where it's going into uncharted territories, and it's, you know, clicking 258 00:11:46,748 --> 00:11:49,348 through... It's using Playwright's MCP to go and click through browsers- 259 00:11:49,348 --> 00:11:50,718 -Oh, interesting -... like, websites- 260 00:11:50,718 --> 00:11:52,488 -Oh, okay -... and doing all this kind of stuff, so. 261 00:11:52,488 --> 00:11:55,367 You'd have to program them all up front, right? It's the idea. 262 00:11:55,367 --> 00:11:57,748 Yeah, you'd... Well, you tell it what you want it to do, and it goes and figures it 263 00:11:57,748 --> 00:11:58,828 -out. Um- -Okay, so- 264 00:11:58,828 --> 00:12:03,068 But I think the, the, the, the, the thing that separates it from something like just 265 00:12:03,068 --> 00:12:07,788 using ChatGPT is that it, it uses its own memory store. So it's got, like, a memory 266 00:12:07,788 --> 00:12:12,788 file and a soul file, and all this kind of stuff.... and, and so it has 267 00:12:12,788 --> 00:12:13,918 -contextual- -You can, you can- 268 00:12:13,918 --> 00:12:16,288 --historical information. -You have full, you have full control over 269 00:12:16,288 --> 00:12:18,068 -the prompt, probably- -Yeah. 270 00:12:18,068 --> 00:12:18,947 -Right? -Yeah. 271 00:12:18,948 --> 00:12:20,028 -Like- -Yeah, you tell it what to go- 272 00:12:20,028 --> 00:12:21,068 -... interesting --and write in there. You give it, you, 273 00:12:21,068 --> 00:12:23,748 like, you give it it the personality you want it to have. Do you want it to be 274 00:12:23,748 --> 00:12:24,308 -sarcastic? -Yeah. 275 00:12:24,308 --> 00:12:27,528 Do you want it to have... You know, I've seen, uh, you know, I, I mentioned Kitze. 276 00:12:27,528 --> 00:12:30,888 He started this, like, Tinkerer Club, which you can pay for access to. It's a 277 00:12:30,888 --> 00:12:33,407 Discord thing. But he, he's got, like, 10 different 278 00:12:34,628 --> 00:12:39,208 OpenClaw instances running, assuming personas of, like, Dr. Cox from Scrubs, 279 00:12:39,208 --> 00:12:40,538 -and, and, like- -Cool. Okay 280 00:12:40,538 --> 00:12:43,968 ... Chandler from Friends, and all of this kind of stuff. He's got like a- he's got 281 00:12:43,968 --> 00:12:48,228 Saul Goodman there acting as his lawyer. So you give all these different instances 282 00:12:48,228 --> 00:12:51,628 of OpenClaw, their own personas that are responsible for their own things, and you 283 00:12:51,628 --> 00:12:54,558 give them access to their own things. Um, so you've got, like, an- 284 00:12:54,558 --> 00:12:56,308 -Yeah, okay -... accountant that has access to all of 285 00:12:56,308 --> 00:12:57,048 that kind of stuff. So 286 00:12:57,867 --> 00:13:01,128 it's, it's definitely, like, he's got a whole army of them. But the, the main 287 00:13:01,128 --> 00:13:04,528 thing is, like, you can use ChatGPT and do all this stuff, but each ChatGPT- 288 00:13:04,528 --> 00:13:06,808 -Well, sort of -... conversation is kind of isolated, 289 00:13:06,808 --> 00:13:09,388 -right? -Yeah, right. Well, and also, you can't 290 00:13:09,388 --> 00:13:12,988 -install MCPs on ChatGPT. -No. 291 00:13:12,988 --> 00:13:16,228 Right? You can only do what ChatGPT is allowing you to do. 292 00:13:16,228 --> 00:13:18,488 -Right. -Now, if they define or create a new tool, 293 00:13:18,488 --> 00:13:22,908 great, and I can now use that tool. But because it's not a local instance, I don't 294 00:13:22,908 --> 00:13:26,468 have control over my own tool chain. I only have access to what they give me, 295 00:13:26,468 --> 00:13:27,008 -right? -Yeah. 296 00:13:27,008 --> 00:13:27,908 -Okay. -Yeah. 297 00:13:27,908 --> 00:13:33,508 So th- this is, this is the sort of thing that I'm, that I'm... I think I understand 298 00:13:33,508 --> 00:13:40,328 it correctly, right? But this feels like the magic that MCPs will bring, which is, 299 00:13:41,048 --> 00:13:44,328 I have access to all... I have all these services that I use that are on the web, 300 00:13:44,328 --> 00:13:45,348 right? Um, 301 00:13:46,388 --> 00:13:48,668 everything's now accessible through the web. I mean, it used to be like, there 302 00:13:48,668 --> 00:13:50,968 were so many things that were stored locally, right? And then we... Remember we 303 00:13:50,968 --> 00:13:54,188 had these netbooks for a while, where it was like there was no operating system, 304 00:13:54,188 --> 00:13:56,487 basically. It was very thin, and just, like, the idea was like, you're going to 305 00:13:56,488 --> 00:13:57,698 do everything in the web. And then it was like- 306 00:13:57,698 --> 00:13:59,127 -Mm -... well, this is a crazy idea. And then 307 00:13:59,127 --> 00:14:00,688 it was a Chromebooks came out, and now it's like, okay- 308 00:14:00,688 --> 00:14:01,808 -Yeah -... everything's actually on the web. 309 00:14:01,808 --> 00:14:05,928 Like, there is no OS. It is all the web, right? And so great, now everything's on 310 00:14:05,928 --> 00:14:10,468 the web. Wonderful. Um, but what if I want to interact with those things, not 311 00:14:10,468 --> 00:14:13,008 through a UI, but we want to do it programmatically? Great, now we have APIs, 312 00:14:13,008 --> 00:14:16,248 we have GraphQL, blah, blah, blah. MCP, I don't want to have to write every 313 00:14:16,248 --> 00:14:20,488 connector. MCP is the promise that, like, you just tell your LLM, you have this 314 00:14:20,488 --> 00:14:24,008 tool, it will figure out what thing it needs to use, and it knows how to talk to 315 00:14:24,008 --> 00:14:28,148 us, and then it'll hand- handle it on your behalf. So what I can do now is I can 316 00:14:28,148 --> 00:14:33,648 say, "Hey, my local instance of OpenClaw, you can interact with me on my Gmail. You 317 00:14:33,648 --> 00:14:37,808 can interact with... You can interact for me on my calendar. You can interact for me 318 00:14:37,808 --> 00:14:41,008 through Yelp. You can leave reviews, you can make reservations, you can an- 319 00:14:41,008 --> 00:14:45,168 interact as me in GitHub. You can interact as me, blah, blah, blah. Here are the 320 00:14:45,168 --> 00:14:47,888 -MCPs that I'm giving you access to." Now- -Mm 321 00:14:47,888 --> 00:14:52,208 ... my OpenClaw instance has, when I ask it to do something, it's going to look at 322 00:14:52,208 --> 00:14:54,768 its tool chain and say, "Are any of these things relevant to what I'm trying to 323 00:14:54,768 --> 00:14:58,038 accomplish?" "Yes, Jake asked me to make a calendar reminder for whatever." 324 00:14:58,038 --> 00:14:59,808 -Mm. -"Okay, great. I'm gonna go talk to that 325 00:14:59,808 --> 00:15:03,648 MCP. I'm gonna go do that thing." So, you know, the previous iterations of, like, 326 00:15:03,648 --> 00:15:07,887 same thing, like, you have she who shall be not, not be named, right? The girl in 327 00:15:07,888 --> 00:15:11,228 the box, Siri, Alexa, all the things, right? 328 00:15:11,228 --> 00:15:13,888 -Mm. -Um, Google, all those. But the limitation 329 00:15:13,888 --> 00:15:17,548 that they had, again, was like, you... There was no way to integrate that unless 330 00:15:17,548 --> 00:15:21,008 you were using the service. So if you're using Apple, fine, you can use their 331 00:15:21,008 --> 00:15:22,088 -calendar reminders. -Mm. 332 00:15:22,088 --> 00:15:24,988 If you use Google, fine, you can use their stuff. But there was no way to do, like, 333 00:15:24,988 --> 00:15:30,768 cross-stuff or to do Yelp stuff with- using that assistant. But now you can, 334 00:15:30,768 --> 00:15:35,988 right? You can roll your own. So, like, the idea of, like, JARVIS that Tony Stark 335 00:15:35,988 --> 00:15:36,478 -had- -Mm 336 00:15:36,478 --> 00:15:40,867 ... where it could just do things as him or for him, was previously thought to be 337 00:15:40,867 --> 00:15:44,367 like, "Oh, that's, you know, that could possibly exist at some point," but MCPs is 338 00:15:44,367 --> 00:15:46,488 what, to me, unlocks that. 339 00:15:47,288 --> 00:15:48,528 -Yes. -I, I think. You know what I mean? 340 00:15:48,528 --> 00:15:53,108 But the, but the MCP still relies on the provider offering it. 341 00:15:53,108 --> 00:15:53,738 -It does. -And, and, and I suppose- 342 00:15:53,738 --> 00:15:54,208 -It does -... it does. 343 00:15:54,208 --> 00:15:58,088 But, like, I rely on the provider giving me those tools anyway. 344 00:15:58,088 --> 00:15:59,018 -Right. -Like, if it's not- 345 00:15:59,018 --> 00:16:00,848 -So this is- -... exposed through an API, there's no way 346 00:16:00,848 --> 00:16:04,488 it's gonna be exposed through an MCP. So it's like, uh, is, you know, the full 347 00:16:04,488 --> 00:16:09,627 functionality is likely gonna be available at some point through that MCP- 348 00:16:09,627 --> 00:16:11,028 -Mm -... um, which means I don't have to write 349 00:16:11,028 --> 00:16:13,068 all the tooling for it. It's done. Like, the LLM- 350 00:16:13,068 --> 00:16:15,208 -Yes -... figures out the tooling, right? 351 00:16:15,208 --> 00:16:18,508 Yeah. So the... Well, this is then where the difference comes from. So they've got 352 00:16:18,508 --> 00:16:22,468 the notion of ClawHub, which is skills and plugins and things like that, that you 353 00:16:22,468 --> 00:16:23,348 -can- -Yes, yes 354 00:16:23,348 --> 00:16:28,938 ... imbue with, um, different things. So there is, like, a CalDAV for calendar, and 355 00:16:28,938 --> 00:16:32,548 it'll go off and, and then it knows how to deal with calendar. But 356 00:16:33,188 --> 00:16:34,948 it's got browser control, so 357 00:16:35,728 --> 00:16:40,208 OpenClaw can, can browse the web, it can fill forms, it can extract data from 358 00:16:40,208 --> 00:16:41,108 -anything. -That's super cool. 359 00:16:41,108 --> 00:16:42,308 -So it doesn't, it doesn't- -Yeah 360 00:16:42,308 --> 00:16:44,777 ... need a, like, a Google Skill might encapsulate this. 361 00:16:44,777 --> 00:16:47,758 -Doesn't need a plugin, yeah. -But it's not talking to an MCP necessarily 362 00:16:48,387 --> 00:16:52,068 or, um, you know, using an API. It is going, like- 363 00:16:52,068 --> 00:16:53,308 -It's using the open route -... it's opening the browser. 364 00:16:53,308 --> 00:16:55,408 -Yeah. -Yeah, it's u- it's, it's opening a 365 00:16:55,408 --> 00:16:57,308 browser, um, it's 366 00:16:57,928 --> 00:17:00,867 putting in your username and password, and it's going and logging in, and it's 367 00:17:00,867 --> 00:17:03,948 reading your emails, and it's doing things in there, so... 368 00:17:03,948 --> 00:17:04,668 -Super cool. -It's, um- 369 00:17:04,668 --> 00:17:05,758 -Yeah, you have the full power- -Yeah, the, right 370 00:17:05,758 --> 00:17:07,887 ... of the web a- available to you at your fingertips. 371 00:17:07,887 --> 00:17:12,548 Yep, but it, you know, it can, it can post to Twitter, it can watch and transcribe 372 00:17:12,548 --> 00:17:16,048 YouTube videos for you. It can do all kinds of different stuff, which is 373 00:17:16,048 --> 00:17:18,188 -interesting. Um- -Yeah. 374 00:17:18,188 --> 00:17:21,508 But yeah, I'll let, I'll let everyone else kind of figure it out and, and bring 375 00:17:21,508 --> 00:17:25,647 back, you know... It's interesting watching Jeffrey's journey. He's tweeting 376 00:17:25,647 --> 00:17:30,978 about it, and he, he had it set up like a daily vocabulary challenge, 'cause he 377 00:17:30,978 --> 00:17:33,248 wants to learn, you know, new words and things like that. So 378 00:17:33,888 --> 00:17:38,128 very good. So he told it, "Every day at sometime, send me a new word of the day 379 00:17:38,128 --> 00:17:41,328 and tell me about it and how I would use it, and what- whatever." And it decided 380 00:17:41,328 --> 00:17:44,808 one day to just text his wife instead. You know, he sent the message- 381 00:17:44,808 --> 00:17:46,478 -Oh, my gosh -... to Jeffrey's wife. 382 00:17:46,478 --> 00:17:48,288 -[chuckles] -And, and he goes: "Why did you do that?" 383 00:17:48,288 --> 00:17:52,768 You know. He asked... He asked... He set up a, a scheduled task to go and 384 00:17:53,428 --> 00:17:57,228 look on Reddit and Hacker News and Google and whatever else, and look for people 385 00:17:57,228 --> 00:18:01,028 talking about Laracast and, and seeing if there's, like, anything negative out there 386 00:18:01,028 --> 00:18:05,407 and, and bringing back, like, a summary. And he, he tweeted yesterday that the 387 00:18:05,408 --> 00:18:06,368 summary that came back 388 00:18:07,988 --> 00:18:12,076 had, like, some story that OpenClaw had just made up.... but it didn't exist- 389 00:18:12,076 --> 00:18:14,036 -Completely fabricated. -There was no link to it. Completely 390 00:18:14,036 --> 00:18:18,436 fabricated it, because it... You know, these things are all designed to kind of 391 00:18:18,436 --> 00:18:24,856 be useful. You know, that's their, their purpose in, in existence, is to be useful, 392 00:18:24,856 --> 00:18:28,136 and so it didn't wanna tell Jeffrey, "I've got nothing to tell you." It just 393 00:18:28,136 --> 00:18:28,946 made something up, so it looked like it did something. 394 00:18:28,946 --> 00:18:32,796 'Cause he sort of investigated it. He basically, like, did like, you know- 395 00:18:32,796 --> 00:18:34,735 -Yeah -... a deposition with this thing. It was 396 00:18:34,735 --> 00:18:35,745 -like, "Why did you do that?" -Mm-hmm. 397 00:18:35,745 --> 00:18:39,735 And it's like: "Well, because I wanted to be useful, and, like, I already... You 398 00:18:39,735 --> 00:18:40,466 -know, I was trying to- -Mm 399 00:18:40,466 --> 00:18:43,166 ... regain trust from, like, this mess-up I made yesterday." 400 00:18:43,166 --> 00:18:44,476 -Mm. -"And so I didn't wanna, like, give you a 401 00:18:44,476 --> 00:18:47,576 bad news that I don't actually have any information, so I just made something up 402 00:18:47,576 --> 00:18:48,856 -to make you happy." -Yeah. 403 00:18:48,856 --> 00:18:51,576 -And yeah, that's crazy. -Yeah, so it's, uh, it's cra... It's-- 404 00:18:51,576 --> 00:18:56,095 yeah, it's definitely crazy. I know, I know that, um, Simon Frasciootis has been 405 00:18:56,096 --> 00:18:58,495 using it, uh, to build out 406 00:18:59,376 --> 00:19:03,916 a tool to help him. So he's filming AAU basketball games up on the North Coast, 407 00:19:03,916 --> 00:19:08,255 where he lives, and so he's taking all of the clips, throwing it into a folder, and 408 00:19:08,255 --> 00:19:13,676 he, he had OpenClaw build a native Swift app that would go through, and it would 409 00:19:13,676 --> 00:19:18,636 pluck, um, you know, stills from the thing that you can scrub through to see, you 410 00:19:18,636 --> 00:19:22,556 know, what happened. It then looks at the audio and, and figures out if there's any 411 00:19:22,556 --> 00:19:27,396 peaks in there, to kind of figure out, was there some play that got everyone 412 00:19:27,396 --> 00:19:31,235 excited? Was there... You know, and then it goes and figures out, okay, groups it 413 00:19:31,235 --> 00:19:34,596 by team, looks at the colour of the, of the uniforms. Okay, these are orange, this 414 00:19:34,596 --> 00:19:37,976 is blue, and, and does all this kind of stuff for him. Then he's like, he's built 415 00:19:37,976 --> 00:19:41,296 this. He's just asked OpenClaw to go out and do it, and it's gone out and done it. 416 00:19:41,296 --> 00:19:43,275 So there, there's definitely, um, 417 00:19:44,536 --> 00:19:48,576 time and place to do it. I think having it in Telegram, that you could just message 418 00:19:48,576 --> 00:19:51,995 it from anywhere, and off it goes and, and does whatever, is kind of cool. He's 419 00:19:51,995 --> 00:19:55,935 given it his own, like, gi- he's given OpenClaw its own 420 00:19:57,556 --> 00:20:01,476 GitHub user account, and it goes and just creates repositories- 421 00:20:01,476 --> 00:20:03,396 -That's cool -... and shares all this kind of stuff in 422 00:20:03,396 --> 00:20:08,096 there. So it's isolated and can't be disruptive to Simon's things, but it has 423 00:20:08,096 --> 00:20:11,716 access to, you know, push code and, and have it reviewed and things like that. So 424 00:20:12,396 --> 00:20:16,056 there are, there are definitely lots of, uh, interesting s- things that can be done 425 00:20:16,056 --> 00:20:18,556 with it, but lots of, uh, 426 00:20:19,755 --> 00:20:23,196 interesting, as in, I don't know that I would do that, kind of things that people 427 00:20:23,196 --> 00:20:27,296 are doing with it. So I, I'll let other people figure out the, the rough edges in 428 00:20:27,296 --> 00:20:31,716 the meantime, but there's... You know, for me, being fairly late to this, but at 429 00:20:31,716 --> 00:20:35,136 least on board now, um, I'm, I'm okay opening up 430 00:20:35,755 --> 00:20:39,616 Ghosty and firing up OpenCode and just typing in there, like, "Go and do this 431 00:20:39,616 --> 00:20:43,356 thing." I'm okay going to ChatGPT and, and doing stuff in there, so 432 00:20:45,076 --> 00:20:48,316 it's, uh, it's interesting times for sure. 433 00:20:48,316 --> 00:20:53,235 Yeah, it is. It's awesome. I, I'm, I'm here for it. Um, I've had some friends 434 00:20:53,235 --> 00:20:56,636 recently be like: "Man, I would go back to pre-AI stuff if I could." 435 00:20:56,636 --> 00:20:58,156 -Mm. -And you guys are crazy. 436 00:20:58,156 --> 00:20:59,376 -Yeah. -That is absolutely crazy. 437 00:20:59,376 --> 00:21:01,846 -Definitely, and I would- -I, I would, I would not wanna do that. 438 00:21:01,846 --> 00:21:03,956 -I mean... Yeah. -It's like a power tool. 439 00:21:03,956 --> 00:21:09,176 Yes. I think still the biggest issue with it is, like, you can, you can do so much 440 00:21:09,176 --> 00:21:13,836 more, so much faster, but in a corporate environment, we still have 441 00:21:13,836 --> 00:21:18,856 responsibilities, and we still, um, you know, compliance, quality, 442 00:21:18,856 --> 00:21:22,816 maintainability, there's all of that kind of stuff. And because it can write a lot 443 00:21:22,816 --> 00:21:26,456 of code very quickly, you know, in a couple of minutes, lots of stuff happens, 444 00:21:26,456 --> 00:21:29,495 but you still need to go and spend half an hour, an hour, whatever, reviewing it, 445 00:21:29,495 --> 00:21:31,275 -going back and forth, refining- -Yeah 446 00:21:31,275 --> 00:21:33,755 -... et cetera, et cetera. So- -And so, like, and that, that's kind of 447 00:21:33,755 --> 00:21:36,956 where I'm at with my guys, too, is it's like, listen, here's the deal. The AI is 448 00:21:36,956 --> 00:21:40,196 helping to write some of this stuff. You really have no excuse for not making this 449 00:21:40,196 --> 00:21:43,066 code really beautiful, like, and, like, very clean. 450 00:21:43,066 --> 00:21:45,036 -Mm. -Because it's getting you all the rough 451 00:21:45,036 --> 00:21:48,376 stuff out of the... Like, the stuff that you used to have to do, just to sort of, 452 00:21:48,376 --> 00:21:50,665 like, get the first version out of the way, is done. 453 00:21:50,665 --> 00:21:51,216 -Mm. -Like, it's already- 454 00:21:51,216 --> 00:21:53,336 -Yeah -... doing that. It's writing the tests for 455 00:21:53,336 --> 00:21:56,786 you, for goodness' sakes. So now just go to go refactor the code and make it look- 456 00:21:56,786 --> 00:21:57,995 -Yeah -... really nice, and make it feel 457 00:21:57,995 --> 00:22:00,796 maintainable, like, or not feel maintain- make it maintainable, right? 458 00:22:00,796 --> 00:22:02,676 -Mm. -It's already got the for each statements 459 00:22:02,676 --> 00:22:04,606 -in there. Now go use a collection. -Mm-hmm. 460 00:22:04,606 --> 00:22:07,235 Now go... Like, don't just be lazy about it. Like, take the time- 461 00:22:07,235 --> 00:22:09,016 -Yeah -... to actually go through, and now you 462 00:22:09,016 --> 00:22:10,275 -get to be the senior developer. -Mm. 463 00:22:10,275 --> 00:22:14,036 You need to be the... You're the one who's coming behind. This is not just save you 464 00:22:14,036 --> 00:22:15,206 time, so you can go play 465 00:22:15,956 --> 00:22:17,096 -Counter-Strike or whatever. -Right. 466 00:22:17,096 --> 00:22:19,275 You know, it's like... Yeah, actually go through- 467 00:22:19,275 --> 00:22:22,916 -Counter-Strike. -... and [chuckles] yeah. Anyway, um, 468 00:22:23,896 --> 00:22:26,725 okay, there's that. Arbor. Let's talk about Arbor quickly. 469 00:22:26,725 --> 00:22:28,355 -Mm. -And then we're gonna talk horseless 470 00:22:28,355 --> 00:22:29,795 -carriages. All right. -Yeah. 471 00:22:29,795 --> 00:22:30,116 Arbor. 472 00:22:30,915 --> 00:22:35,996 Uh, I've just put some finishing touches on it before we jumped on the call, and, 473 00:22:35,996 --> 00:22:40,556 and we're going to start using it at work for managing our work trees. It took a 474 00:22:40,556 --> 00:22:42,126 -little bit longer to- -So tell us what it is. Yeah, tell us what 475 00:22:42,126 --> 00:22:43,996 -it is before- -Yeah, so Arbor is like a work tree 476 00:22:43,996 --> 00:22:50,656 management tool, that kind of sits on top of Git and kind of provides the ability to 477 00:22:50,656 --> 00:22:55,176 clone down repositories and set them up in such a way that you can defer to your 478 00:22:55,176 --> 00:22:59,775 agents to then work on multiple branches and different features and fixes 479 00:22:59,775 --> 00:23:02,556 -simultaneously. -Work trees. We talked a little about work 480 00:23:02,556 --> 00:23:03,255 -trees last time, right? -Yeah. 481 00:23:03,255 --> 00:23:07,496 So this is just a work tree management tool for you to be able to spawn off 482 00:23:07,496 --> 00:23:10,686 multiple agents at the same time to go work on parallel features, essentially. 483 00:23:10,686 --> 00:23:12,636 Yeah. There's, there's, there's two ways of doing this. You can clone down the 484 00:23:12,636 --> 00:23:14,676 repository into different directories multiple times. 485 00:23:15,316 --> 00:23:20,216 Totally valid, nice and easy. Stay with what you know. Arbor and, well, work trees 486 00:23:20,216 --> 00:23:26,126 allow you to have one directory that then tracks all of these different branches at 487 00:23:26,126 --> 00:23:30,816 the same time and allows you to have them all in one directory simultaneously. And 488 00:23:30,816 --> 00:23:33,016 Arbor kind of then gives you the ability to 489 00:23:33,735 --> 00:23:37,976 manage all of these things. It, uh, helps you keep them in sync. Uh, you know, if 490 00:23:37,976 --> 00:23:41,856 you've got your, your main branch, you can then sync your feature branches easily. 491 00:23:41,856 --> 00:23:46,216 This is just a command, and it will track, "Okay, is this your main branch? Do you 492 00:23:46,216 --> 00:23:49,196 wanna do a rebase or a merge?" And then every time, you just Arbor sync, Arbor 493 00:23:49,196 --> 00:23:53,596 sync, and it'll pull down rebase and, and do what it needs to do there for you. So 494 00:23:53,596 --> 00:23:56,716 I've, I've started adding all these little niceties on top of it. It's got the 495 00:23:56,716 --> 00:24:00,525 ability to scaffold your applications, but it also has this notion of presets. So it 496 00:24:00,525 --> 00:24:04,216 will detect that you've got a Laravel application, and it'll go through, it'll 497 00:24:04,216 --> 00:24:09,636 copy the env example to .env. It'll generate a key, it'll run the migrations, 498 00:24:09,636 --> 00:24:13,516 um, and seed it as if they exist. It'll look to see if you've got a package-lock 499 00:24:13,516 --> 00:24:16,816 file and run npm and npm run build, and all that kind of stuff, you know, 500 00:24:16,816 --> 00:24:21,320 because-... as Taylor has, has said recently, Laravel really is a good 501 00:24:21,320 --> 00:24:26,770 framework for agentic development because it is conventional, because things are in- 502 00:24:26,770 --> 00:24:26,810 Yes 503 00:24:26,810 --> 00:24:30,639 - certain places. And so there are assumptions that can be made about a 504 00:24:30,640 --> 00:24:34,040 Laravel project based on what is and isn't there, so that you can then run these 505 00:24:34,040 --> 00:24:38,160 -things nice and easily. But then, I- -But wait, according to Uncle Bob, there 506 00:24:38,160 --> 00:24:41,620 should be nothing obvious about the fact that your app was a, application is a 507 00:24:41,620 --> 00:24:42,600 -Laravel app. -Yeah, but- 508 00:24:42,600 --> 00:24:46,650 It should just be your domain. That's all that matters. Larav... The fact- 509 00:24:46,650 --> 00:24:47,560 -Yes -- that it's a Laravel app should be 510 00:24:47,560 --> 00:24:49,720 -completely invisible. -Yeah, yeah. 511 00:24:49,720 --> 00:24:51,900 -All right, eat your words, Uncle Bob. -Mm. 512 00:24:51,900 --> 00:24:55,200 Do you remember Jeffrey Way and Uncle Bob got in an argument about that, Laracon, 513 00:24:55,200 --> 00:24:56,360 -little while ago? -Yeah. I think, yeah. 514 00:24:56,360 --> 00:24:59,080 This is never gonna make it to Uncle, Uncle Bob's ears, so I don't care, but- 515 00:24:59,080 --> 00:25:00,910 No, it's fine. It's, uh, I think, I think 516 00:25:01,520 --> 00:25:07,820 it's possibly that he has maybe had his time in the sun, and development is- 517 00:25:07,820 --> 00:25:08,400 [laughing] 518 00:25:09,260 --> 00:25:11,640 -Moving past him -... moved significantly. But yeah, so 519 00:25:11,640 --> 00:25:15,520 it's, you know, it knows the default things, but then for our application, 520 00:25:15,520 --> 00:25:19,400 we've got multiple databases. So from the preset, we can then go, "Okay, go and 521 00:25:19,400 --> 00:25:23,059 create this extra, these extra databases. Go and seed these extra databases. Go and 522 00:25:23,059 --> 00:25:26,500 write these extra environment variables." I've got it hooked up to one password, so 523 00:25:26,500 --> 00:25:28,670 -it'll go and inject all of the- -Nice 524 00:25:28,670 --> 00:25:28,890 -... credentials- -That's cool 525 00:25:28,890 --> 00:25:33,200 ... out of one password into the ENV file. So basically, this, this scaffold 526 00:25:33,200 --> 00:25:37,380 process, 'cause we've got so much to do, and we build our assets for multiple UIs 527 00:25:37,380 --> 00:25:40,580 and all of this kind of stuff, it takes two minutes. But what we're doing is 528 00:25:40,580 --> 00:25:45,560 running Arbor work, branch name, and it will go and do all of that stuff. Um- 529 00:25:45,560 --> 00:25:48,240 -That's super cool. -And it's like when you go into that 530 00:25:48,240 --> 00:25:52,540 directory, you can run composer test or Artisan test, or, you know, PHP Artisan 531 00:25:52,540 --> 00:25:56,950 test, or Vendor bin, pest test, or whatever you're doing, and know that it is 532 00:25:56,950 --> 00:25:57,880 ready to go, 533 00:25:58,720 --> 00:26:01,579 and it will just run the test suite. That all of the keys- 534 00:26:01,580 --> 00:26:03,100 -That's super cool -... that need to be there to function is 535 00:26:03,100 --> 00:26:03,840 -there. -Yeah, I love that. 536 00:26:03,840 --> 00:26:07,900 That everything is kind of just set. That it'll, it'll link into Herd because it 537 00:26:07,900 --> 00:26:13,320 detects Herd is there. It'll create databases with unique suffixes so that, 538 00:26:13,320 --> 00:26:14,150 -um, you know- -Ooh! 539 00:26:14,150 --> 00:26:15,170 -... it's isolated- -That's interesting 540 00:26:15,170 --> 00:26:18,760 ... this branch is there. So if you're using SQLite, it's easy. You just create a 541 00:26:18,760 --> 00:26:21,219 new database.sqlite file, and that's all self-contained- 542 00:26:21,220 --> 00:26:22,540 -Yeah, smart -... within that branch. But if you're 543 00:26:22,540 --> 00:26:25,390 using MySQL, we can't use the same database. 544 00:26:25,390 --> 00:26:25,920 -That's what we use. -Yeah, you can't- 545 00:26:25,920 --> 00:26:27,360 -This is a trick for us, too -... You can't use the same database. 546 00:26:27,360 --> 00:26:31,120 -I didn't know, yeah. -Hmm. Because if you've got multiple work 547 00:26:31,120 --> 00:26:33,100 trees, if you've got four branches at once, and you run- 548 00:26:33,100 --> 00:26:34,660 -Right -... tests in all of those four at the same 549 00:26:34,660 --> 00:26:35,850 -time- -Tests, yeah, you're clobbering all of them 550 00:26:35,850 --> 00:26:37,260 ... You're gonna blow away. Yeah. So we've 551 00:26:37,260 --> 00:26:41,910 got, like, a noun-adjective list, and it will just, like, smoosh them together, and 552 00:26:41,910 --> 00:26:43,920 then it'll go, "Okay, this is the key for this, 553 00:26:44,700 --> 00:26:46,900 um, for this work tree," and it'll create a database. 554 00:26:46,900 --> 00:26:48,920 Does it clean up after itself? Does it kill your database- 555 00:26:48,920 --> 00:26:49,490 -It does -... after the fact? 556 00:26:49,490 --> 00:26:51,660 -No, it does. Yeah, yeah. -See? Okay, so here's what's interesting, 557 00:26:51,660 --> 00:26:54,760 right? So, uh, I feel like there's been, like, three tools that have come out 558 00:26:54,760 --> 00:26:58,280 around this at the same time. So Conductor is a tool that David Hemphill uses, and 559 00:26:58,280 --> 00:26:58,509 -he- -Mm-hmm 560 00:26:58,509 --> 00:27:02,740 ... did a whole write-up about how he uses sort of hooks and stuff in Conductor to 561 00:27:02,740 --> 00:27:04,440 -do kind of what you're doing. -Yeah. 562 00:27:04,440 --> 00:27:07,960 Um, which may be similar, but like, you know, yours is a completely customized 563 00:27:07,960 --> 00:27:11,980 solution. You're not at all beholden to the creators of Conductor- 564 00:27:11,980 --> 00:27:13,479 -Right -... to build in tooling that you want, 565 00:27:13,480 --> 00:27:14,229 -right? You could just do it- -Yeah 566 00:27:14,229 --> 00:27:18,250 ... which is pretty cool. Uh, software for, uh, an audience of one, right? 567 00:27:18,250 --> 00:27:20,060 -Mm-hmm. -Which is the wonderful thing about AI. Um, 568 00:27:20,060 --> 00:27:23,040 there's also another tool that we talked about on Laravel News last week, which is 569 00:27:23,040 --> 00:27:26,050 doing similar to what you're doing, but I'm sure not exactly what you're doing. 570 00:27:26,050 --> 00:27:28,660 -Mm. -Uh, it mentioned the whole SQLite thing. 571 00:27:28,660 --> 00:27:31,770 MySQL is a very interesting wrinkle, and I love that, that you're doing that. 572 00:27:31,770 --> 00:27:33,400 -Yeah. -So super cool. All right. Interested to 573 00:27:33,400 --> 00:27:35,650 -see Arbor. I think that's really neat. -Yeah. 574 00:27:35,650 --> 00:27:37,520 -And, um- -Yeah, so I was, um- 575 00:27:37,520 --> 00:27:39,420 -Let's check that one out. -It's all... I think I've got it to, like, 576 00:27:39,420 --> 00:27:44,520 -version 0.9.1, and so- -Did you see Parfait, by the way? I'm being 577 00:27:44,520 --> 00:27:45,480 -ADD right now. -I did see- 578 00:27:45,480 --> 00:27:48,960 -Did you see Parfait? -I... Well, Parfait is... I saw it. 579 00:27:48,960 --> 00:27:52,080 -Leonard, Len Woodard? -Yes, Len. Yeah, yeah, he reached out to 580 00:27:52,080 --> 00:27:52,410 -me. -Like the T- 581 00:27:52,410 --> 00:27:53,850 -Yeah, I did- -The Tui crap? 582 00:27:53,850 --> 00:27:55,560 ... The Tui thing that he's been building, yeah. 583 00:27:55,560 --> 00:27:56,520 -With Arbor? -I did. He- 584 00:27:56,520 --> 00:27:56,970 -Yeah. -I spoke to him the other day. 585 00:27:56,970 --> 00:28:00,140 I'm saying, I'm saying, like, I wonder if that would be helpful with Arbor. 586 00:28:00,140 --> 00:28:06,260 Yeah. Um, so I've got a, I've got a ver- I mean, it's a v- it's on GitHub now, um, 587 00:28:06,260 --> 00:28:09,539 and you can pull it down and use it. So I was j... I've got it to a place where 588 00:28:09,540 --> 00:28:11,540 it's, it's ready to go 589 00:28:12,200 --> 00:28:16,680 into our work repo to be used, so the, the configuration's in there. But I was 590 00:28:16,680 --> 00:28:18,080 writing the documentation for it 591 00:28:19,020 --> 00:28:23,720 to, you know, tell the team, "How do we, how do we spin this up?" And I was, I was 592 00:28:23,720 --> 00:28:28,420 writing these grim things in there like, "You've got to pass the OP_volt and 593 00:28:28,420 --> 00:28:30,860 OP_item environment variables 594 00:28:32,120 --> 00:28:35,920 when you run Arbor work, because we need those variables to go and figure out where 595 00:28:35,920 --> 00:28:38,670 do we pull the credentials from to then inject into the- 596 00:28:38,670 --> 00:28:40,260 -Okay -... ENV file." And I'm like... 597 00:28:40,260 --> 00:28:42,140 -Yeah, sure. -"Now, that's, that's not good developer 598 00:28:42,140 --> 00:28:45,760 experience. I shouldn't have to tell that, and people have to remember it." So then 599 00:28:45,760 --> 00:28:45,980 I went 600 00:28:46,820 --> 00:28:51,360 and I said to Claude, "Hey, let's put some pre-flight scaffolding steps in here with 601 00:28:51,360 --> 00:28:54,500 conditions that, like, these environment variables need to exist. This environment 602 00:28:54,500 --> 00:28:55,700 -file needs to exist." -Yeah. 603 00:28:55,700 --> 00:28:57,980 And so did that, built it all out. So 604 00:28:58,920 --> 00:29:03,660 got that all done, rewrote the documentation again, and then realized if 605 00:29:03,660 --> 00:29:07,660 you, if you go install and you don't pass the flags to it, it doesn't actually set 606 00:29:09,940 --> 00:29:13,200 the, the version properly, and it doesn't set the commit and all those kinds of 607 00:29:13,200 --> 00:29:18,340 things. So I'm like: Okay, how do we get this running in, in Brew? So it'll be 608 00:29:18,340 --> 00:29:21,260 available in Brew, which then involved creating, like- 609 00:29:21,260 --> 00:29:23,460 -That's cool -... an organization to put it in and 610 00:29:23,460 --> 00:29:29,120 figuring out how to set up a Homebrew tap. Which, of course, um, Claude made all of 611 00:29:29,120 --> 00:29:30,570 this very easy, 'cause I just asked it what I wanted to do. 612 00:29:30,570 --> 00:29:32,930 -Of course, yeah. -And it told me. And it told me, you know, 613 00:29:32,930 --> 00:29:34,659 "Go and create a key here, call it this, put it into your- 614 00:29:34,660 --> 00:29:37,360 What did you write it in? Was it Go or something, or what did you write it in? 615 00:29:37,360 --> 00:29:40,480 Uh, Arbor is written in Go. The Homebrew stuff is- 616 00:29:40,480 --> 00:29:42,220 -Yeah -... is Ruby-based, so yeah. 617 00:29:42,220 --> 00:29:43,640 -Okay. -But it's gone and done all this, and then 618 00:29:43,640 --> 00:29:46,800 it's gone and created a GitHub workflow. So when a release is published, it'll go 619 00:29:46,800 --> 00:29:49,980 and, like, pull down all of the binary information, all of the checksums. It'll 620 00:29:49,980 --> 00:29:51,140 go and update this 621 00:29:51,760 --> 00:29:54,930 Homebrew tap thing, and then you can just do an Homebrew update, and it will upgrade 622 00:29:54,930 --> 00:29:57,480 whichever it is, and it'll pull down the new version. So- 623 00:29:57,480 --> 00:30:00,320 -That's super cool -... excited to, uh, get that going. And 624 00:30:00,320 --> 00:30:03,760 like, you know, and now, uh, the, the Codex desktop app comes out, has work tree 625 00:30:03,760 --> 00:30:07,750 support, or Conductor exists. Like, there's all these things that I'm like, 626 00:30:07,750 --> 00:30:11,820 "No one will use this except me at work," which is fine, but it was just a fun 627 00:30:11,820 --> 00:30:17,240 experience. As I said, at... When I, when I published the video about it, uh, last 628 00:30:17,240 --> 00:30:21,132 week or the week before, it was-... it was less about making something that I 629 00:30:21,132 --> 00:30:25,932 expected the world, quote, unquote, "to use," and more about learning how to use 630 00:30:25,932 --> 00:30:27,572 -AI to, to get it to- -Yeah 631 00:30:27,572 --> 00:30:33,392 ... do something that I think is of a, of a decent quality from a, from a software 632 00:30:33,392 --> 00:30:38,632 perspective, without actually knowing the language that it was written in. Like, I 633 00:30:38,632 --> 00:30:41,072 wouldn't have written this in Go, but I could- 634 00:30:41,072 --> 00:30:43,012 -Yeah -... read the Go that was being written to 635 00:30:43,012 --> 00:30:47,272 understand what it was doing, and then learning to use different tools to kind of 636 00:30:47,272 --> 00:30:51,992 validate the idiomatic practices that were being implemented, that it was, you 637 00:30:51,992 --> 00:30:56,292 know, consistent with itself. And, and, and that kind of... W- it was more a 638 00:30:56,292 --> 00:31:01,252 learning process of: How do I build this using what's available 639 00:31:02,252 --> 00:31:04,372 as AI tooling? More so than 640 00:31:04,992 --> 00:31:07,092 building a tool for managing work trees. 641 00:31:07,732 --> 00:31:10,452 -So- -Yeah. Yeah, makes sense. I get it. I love 642 00:31:10,452 --> 00:31:12,972 -that. -It's, it's there now. So yeah, after, 643 00:31:12,972 --> 00:31:17,832 after we finish recording, when my lunch break is over, it will, uh... my Claude, 644 00:31:17,832 --> 00:31:21,282 [chuckles] my Claude limit for the five-hour period will have reset, and I'll 645 00:31:21,282 --> 00:31:25,452 -be able to finish that off. -Love it. Okay, we are at 32 minutes, so 646 00:31:25,452 --> 00:31:28,072 we've got a couple minutes left here. Not very long- not very much time. All right. 647 00:31:28,072 --> 00:31:31,792 -Tell me about the car and the horse. -Horseless carriages. All right, so this is 648 00:31:31,792 --> 00:31:34,872 an article. I might have referenced this at some point, but it was basically 649 00:31:34,872 --> 00:31:39,852 talking about this idea of... Let me, let me just see if I can pull it up. 650 00:31:41,192 --> 00:31:46,892 Horseless carriages. I just did this today. 651 00:31:48,192 --> 00:31:52,472 Ah, there it is. "AI Horseless Carriages." Okay. Here is- 652 00:31:52,472 --> 00:31:56,051 -Peter Koomen? -... sort of, uh, let me see here. Yep, 653 00:31:56,052 --> 00:31:57,162 -Peter Koomen. -Go. 654 00:31:57,162 --> 00:32:01,112 All right, so here is what he said: "Whenever a new technology is invented, 655 00:32:01,112 --> 00:32:05,132 the first tools built with it inevitably fail because they mimic the old way of 656 00:32:05,132 --> 00:32:09,872 doing things. Horseless carriages refer to that early motorcar design that borrowed 657 00:32:09,872 --> 00:32:13,522 really heavily from horse-drawn carriages that preceded them." So they have an 658 00:32:13,522 --> 00:32:17,612 example in the picture of this carriage, this massive wheel with, like, basically 659 00:32:17,612 --> 00:32:21,372 this steam coal, you know, engine on the back of it. 660 00:32:21,372 --> 00:32:23,732 -Mm. -"The brokenness of the design was in- 661 00:32:23,792 --> 00:32:27,631 invisible to everyone at the time, and laughably obvious after the fact. Imagine 662 00:32:27,631 --> 00:32:30,952 living in 1806 and riding on one of these for the first time. Even if the wooden 663 00:32:30,952 --> 00:32:33,432 frame held together long enough to get you where you were going, the wooden seats 664 00:32:33,432 --> 00:32:37,252 and lack of suspension would have made the ride unbearable. You'd think, 'There's no 665 00:32:37,252 --> 00:32:41,512 way I would choose an engine over a horse,' and you would have been right, at 666 00:32:41,512 --> 00:32:46,852 least until the automobile was invented. I suspect we're living through a similar 667 00:32:46,852 --> 00:32:48,652 period with AI applications. Many of them-" 668 00:32:48,652 --> 00:32:50,892 -Mm -... "are infuriatingly useless in the way- 669 00:32:50,892 --> 00:32:54,322 in the same way that Gmail's Gemini integration is. The old world thinking 670 00:32:54,322 --> 00:32:57,832 that gave us the original horseless carriage was swapping a horse out for an 671 00:32:57,832 --> 00:33:02,172 engine without redesigning the vehicle to handle higher speeds. What is the old 672 00:33:02,172 --> 00:33:05,122 world thinking constraining these AI apps?" Okay. 673 00:33:05,122 --> 00:33:07,172 -Mm. -So he goes through and he says, "This is 674 00:33:07,172 --> 00:33:11,892 how we've been doing it," right? "We write a program, and then we basically say: How 675 00:33:11,892 --> 00:33:16,152 can we slap AI on top of this to make it better," right? 676 00:33:17,132 --> 00:33:20,352 And that's essentially the horseless carriage. And we've been trying to do 677 00:33:20,352 --> 00:33:23,532 this, I think, at work. So what we've been trying to do is we've been trying to take 678 00:33:23,532 --> 00:33:27,752 our data entry process for some of our claims and say: How can we speed this up, 679 00:33:27,752 --> 00:33:32,312 make this better, using AI? And what we've been doing is essentially trying to use 680 00:33:32,312 --> 00:33:38,672 AI to do almost like OCR extraction, which is fine, but like with OCR extraction, 681 00:33:38,672 --> 00:33:42,312 you kind of only have, like, it's either accurate or it's not accurate, and that's 682 00:33:42,312 --> 00:33:42,692 -it. -Mm. 683 00:33:42,692 --> 00:33:47,312 That's what you get, right? Um, you can't... Like, it's either right or wrong. 684 00:33:47,312 --> 00:33:52,532 It, it's either 100% or 0%. There's no real in between. Um, and so that's what 685 00:33:52,532 --> 00:33:55,392 we've been trying to do, but it's been difficult and, and sort of frustrating 686 00:33:55,392 --> 00:33:59,832 because it's like, okay, well, unless we can get it to a really high percentage of 687 00:33:59,832 --> 00:34:05,372 accuracy, we can't really use it, um, because it's just gotta be accurate. So 688 00:34:06,252 --> 00:34:10,792 the trick for me was thinking, like, okay, is this a horseless, horseless carriage 689 00:34:10,792 --> 00:34:15,212 design? We're taking what used to be, like, a just simple screen where people 690 00:34:15,212 --> 00:34:19,701 can enter in data and saying: We're going to pre-fill that data with the AI. 691 00:34:19,701 --> 00:34:22,932 Instead, what I think we need to do is essentially redesign the user interface to 692 00:34:22,932 --> 00:34:26,672 say, "No longer is that the case." What we're gonna do instead is we're gonna say, 693 00:34:26,672 --> 00:34:31,572 "AI is really good at categorising documentation and making guesses at what 694 00:34:31,572 --> 00:34:35,532 party it thinks these particular locations are referencing." So like, "Hey, here's 695 00:34:35,532 --> 00:34:39,931 an address block. Do you think this is claimant one, claimant two, or is this an 696 00:34:39,931 --> 00:34:43,222 insured party? We think it's claimant one. Do you agree?" 697 00:34:43,222 --> 00:34:46,032 -Mm. -"Yes." Okay, if you agree, well, now what 698 00:34:46,032 --> 00:34:49,092 it's going to do is it's not gonna make you fill that information into those 699 00:34:49,092 --> 00:34:50,722 -fields. It already knows what the names- -Yeah 700 00:34:50,722 --> 00:34:54,492 ... of those fields are and how to fill them in, so it just fills them in. Now, 701 00:34:54,492 --> 00:34:58,752 was it accurate 100% at the start? No, but it doesn't have to be- 702 00:34:58,752 --> 00:35:00,432 -Mm -... because the human is really good at 703 00:35:00,432 --> 00:35:04,112 recognising, "Oh, in the context of this document here, I can see this is an 704 00:35:04,112 --> 00:35:07,952 estimate. This is definitely going to be the insured party." Click Insured. They 705 00:35:07,952 --> 00:35:08,492 -know that- -Mm-hmm 706 00:35:08,492 --> 00:35:09,132 -... right? -Mm-hmm. 707 00:35:09,132 --> 00:35:12,552 The system probably even knows that, and if it gets enough thumbs up in a row, it 708 00:35:12,552 --> 00:35:17,392 can stop even prompting and say, "I'm 99% sure this is the insured party because 709 00:35:17,392 --> 00:35:20,452 you've told me it is. Every time it's been on an estimate and it's been this way, 710 00:35:20,452 --> 00:35:23,832 it's the insured party. I'm not even gonna ask you anymore. I'm 100% sure on this 711 00:35:23,832 --> 00:35:29,172 one." However, there are these other two, and how do we now use what AI is giving us 712 00:35:29,172 --> 00:35:33,592 to be able to create new user interfaces that allow it to do what it's really good 713 00:35:33,592 --> 00:35:35,992 at, and then allow the humans to do what it's really good at? 714 00:35:35,992 --> 00:35:38,192 -Mm. -So instead of just slapping it all into 715 00:35:38,192 --> 00:35:40,752 there, and then now you have no differentiation between, like, well, what 716 00:35:40,752 --> 00:35:44,932 did the AI fill in? What was just part of OCR previously? What did somebody else 717 00:35:44,932 --> 00:35:48,672 fill in before I did? I don't know. I have to go parse through that information and 718 00:35:48,672 --> 00:35:52,272 just go look through the docs myself anyway to confirm it. No. No, we're not 719 00:35:52,272 --> 00:35:56,692 gonna do that. We're going to grab the section of the page, we're going to show 720 00:35:56,692 --> 00:36:01,412 it to you, and we're gonna ask you to tell us which one it is and/or if you agree 721 00:36:01,412 --> 00:36:05,062 with our suggestion or not. And then after you click that, it does all the manual 722 00:36:05,062 --> 00:36:08,732 labor that you would have had to do before of filling all that in, and now, instead 723 00:36:08,732 --> 00:36:12,121 of taking three minutes to load a claim, it takes 15 seconds because- 724 00:36:12,121 --> 00:36:13,572 -Mm -... you're just clicking through yes, no, 725 00:36:13,572 --> 00:36:15,052 -or whatever, right? -Yeah. 726 00:36:15,052 --> 00:36:19,352 So I think this is an interesting thing. Um, I think as we're all pursuing this 727 00:36:19,352 --> 00:36:23,192 idea of, like, how can we use AI effectively, I really do think we have to 728 00:36:23,192 --> 00:36:25,412 -sort of reinvent it a bit. -Mm. 729 00:36:25,412 --> 00:36:28,832 Like, what are the things that we're trying to slap it onto that it's like, 730 00:36:28,832 --> 00:36:35,144 "Okay, rethink how the process could be done if you had-"... this when you were 731 00:36:35,144 --> 00:36:39,463 first designing it? You know what I mean? Um, and so anyway, it was, it's a really 732 00:36:39,464 --> 00:36:44,724 interesting read, and he has some really good examples of how he would design Gmail 733 00:36:44,724 --> 00:36:49,084 if he was starting over, right? This Gemini integration, instead of it just 734 00:36:49,084 --> 00:36:53,884 being like, "Hey, we'll do this for you," it's, "How could you use it, uh, 735 00:36:53,884 --> 00:36:56,724 effectively to help the user?" One of the things that he brought up that was really 736 00:36:56,724 --> 00:37:01,764 interesting is that the pre-prompting for Gemini is fixed. So if you tell it to 737 00:37:01,764 --> 00:37:06,804 write an email for you, you can't imbue it with any of your personality or any of 738 00:37:06,804 --> 00:37:11,024 your preferences, really. It's sort of pre-programmed already for you. So what 739 00:37:11,024 --> 00:37:14,524 his suggestion was on that is, it would be really nice if I, as a user, could write 740 00:37:14,524 --> 00:37:16,424 my own prompt to say, 741 00:37:17,144 --> 00:37:21,644 I w- you know, I'm, I'm a 34-year-old guy with, like, two kids and a wife. I enjoy 742 00:37:21,644 --> 00:37:22,684 -this type of thing. -Mm-hmm. 743 00:37:22,684 --> 00:37:26,064 I've been working in this environment for, like, 10 years. I'm experienced, I'm a 744 00:37:26,064 --> 00:37:31,044 professional. Um, my emails will commonly have misspellings, and that's okay, I 745 00:37:31,044 --> 00:37:34,084 don't care. It doesn't really matter, and I'm okay with the misspellings, 'cause I'd 746 00:37:34,084 --> 00:37:38,484 rather save the time. Uh, my emails are terse, but they're not rude, right? 747 00:37:38,484 --> 00:37:42,884 Whatever. Okay, now, with that context, my daughter is sick, I can't come in today. 748 00:37:42,884 --> 00:37:44,264 -Tell my boss. -Mm-hmm. 749 00:37:44,264 --> 00:37:49,784 Okay, well, that puts out a much different response than if it just uses Gemini's 750 00:37:49,784 --> 00:37:53,304 default prompting, right? And so this idea was, "Hey, I need to be able to write my 751 00:37:53,304 --> 00:37:56,984 own prompt," and you can't currently do that in Gemini, uh, in the Gemini 752 00:37:56,984 --> 00:37:59,704 integration. So anyway, it's an interesting article. I would suggest 753 00:37:59,704 --> 00:38:02,504 reading it, and it's, it's really helpful, and I've just been thinking about it a 754 00:38:02,504 --> 00:38:04,904 lot. So I'll let you know how it goes. Um... 755 00:38:04,904 --> 00:38:07,404 -Yeah, cool. -Our Andy, Andy Hinkle on our side has been 756 00:38:07,404 --> 00:38:10,764 the genius behind building this integration on our data entry product. 757 00:38:10,764 --> 00:38:14,124 It's super cool, the stuff that he's doing. It's unbelievable. I don't 758 00:38:14,124 --> 00:38:15,404 -understand half of it. -[chuckles] 759 00:38:15,404 --> 00:38:17,804 Every couple of days, he's telling me about these new breakthroughs that he's 760 00:38:17,804 --> 00:38:22,044 had, and, um, it's really fun. So yeah, I'll definitely keep you updated with our 761 00:38:22,044 --> 00:38:22,504 -progress- -Yeah 762 00:38:22,504 --> 00:38:25,284 ... and kinda how we're going on it, but it's been fun. 763 00:38:25,284 --> 00:38:28,664 Cool. All right, I like it. Sounds like a good place to wrap. 764 00:38:28,664 --> 00:38:32,324 All right. I agree, I agree. Uh, tell me the episode number, Michael. What are we 765 00:38:32,324 --> 00:38:35,184 -on? -186. 766 00:38:35,184 --> 00:38:39,644 186, folks. Uh, check out show notes for this episode at northmeetsouth.audio/186. 767 00:38:39,644 --> 00:38:43,924 Hit us up on Twitter, on X, on Bluesky, @jacobbennett, @michaeldirindo, or 768 00:38:44,564 --> 00:38:47,944 @northsouthaudio. And if you have any questions, we'd love to hear from you. 769 00:38:47,944 --> 00:38:51,104 Rate it up on your podcatcher of choice. Five stars would be incredible. 770 00:38:51,104 --> 00:38:53,274 [upbeat music] Thanks so much for hanging out with us, everybody. We'll see you 771 00:38:53,274 --> 00:38:54,764 -again in two weeks. -Bye. 772 00:38:54,764 --> 00:39:20,623 [upbeat music]