Is Prompting the Future of Coding? - Ep. 20 with Nick Dobos

[00:00] You built the number one programming GPT Grimoire, which is a coding assistant. It's like a wizard-based coding assistant. And you've done over a million chats, which is wild. Coffee website. I'm just going to hit N. Oh, your coffee theme website is back. Click our website and boom, there we go. Wow. That's really cool. [00:30] you [00:31] Nick, welcome to the show. [00:33] Hey, thanks for having me. [00:34] Excited to be here. [00:35] Yeah. Awesome to have you. [00:38] Because we've been Twitter mutuals for a really long time, and you're just constantly dropping cool stuff on Twitter, and you have this sort of mysterious... [00:47] mystique to you that I quite appreciate and it's fun to actually like put a face to the mystique. [00:55] Well, thank you. Yeah, you know, I have fun. I don't know. I just play around with it. I think I started working at Twitter and then I just got into posting on it and now I'm addicted. So. [01:05] having fun with it and people seem to enjoy all these random experiments I've been sharing and having fun with so [01:11] "Yeah, no, it's cool to meet you [01:12] Hang out with you. [01:14] Cause yeah, I've been, you know, you're writing for a long time with every and everything and [01:18] yeah i've got a chance to work with our mutual fan nathan so [01:21] Yeah, excited to be here and meet you. [01:24] Amazing. That's so awesome. So for people who don't know, you are a programmer. You built the number one programming GPT, Grimoire, which is a coding assistant. It's like a wizard-based coding assistant. And you've done over a million chats, which is wild. So I think one of the

[01:47] foremost experts in the world on building custom GPTs and doing that successfully and then also [01:55] using ChatGPT for programming. And I think that's sort of where I want to start. [02:01] Can you talk to me about like your introduction to AI for programming? Like how did you get into it? Yeah, I mean, I guess, you know, I've been programming for a long time, you know, I guess like for 11, almost 12 years now professionally. And so, you know, I've done a few small, I've done like a few small AI projects throughout the years, you know, mostly for like games and little things like that. Nothing too crazy. Yeah. [02:24] Um, [02:25] You know, I think [02:26] having done that, I realized that a lot of it is kind of just like very house of cards, smoke and mirrors. [02:32] But kind of with this newest wave, I had been working as a programmer, mainly doing iOS apps, [02:38] And I saw a GitHub copilot and just like, oh, that's a cool tool, you know, because like, [02:43] As you're programming, you have a little bit of like autocomplete and it kind of like works with just like the things on your phone where it's just finishing the sentence or [02:51] kind of guessing what's filling in. But yeah, I started playing around [02:55] with GitHub Goflare and was really just surprised at how well it works. [03:00] Do all these little tricks with it where you can even like [03:02] write a comment of just like what you want it to do and then it'll like finish like big sections. [03:08] So yeah, I started playing around with that. I think I had some [03:11] Lucky timing because I've been working at Twitter and then the whole Elon thing happened and that was a whole mess. [03:17] So I got laid off from that and then just had some time to play around and learn new things.

[03:22] So, ChatGPT came out in November that year, I think. [03:27] And really, I started playing around with it, having some fun with it. [03:30] And then, especially when GPT-4 came out, tried it that day and was like, "Hey, can I use this for programming?" [03:36] I had tried. [03:38] you know, I'm using it for [03:40] Using GitHub Quokka and kind of found a little bit of success with that and was really just blown away. [03:45] with how well it worked. You know, I was like, [03:48] immediately like, wow, if I had like asked, you know, a student or like someone who's learning, you know, computer science, like, [03:54] to do this, I feel like they would be at a college level or higher. So like, whoa, [03:58] How can I use this to start filling in pieces of things I'm working on? [04:03] So yeah, I started doing this ad and really just having fun with it, playing around with it. [04:07] So, [04:08] I think one thing that maybe separates me out from a lot of people who are already working in this space is that I really had not been focused on kind of [04:16] the big AI pushes happening beforehand. So, you know, I'm almost like a little bit clueless and catching up on that end, but [04:22] maybe having a little bit more fun from a front-end and artistic perspective and having a lot of fun [04:28] Yeah, just like exploring with it. [04:29] You know, I did my like coding wizard GPT and I'm just like, it's so silly. Like I picked the silliest branding ever for this thing. [04:37] And I don't know, people just seem to like it. You know, and honestly, just playing around with it. You know, seeing, can I turn this into a tool that's, [04:43] useful and entertaining and a little bit fun. [04:46] That's really cool. I love it. I mean, I do think just that spirit of exploration is like, it's the thing I love about AI stuff. And I think it's just the thing that

[04:55] the people that are doing the best and most interesting work in the space right now just have that sense like that curiosity and you definitely definitely have that and that's why i love i love seeing your stuff yeah and it's just so much fun to play around with so yeah i honestly can't even help it totally so i'm wondering like maybe you can just show us grimoire like i want to i want to see it in action i want i want you to show us show us how it works [05:19] Yeah, I mean, like, this whole thing kind of spawned out of a, like, I guess I had started playing around with, you know, kind of the custom instructions just as a way to kind of, you know, shape your, out of those just as a way to, like, shape the, [05:35] responses. [05:36] And I realized you could kind of format it in all these different ways and kind of like build hockey. So I kind of started playing around with ideas there. [05:42] And I'd also been [05:43] kind of, you know, as I was learning to do this and learning to, you know, kind of, how do I ask questions to like get code out of it? [05:51] like was coming up with like, oh, can I, you know, is there an easy way to do this? But I've also just been, you know, like trying to write articles, kind of explaining that. I was like. [05:59] when the GVT store came out, I was like, wait, can I just like shove those articles into a chat bot and like turn that into an interactive experience? [06:06] So it almost kind of turned out like somewhere between an assistant and maybe like [06:10] a tutorial or like a curriculum. [06:13] um [06:14] But honestly, I just kind of started throwing out little shortcuts, little tips and tricks. [06:19] as it's kind of gotten a little bit more popular and you, you know, like partner with, you know, people doing different things. [06:23] So like, [06:24] Yeah, basically it's a coding wizard.

[06:28] And I guess [06:29] Part of my thought here is that [06:31] where [06:31] I think that [06:33] LLMs have really created this weird new, I guess, [06:36] branch of programming. [06:38] Um, [06:39] that [06:40] I almost want to call it like a branch of abstract programming. I guess I've been referring to it as like prompt gramming. It feels like it's almost maybe a new form of programming languages we've had. [06:51] Things like binary, you know, like, [06:54] assembly language, like kind of low-level programming languages, you know, we've gotten into high-level programming languages. [07:00] And depending on how loosey you want to get with the definitions, even things like no code, [07:05] Or even just like typing and documenting on a computer, like is that programming? Like, I don't know, maybe you're kind of giving commands to a computer to shape it and make it do something. [07:15] So, [07:15] it seems like LMS are just like a way to do that. So like, [07:19] um [07:20] you know, there's [07:21] a whole bunch of different projects you can do, a whole bunch of different hotkeys you can use. [07:25] But one of the simple ones, which I just thought would be a fun one for most people to start with, is just doing a little link-in bio site. [07:32] So. [07:32] For example, I have this just one here. [07:35] really just like create sites in seconds. It's kind of crazy. [07:38] So I'll just click on this one. [07:40] But... [07:42] um yeah this should just pull out a little description from it [07:46] And basically we're going to make a website just like out of thin air. It's just, it still blows my mind that this is so, [07:52] easy and possible. It feels like [07:54] take a seed of an idea and expand it into [07:57] a full thing.

[07:59] That's interesting. So tell me about like, while it's building, I'm really into this idea of like, prompt gramming or just generally coding with AI being like a new layer of programming abstraction on top of scripting languages, which are built on top of like C and C++, which are built on top of assembly, and then assemblies built on top of binary. [08:29] looks like as a language and why you think it should count as a language. [08:35] Yeah, I guess, you know, it's definitely going into areas that I think are new and maybe some people say it shouldn't be a language or it should. Like, I don't really care much about how the labels are drawn as much as I just need some way to try and explain it. [08:50] the different layers, 'cause I don't know, it feels like this new area where, you know, I can just like give it a prompt in English, it'll spit out a bunch of code, [08:58] And I can then go use that code and either work that into my project, like run it in a code interpreter, like something like that. [09:04] um [09:05] So like, you know, here I've got this little... [09:07] thing kind of queued up and you know one of my favorite hotkeys i was actually able to partner with netlify and i was just [09:12] let me just stop you right there so i want to i want to go up and just go through what it did for for you um so basically you said link in bio create your website openpt1md [09:24] to Lincoln bio site. So, [09:27] That's the prompt that it was basically pre-filled. So tell me what that prompt means in the

[09:35] GPT setup that you have. Yeah, so maybe to like back it up a little bit. Here I was going to flip into a new conversation, you know, there's [09:42] so many ways I think you can structure GPTs and some of it's just like [09:46] you know, [09:47] giving it like a simple shape of just like, "Here, respond like you're a pirate," or, you know, "respond like you're super happy," or, you know, "respond as if you're like a writer," or something like that. [09:58] But I think some of the other ways is you can really [10:01] Um, [10:01] like. [10:03] I guess, predefined ideas and put them in. And something that I tried to do with this, that I think is way different from a lot of other people doing is kind of [10:11] giving it a template. So like I have a whole bunch of like kind of pre-built projects. So this link in BioSite is just like the conversation starter button. And I was just kind of triggering one of the projects. [10:22] But basically, you can, like, come up with any idea, like, I don't know, [10:25] I have a coffee next to me, so I'm just gonna be like, coffee website. [10:30] I was going to do that like two words and we'll just like see what comes out. [10:34] And so it's like, okay, I mean, how are we going to build a copy website? Let's figure out like the structure. [10:39] like the pieces of it, [10:41] Almost like [10:42] breaking the task into more parts and [10:46] Yeah, I guess. [10:47] expanding on that initial thing. [10:49] And so here, we're gonna make a homepage, we're gonna make a menu, we're gonna make a blog, we're gonna make a contact page. [10:56] and we're gonna figure out what to do, [10:58] And now there's a few options here. [11:00] And so [11:01] I'm just going to hit N because that's one of my favorite ones. Because you can just make websites instantly.

[11:06] What does N do? [11:08] So, yeah, N is one of the hotkeys. There's a whole bunch of hotkeys, which like, if I want, I can show you those, like, if you want to show those all like read out, you just like press K and I'll show you all of them. But this one works with Netlify and basically just like takes the code and instantly puts it online. [11:23] So you'll see in a second, we'll just have like a website just spit out just like that. [11:28] Wow, that's fascinating. So Netlify is sort of like a hosting provider, so it's going to write the code. Exactly. Netlify and then put it online. [11:37] Yeah. And so, you know, [11:38] The way it's set up is that website will kind of be ephemeral and will just like delete itself after one hour, or you can like claim it and like sign up with Netify and you know, they're. [11:47] They have like a hosting fee. So here, you know, we came back, oh, your coffee theme website is back. [11:52] Wow. [11:53] click our website and boom, there we go. You know, is this the most beautiful website in the world? No. Right. That's really cool. Right. But you know, we have [12:02] Just a thing. [12:03] You know a little thing set up you have our little menu. I guess we have three drinks in this coffee site and [12:09] I guess a little blog here, [12:12] And, you know, obviously this is not like... [12:15] you know, a billion dollar business website, right? But for something that took, you know, half a second and I did [12:21] two words and one button to like spit this out. [12:25] um [12:26] So yeah, I'm having lots of fun in just doing like what's like [12:30] you know, [12:31] small, simple projects I can do. Because right now we are definitely limited by kind of [12:35] how much

[12:36] output can come out of these and the context limits there. [12:39] Um, [12:40] But at the same time, it's really fun. And like, here's another fun trick. [12:44] Um, let's like... [12:46] make a QR code. [12:50] And let's see if we can get the, all right, no, you don't want a generator, stop it. [12:55] Okay. [12:56] Over here we will edit this, I guess. [12:59] All right. [13:00] for the website URL. [13:03] Let's see if this works. Normally it does. [13:08] So it's saying creating a QR code for the website URL is a practical... [13:13] Is it not right? [13:14] All right, let's see. [13:16] Here we go, let's implement this plane in Python to get this thing. Here we go, generates QR code. [13:23] Thank you. [13:24] Alright, now I will run it. There we go. [13:27] It had to think for a second, but yeah, this was the little URL that we just created. [13:32] So, I think that's a good question. [13:33] So yeah, double check that. Yeah, there we go. [13:36] And do we get it? [13:38] And then let's see. Okay, it's gonna make us download it. [13:42] Bring it down. [13:44] Renown the file. [13:45] Open it. [13:48] And boom, there's our QR code for the website we just made. And I can just go share that with anyone. [13:53] Like, I don't know. Is that like, I don't know if this makes any sense, but you can do all these like crazy little [13:59] I guess motions that seemed so difficult, like are now super easy with prompts. [14:03] No, that makes a ton of sense. And even just that little thing where it's now pushed live to a website, it just takes that little step of like, oh, I got to paste this into Replit or something like that and takes it away, which I really love. I think that's so cool. I want to go back to the link in bio thing because now that I sort of get the general gist of how this works, I want to understand like...

[14:27] So you have these pre-built projects inside of Grimoire, and one of them is this [14:34] is this link in bio thing. And link in bio, it sounds like you've created a link in bio project that lives inside of the GBT, like you uploaded it as maybe a file in the knowledge base. [14:47] And then when you ask for a link in bio site, it knows to take that project and that tells it what to do. Is that how it works? Yeah, exactly. You know, the basic idea here was that I wanted to, I guess, [14:58] I wanted to solve like the blank canvas problem. Cause I feel like some people are just like, okay, I have this chat thing. Like, what do I ask it? What do I do with it? [15:05] And especially with programming stuff, [15:08] There's just so many tutorials out there. [15:11] explaining a gazillion things are making it so difficult. And I was like, can I just give people a list of fun projects to do? [15:16] So I have this other one where if I just hit P, this should work. We'll see. It's been a little unreliable. [15:22] But [15:22] It should actually print out this chapters and parts. And so these are knowledge files that I uploaded as part of Grim R. [15:30] which is like, yeah, I guess [15:33] the same way that you know maybe your company has some pdfs and you want to go in and like read your pdfs or something and like ask your pdfs questions [15:41] But I decided to put... [15:43] like a curriculum of sorts or kind of a launching point of just like, okay, if you wanted to learn the program like zero to 100, what does that look like? [15:52] And I guess also final question is like, [15:54] I think [15:55] it's very clear to many people that, you know,

[15:57] LM's and GPT-4 and other things like it are immediately going to be like transforming coding. [16:05] But the question is like, [16:06] Do you like learn that first or like how does this get integrated? Because I think, you know, [16:11] As someone who's been doing it for a while, you know, I have quite a bit of, you know, [16:15] background knowledge around coding. So applying it to coding becomes much easier. [16:19] But. [16:20] If someone's brand new to coding, like what do we teach them? You know, how does that fit in? I think, you know, a lot of teachers are struggling with that. Like, I don't know if this is the correct answer. [16:29] But I was like, can I give someone a curriculum to go from, you know, [16:33] zero to a full programmer. Like, what does that look like? Wow. Okay. So, and so take us through this curriculum, uh, at a high level. Like it seems like it's written in a, almost like you're going to Hogwarts type way. [16:45] Yeah, exactly. You know, I really leaned into just making this whole like wizard theme and making it fun and [16:50] I think also I'm trying to have, just really just to make it more interesting to myself, [16:54] Oh no, I did a break. [16:55] network errors. [16:57] Um, [16:58] But... [16:59] Yeah, you know, because if so many of the dev tools out there just like robot themes, it's like, all right, we're going fantasy, we're going all the way and yeah, just [17:05] Trying to do it like that. [17:07] So yeah, you know, kind of, [17:09] have like a first part which is just like basic things like here like hello world and pong and just like [17:14] How does GitHub work? How does the command line work? Things like that. [17:19] But then also just like, I think moving into new territory where it's like, there's all these new creative tools springing up like, [17:25] from Midjourney to like Suno to like Runway. So you can kind of do like,

[17:31] prompt to all these different forms of media now. [17:34] Um, [17:35] You know, it seems like kind of [17:36] I guess like the first half of like the LOM, the like interpretability later, [17:41] layer of [17:43] being able to, you know, [17:44] into it what you're trying to ask from it and like build something out of it. [17:48] has really cracked open all these different media forms. [17:51] So yeah, I was just trying to be like, "Okay, can I collect all those into a cohesive way?" [17:57] There's kind of recommendations like that too. That's kind of this like, [18:00] Beginners incantations stuff. [18:02] Um, [18:03] You know, so then also [18:05] Also then, moving into more advanced programming stuff, [18:08] Then getting into these forbidden spells, this is kind of, I think, [18:13] maybe the most exciting like dev tool stuff for like if you're [18:18] a professional developer who's been doing this for a while and are now learning to use it with AI. [18:23] And I've been having a lot of fun using cursor. I don't know if you've ever used that app before. I have, yeah. [18:29] Yeah. [18:29] is so much fun. It is so crazy. Tell us about it for people who don't know what it is. Yeah. So Cursor, it's basically an app that you use to write code. It might be an IDE or it might just be a text editor, depending on how you define it. [18:43] but it's like a branch of VS code, which is a very popular [18:49] um [18:49] you know, tool put out by Microsoft. So like it supports like basically all sorts of programming. [18:55] and it's kind of built with like AI first. [18:57] and [18:58] Yeah, it's cool because it like uses ChatGPT, but it kind of flows it into like a programmer flow.

[19:05] very well. So I think I split my time between Grimoire and [19:09] cursor basically like for my day job. [19:12] um [19:12] So yeah, I'm having some fun there and just like trying to teach people how to use that. I think there's so much interest with everyone learning all these new tools and figuring out how to use them. [19:21] So yeah, just trying to share what I learned there. [19:23] And that's like, so give me an example of like in the forbidden spells section, like what? [19:30] you know, if you're teaching someone about cursor, like what are the kinds of things that I'm going to be taught? [19:34] Yeah, so here, let me like do this new thing. So I think it freaked out when [19:38] there was a network error, but if I just do a new chat, [19:41] I can go-- let's see. [19:44] Give a look at... [19:46] Wait, was it part four? I forgot. [19:49] Let me actually just print out the projects. [19:54] And... [19:56] this should show all the different things. But yeah, for the cursor one, I'm really just trying to show people how the app works because [20:02] There isn't even really a good tutorial, I feel like, even on their website. Like, it's just so new and they're changing things and building things. [20:09] Um, [20:10] that even just like [20:12] you know, teaching people like, oh, like, here's the command K hotkey, you know, like, how do you use that? You know, how can you [20:19] um [20:20] you know, experiment with ways to kind of like put multiple prompts together [20:24] and yeah, kind of put things [20:27] Yeah, I guess use it together in like a coherent flow. [20:31] Got it. [20:33] And so...

[20:35] Basically, so forbidden spells is sort of using things like cursor, but it seems like there's a lot of... [20:43] things going on here. Like you're teaching people a lot of [20:46] new concepts from like making simple games to doing back end stuff to learning how to use databases. Like this is a fairly comprehensive guide. How many people are using this to learn how to how to code? Do you know? I'm honestly not sure. Like I wish ChatGPT had better stats. Like I'm trying to figure out like, [21:06] how often people are using it. And part of what a product I'm working on now is actually kind of building it into a standalone app, so I can have kind of better stats that way and you understand what people are using it for. But yeah, I was, [21:18] Part of this was like, can I... [21:21] almost like, can I, um, [21:23] I guess going back to some of the Harry Potter like kind of language, like, can I like build myself a horcrux? Can I like take a big part of my like expert knowledge I've like built up over the years and like, you know, turn that into a course, give that back to the community. [21:37] and do that in a weird interactive chatbot. [21:40] So I definitely even have sections that are [21:44] basically just me trying to summarize an entire college curriculum [21:49] um [21:50] of computer science into a few chapters. [21:54] but [21:55] It's interesting because I don't even have to actually write the whole thing because of ChatGPT. I can just kind of like give it a seed to go off of. [22:02] You know, I can kind of be like, [22:03] teach them this, this, this, and this, and I don't have to do anything else, right? The machine will just do the rest.

[22:08] Um, [22:09] And yeah, it's going to give people a starting point to riff off of. [22:12] So yeah, it's fascinating. [22:15] That is really fascinating. What else should people know about coding with grimoire? Like what else? [22:22] What else can they do? What else can they build? What kind of what style of building is it most useful for? [22:31] - Yeah, I mean, I definitely set this up with like a curriculum and like some things to like get people started, but at the same time, I also wanted to build it [22:39] to be a tool that you can use just for general programming questions. So, you know, the way I use it lots of times is, [22:46] I'll just be working on a problem and I'll just have a quick question like, "Oh, this thing isn't working." [22:52] Like, you know, how do I make this text like underlined? You know, sometimes there's even simple questions like that. [22:59] Um, [23:00] that you know sometimes just like oh being able to look things up really quickly [23:04] But also, I think more and more I'm finding ways it's [23:08] you can kind of [23:10] use it to bridge gaps between ideas where if you're like, okay, I want something to like do this. [23:16] and then use it to break down your problem into smaller steps. [23:20] Um... [23:21] So yeah, I'm not sure if that totally answers your question. No, it does. I'm curious, like if you compare that to like copy pasting the same question into like vanilla Chachabut tea, like what what's the difference in your mind? Like, why is it better? Yeah, so some of it's just like the way I've written system prompt, like.

[23:41] you know, [23:42] A lot of it is just like [23:43] Some of it is like to actually do more planning. You'll notice like, even if I do like a simple thing, like, [23:51] Now do I-- [23:52] make a game in Swift. Like I'm an iOS programmer, so I do a lot of stuff in Swift. [23:58] Um, [23:59] But [24:01] I have a little silly intro and stuff like that, so some of it's a tiny bit of flavor, but really I'm trying to [24:08] bump it into being more truthful, more helpful, more thoughtful. [24:12] And so some of that, like there's actually some like planning steps that I'm trying to lean it towards. [24:17] um [24:18] which [24:19] is interesting 'cause I'm not sure if LLMs are going like, [24:24] you have to kind of nudge LLMs to like get the best out of them in a way it seems like. [24:28] You can't just say, like, give me this thing. You really need to almost, like, [24:34] queue it up and kind of aim it first. [24:37] And so part of that is like, you know, telling it, oh, you are like, you are an expert writer. Like now write me, you know, a thing in this style, as opposed to just like write me a thing. [24:47] uh [24:48] So some of it's like that, but the other thing is like, I really found myself just like, [24:52] not wanting to type a lot, so I made all these hotkeys to just do things quickly for me. [24:56] um [24:57] So, you know, like one thing that I found I would do a lot is I would like [25:02] You know. [25:03] kind of chat through some questions, [25:05] and have some code. And so I was just like, okay, [25:08] Can I just like save that code? So like here's one, if I just press Z, right? I can then take these files when I like just automatically get like zipped up and just download for me.

[25:18] Um, [25:19] That's super cool. I love that. I actually have not seen that as a design pattern that frequently, but it's kind of genius because it's so true. There's so many times where you're asking it to do the same kind of thing all the time, and it should probably know what your follow-up question is going to be. One for me is when it only prints out part of the file that I need to change, and I [25:42] hey, can you reprint the entire file for me so I can just copy paste it? Stuff like that, just having a little hotkey for it might be really useful. Yeah. And so I think that's kind of where some of the ideas started from. I started kind of putting these in the... [25:57] like custom instructions, and I was like, "Okay, wait, can I just like make a full suite of hotkeys for like programming?" So yeah, I have like, I think like 20 in there now. [26:05] for all sorts of different things. [26:07] It's kind of neat because it's also trying to guess like what keys you're going to use. [26:13] which I think also I've noticed has this [26:16] kind of strange [26:18] Like, you know, kind of example here, right? You know, it started spitting out, answering my question about like my little coding thing. [26:23] Um, [26:24] But kind of near the end, you know, it's like, okay, do you want to do this? Do you want to do this? Do you want to do this? [26:29] It almost like... [26:30] reels the thought back in a little bit and kind of puts you back on like a [26:36] like a sane path. And I think it does this interesting thing where it maybe keeps you from going off the rails a little bit. [26:42] Because... [26:43] There's definitely... [26:44] like... [26:46] if you

[26:48] like get weird responses from it. And he's like, keep going down. Like you can go down a rabbit hole, just like weird answers that make no sense. And like kind of run in circles. [26:56] So, [26:57] almost in a way it kind of brings it back a little bit. Like, [27:01] and kind of reels the thought back in, I think. And it's just an interesting way to structure the prompts. [27:06] Also in a way like reminds it that it's like still programming and that we want to be programming. [27:12] That's really interesting. I've definitely found that too. I haven't done it with sort of these like hot key footers. I really want to try that though. But what I found is having it output a plan at the beginning, like you were talking about, where it's just like, first, I'm going to do this, then I'm going to do this, then I'm going to do that is like a really nice way to get it to like sort of stay on the rails. But yeah, it seems like the hot keys like add something new. [27:42] is like that planning [27:44] And [27:45] I think sometimes the programming questions I ask, just being able to switch it into practical steps of like, [27:53] I need to do this, I need to do this. Like, oh, I don't like, you know, step three and four, I'm going to do this instead. [27:58] it's weird because you can kind of almost like [28:02] Yeah, I think... [28:03] plan out [28:05] I find it really useful as a planning partner that way, I suppose. [28:08] Yeah, that's really cool. I want to get into a little bit of like... [28:12] Um, [28:13] your thoughts on building gbt's because as a person obviously like you've built something that's super successful there's very few people that have built a gbt that's reached this level of scale like tell us a little bit about like what you've learned so far

[28:31] Yeah, I mean, I am honestly like still trying to figure it out. It's very much an experiment, but I [28:39] I think people are vastly underestimating GPTs and especially what you can do with Code Interpreter. [28:45] like there that thing unlocks so much crazy stuff [28:49] And... [28:49] I'm like, I guess a little surprised by the GPT store and kind of the GPTs that are getting featured. [28:56] Because I feel like [28:57] we can go way fancier and way crazier. And a lot of them just feel like, they're not as interesting as I would have hoped. [29:04] Um, [29:05] But [29:06] Yeah, I think just code interpreter in general unlocks like a crazy amount of stuff. We're just like, yes, like spin scripts out of nowhere and just like do all these kind [29:15] bizarre things. I'm really waiting for someone to [29:18] make like a [29:20] one that uses a sprite sheet and actually pulls images out. I've been trying to figure out if there's a good way to do that. [29:25] Because I know it's possible. You can do some really fun things with it there. [29:28] But... [29:29] Yeah, I guess I'm also-- [29:32] trying to figure out where OpenAI is going with this. [29:35] Because... [29:36] I don't know if it's been terribly successful for them in the way that they wanted, like, [29:41] if it's driving growth for them. And I guess I feel like that's kind of their main goal. [29:45] But [29:47] It has definitely expanded the scope of what they can do in a big way. [29:51] and also just connected it to so many [29:54] different things. [29:55] which I think is going to be a bigger and bigger part where [29:58] You know, maybe there's like multiple steps of multiple GPT's involved.

[30:03] But... [30:04] Yeah, it seems... [30:05] I guess, what else do you want to know about building GPTs? I'm just playing around with them, figuring it out. [30:12] Well, let's say like I'm [30:14] I'm a person who's like... [30:15] I kind of know what it is to build a custom custom GPT like to build my own custom version but like [30:22] when you think about what makes a problem or an idea for a custom GPT a good idea maybe we could start there like how do you know whether it should be a custom GPT or not? Yeah honestly I don't know I feel like I'm just guessing I think there's definitely [30:39] ones that I'm building for myself and ones I'm building for others. I think grimoire especially is one that I built with others in mind in a big way where I'm like, [30:48] kind of giving this tutorial, giving this curriculum, and with the idea that people are [30:52] who might not know programming might be using it. [30:55] Whereas some of the other ones I think that I'm playing around with are much more personal. [31:00] kind of playing out with like this one [31:03] to, I guess, like understand my emotions better, like kind of as a self journal in a way where it's like, [31:08] If I write a journal to this, can it interpret my emotions in a way? [31:12] um [31:13] But I think in a lot of ways, [31:16] the programming. Yeah, I guess some of this just figuring out what GPT is even good at. I think that's what is interesting about the GPT store too is like, [31:22] seeing some of these big categories of all of like, [31:25] Research is definitely one of them. [31:27] I think, and just being able to interpret things and learn things from big swaths of data.

[31:33] I think some of the scientific research ones are really cool. [31:38] Um, [31:39] But yeah, I'm also like, where are all the games? Like, where are all the crazy, wacky ones? [31:43] um [31:44] I feel like, yeah, there's so many weird code interpreter ones. [31:48] And I guess it's really just experimenting around and playing with things. [31:51] At this point, it's so new that no one really knows what we're doing with these. [31:55] And, you know, even like open eye themselves, like, you know, [32:00] after they have versions that are more powerful that they're playing with internally before they release them, [32:06] But... [32:07] you know, I don't say that like they are the best in the world at driving them either. Right? Like, [32:12] If you think about [32:13] a video game, for example, like, [32:15] you know, say like you are the game developer behind StarCraft, right? [32:19] that doesn't mean you're the best Starcraft player in the world. So what does it mean to drive these AIs to pilot them? [32:28] And [32:29] you know, how can you then use [32:31] kind of [32:32] prompts to steer them because it feels like [32:35] in many ways the prompts are kind of [32:37] taking like its big open space and trying to constrain it into a more narrow [32:42] and structured space [32:45] um [32:46] Yeah, and I think... [32:47] Maybe some of this comes from [32:49] under like [32:51] understanding how [32:52] generative and kind of random experiences happen in games as well. [32:56] I've definitely been a gamer my whole life, [32:59] built like a little bit games and you don't really do it for a living but [33:02] It's something I always think about and

[33:04] One thing that I've always really found fascinating with [33:07] games that have a lot of randomness in them, [33:09] which is kind of what ChatGPT is, right? It's like a statistical randomness of like, [33:13] all this internet data cobbled up [33:16] is [33:16] It's really hard to design an output for that. [33:20] If you have a game that's random, maybe you're playing... [33:24] Diablo or something, and you're generating a [33:27] like dungeon map to explore. [33:29] How do you actually know that that dungeon map is interesting and fun to run around in? [33:34] like you kind of it's really hard to do like as a game designer like how do you ensure that [33:39] when you're queuing people up to walk through a random something that's going to work um [33:45] So, [33:46] Yes, I think so. Some of it's like, [33:48] trying to decipher and [33:50] understand [33:52] where people are going with these. Are they going to walk down the spooky hallway? What happens if a ghost jumps out? I don't know. [34:00] But it's like, yeah, kind of understanding is like, [34:02] what are people doing with these and how do we [34:05] focus and narrow them towards like what we want to do. [34:08] Yeah. And I'm curious, like what you've learned about, like, for example, figuring out how to build the system prompt. [34:15] for a particular GPT, like what are the things, what are the things that you've learned that have been useful for you? Yeah, really a lot of it is experimentation and just trying things. You know, I know some people, and especially I think the more [34:28] Technically, my people are really into like doing evals and testing and stuff like that. [34:32] I think a lot of it's really just like sitting down and fiddling with it.

[34:35] The way I approach building these is I kind of [34:38] I have an idea, I write a system prompt, kind of just defining loosely what I wanted to do. [34:44] And then I kind of just [34:46] Ask it questions, see what comes back. [34:48] You know, just trying it over and over again, you know, when I think, [34:53] When the GPT-S4 first came out, I was hitting my cap on that thing like four or five times a day, just testing Grimoire, trying to run through different things. [35:01] Wait. [35:02] I was doing as much as I can because I was just so curious and having fun with it. [35:06] um [35:07] And some of it's just [35:09] learning [35:10] like how to prod these things. [35:13] and how to direct them. And it's interesting. Some of it's like [35:18] learning kind of how they were built. [35:21] there's definitely been insights that have come from that and kind of digging into the [35:26] you know, [35:26] the theories of how these things are built. [35:29] Um, [35:30] But I think it's also just [35:31] trying to kind of in this weird way [35:34] Um, [35:35] understand the zeitgeist of the world, really. Like, almost like you're trying to [35:42] Like you're trying to play Family Feud instead of Jeopardy, where it's like, [35:46] survey said like a hundred people think this way. So like, this is the way it's going to behave. And you're like, how do I like suss out that behavior and, [35:54] directed towards that. [35:56] It's yeah, it's they're quite bizarre, but quite fun. And you said like, you're really you don't feel like people are really pushing it to its limits. And in particular, there's a lot of stuff to be done with code interpreter. Can you can you tell me more about that? Like, what do you think?

[36:11] is possible now. [36:13] that's exciting to you that you don't think people have picked up on yet? Yeah, I mean, I think just in general, people are completely underrating this technology in a lot of ways. And there's certainly limits. I'm not. [36:24] I'm not saying that this is the perfect thing and there's no problems with it, but [36:28] you know, there's being able to like do this quickly [36:31] unlocks so much for so many people too because I think [36:35] Um, [36:37] It's going to massively lower the barrier to how [36:40] difficult it is to program a computer. [36:43] right like [36:44] Um, [36:45] You know, if my mom can go, I want a website that does this and it spits out a website like does that, you know, with her barely even doing it, you know, I think that's incredible and unlocks like. [36:56] makes it easier for people to do things, makes it faster for people to do things, [36:59] But I think part of that is also just [37:01] unlocks totally new use cases and really transforms how [37:05] we're building software too, especially on big teams. Like if things are going to be so crazy, [37:11] like Facebook had this like crazy paper come out where they're like, [37:14] generating test cases for like massive code bases, like millions of lines of code that they're spitting out and generating. [37:21] And I think even they put it like 70% of them into production now, it just has a way to like [37:26] test and like, [37:27] harden their code base. [37:29] Um, [37:30] So, [37:31] Yeah, I think this is gonna be something where I think it's gonna take like 20 years for us to really hit the limit of prompting and [37:38] kind of [37:38] what we can pull out of these models too. While we're still improving on them at the same time,

[37:44] and seeing how they do it, how they change. [37:47] And [37:48] Yeah, I think something that clicked for me really early on that I think maybe [37:52] not a lot of other people have grasped as like, [37:54] just how much we can take it and transform it into new actions. I think [38:01] I'm really fascinated with the way that [38:05] some of these new projects, [38:08] are kind of going because it seems like ChatGPT is really working towards like a tool using model. [38:14] I also really liked the way that [38:18] I've seen the little rabbits, the little like orange AI [38:22] like boxes. [38:23] Oh, I got hearts. [38:27] But, um, [38:28] I really liked how they are talking about what they're doing. Like I bought one just because of the way they were talking about it. [38:34] because they're really talking about large action models instead of large language models, where it's like, [38:39] Can I just like, [38:41] say something and then turn that into like a command that rips through other systems. [38:46] So the basic thing is like, [38:49] Can I, you know, [38:50] like make it like buy me a pizza or like book me an uber or you know fill out a spreadsheet [38:55] And once we, you know, I think something that also, it's so crazy that we can, [39:00] kind of transform text in this kind of structured, sensible way. [39:04] And that unlocks so many, I think, interesting use cases [39:08] and also just ways to build apps that are kind of [39:11] So, [39:12] fuzzy and unstructured and a little bit more abstract.

[39:16] So, yeah, I'm really excited to see what people do with it. I think the sky's the limit. That's really cool. So I guess you sort of think that Code Interpreter is... [39:24] I mean, one, it just has a ton of interesting use cases by itself, but also as one example of AI using tools in general, it seems like you're excited about the ability for language models to actually start to control different parts of your computer or different parts of your phone, all that kind of stuff. And once that happens, there's a lot more opportunity to build interesting things. Yeah, absolutely. I think that's going to be a huge portion. I think that's kind of where it seems like Apple's going. [39:54] out recently or like [39:55] talking about models that can see really fine details of where UIs are and how to click on things. [40:02] Yeah, I'm almost like, [40:04] Can we build this like new layer on like top of existing apps and internet? [40:10] Because I guess like, [40:11] know i've been talking a lot about how it's useful for coding but it's also really useful for using other apps that are just complicated [40:18] Right? So like one example that I ran into was I was trying to like edit a video and you know, it's an app that I was using, like couldn't do the one effect I wanted. I wanted like this fancy reverse thing. [40:27] Um, [40:28] And so I have, you know, I have Photoshop and like the Adobe suite. So I was like, okay, I'm gonna try using After Effects. I hate After Effects. I have no clue what I'm doing in that app. It's really complicated. Like. [40:37] He does a big video editor. I was like, okay, what if I just asked chatGBT how to do it? And it's like, spit out a perfect tutorial of like, okay, to do it, you just go click here, click here, click here. It took me like five minutes and...

[40:48] Like if I try [40:50] Like if I tried to find a tutorial, like it would not have been that good. Even if I found an exact tutorial, like do this thing. [40:57] And so it was like, [40:58] it makes it just so much easier. And all these different like ways, this just got sped up. [41:04] So, [41:05] Yeah, I'm almost like, can we build a layer that like operates on top of all of this that kind of drives other apps? [41:11] There's some really cool projects like Open Interpreter, [41:14] you have to kind of work more into your system and like, [41:17] can generate scripts that will open your email, you know, and like kind of control your computer in these really interesting ways. [41:24] So yeah, I think it's a really open... [41:26] a new way to communicate with a computer in this [41:29] weird and abstract and kind of fuzzy way. [41:32] which I don't know if they totally solved [41:35] interpretability and like, more linguistic programming, like, again, like, I don't feel like I'm enough of an expert on [41:41] some of these like deep AI topics to really comment on that. They're just like, it seems like it works and you can do fun stuff with it. So I'm going to have fun playing with it. I feel the same way. I just want to play around with it. And I sort of, yeah, I think there's this like, on the one hand, there's this thing you said earlier about it's sort of a higher level programming language that's above scripting languages, which, you know, it's the next layer of the programming stack.

[42:11] you know, [42:12] a new way to interact with computers overall. And I think that's probably [42:21] They're probably maybe even the same point to some degree because like the programming language that you're using to program with like prompting is just natural language and they're both the out both being able to program and being able to use a computer are both the output of being able to use natural language and have the computer know what you're talking about. [42:43] which is just a new interface. Yeah, and that is so fascinating because it's like a lot of what programming is, is yeah, figuring out how to take these like vague ideas of like, what should it be and like making it concrete and giving it actual form. [43:00] And so, yeah, kind of, you know, just like, [43:03] automatically do that is so powerful and really interesting in that way. [43:08] So... [43:08] Yeah, it's cool stuff. [43:10] I know you said earlier that you built a GPT that you use to kind of unpack your emotions. Would you be willing to show us that? You don't have to show us historical chats, but I'm kind of curious to see what you did. [43:22] - Yeah, so yeah, I can pull it off. I need a better name for it. I'm just calling it like the Emotion Shaman. I've been on with so many silly fantasy names. [43:31] I don't know if you know this, but I teach this course with a clinical psychologist, Dr. Gina Gorlin, called Maximize Your Mind with Chachibit. And a lot of it is this kind of thing. So I'm like a huge nerd for this. Yeah, I would love to nerd out with you on this because it's so interesting. Yeah, I guess the main thought with this one, and maybe I'll even just like open someone up and like show you how I did it.

[43:56] Um, [43:56] Was like... [43:58] To me it was like, [43:59] I am definitely a person who I like, [44:03] I suppress my emotions a lot. And I definitely like when I'm stressed, I do like of the like, fight, flight, or freeze. Like I like freezing and like hiding. And so I'm like, my stress response is usually like suppressing things and going away from things. So I'm like, [44:19] I'm not, I'm bad at interpreting my own emotions. So I was like, [44:24] So like lots of times when I'm like in therapy, talking to therapists, [44:28] The way that they help me is like, [44:30] helping me like even just like name and identify my emotions. [44:35] like. [44:36] So I'm like, am I angry? Like, I don't know. [44:38] And so like, and even things just like, [44:43] you know and obviously there's other ways to do this like a lot of times i'll like journal it out and like think about it that way it was like [44:48] Wait, can I like... [44:49] have the journal talk back to me. [44:51] So here's one that I made and really all I did with this was like, you know, I gave it some basic constructions of like, [44:57] basically just trying to give it ideas of like, you are like related to therapy and just like, basically just listed as many therapies as I could think of. [45:05] And it's like listed tons and tons of different emotions. Like I found a bunch of different like, [45:10] the motion wheels and stuff. [45:12] I'm just like, [45:13] trying to nudge it towards the direction [45:16] um [45:17] you know, of these different emotions so that like when I'm going through something, you know, I can be like, [45:22] You know, like, I don't feel great today, like blah, blah, blah, like someone so happened, you know.

[45:27] Like... [45:28] I don't know, I didn't sleep well or something. I don't know, this one's even super minor, but [45:33] Um... [45:35] Sleep well, like I am kind of crabby. [45:39] Um, [45:42] And so just as a way to like, [45:44] almost echo my thoughts back at me in a way. [45:48] Um, [45:48] Because I guess... [45:50] I find it interesting that I can use a machine to kind of like [45:54] twist my thoughts in a way and kind of reframe them back towards me. [46:00] I'm curious, like, [46:01] If that's a way to [46:03] you know, think differently and see [46:07] my life differently, I guess. [46:09] Totally. I love that. I mean, I've, I've so many thoughts on this and I'm, I'm so on the same track. Like one of the things that we talk about a lot in this course is like using chat GPT as a mirror and a mentor. And the mirror part, I think is sort of what you're getting at, which is. [46:39] to have with a computer. And one of the things I'm getting from what I saw you do is... [46:46] Are you familiar with Lisa Feldman Barrett's work on emotion? Not by name. I don't know. So she's a neuroscientist. She wrote this book called How Emotions Are Made. And one of her like...

[46:57] one of her like research areas is this idea of emotional granularity, which basically means that people who can use more precise words for the emotions that they're experiencing typically have a better ability to cope or deal with the emotions they're experiencing. And so it looks like what you've built for yourself is like kind of consistent with that research. Yeah, almost like exactly that. [47:27] idea where it's like, if I can like name the emotion, you know, and just like have more clarity about it. And even just like, [47:33] you know, understand, like, this is how I'm feeling. And it fits into like, these stories in the world, like, that, [47:41] I think helps you understand it. Have you played with the Hume demo? I have. It is so cool. I was just about to bring that up because I interviewed the CEO of Hume for this show, which is coming out tomorrow. Oh my gosh, I'm definitely listening to that one. Yeah, I feel like you, first of all, you should definitely watch that one. But I do feel like, especially if it's about naming emotions, it's kind of wild. I don't know if you play with their vision API, but it's kind of wild. Like it will, it reads your face [48:11] this is what's on your face. And I feel like that would be so useful for like this kind of journaling exercise. - Yeah, I'm like so curious to run that on myself and like see kind of how knowing how a computer reads me. [48:24] if that like changes how I behave or how I think about myself, it might be weird and like, just make me like overly like, Oh my God, am I acting sad right now? But I'm also just like, what, like, what does that, you know,

[48:35] Can that almost like... [48:37] Yeah, I think... [48:38] for me and help me understand myself. Yeah, Hume is really cool. It's definitely an oh wow moment if you haven't tried that one. [48:44] Totally. One of my favorite things to do on this, like in this line of things, because there's like, it's like an infinite territory, which I love because there's so many different psychological modalities. There's so many different ways to understand yourself and to tell stories about yourself and to think differently and all that kind of stuff. But one of the things that's been super fun for me is... [49:02] is simulating one-way door decisions with AI. What's a one-way door decision? What do you mean when you say that? So a one-way door decision is something like, do I take this job or not? Or do I quit my job to start a company? Or do I get married? Or do I break up? Or do I do X, Y, Z where it's like, [49:22] very hard to go back. [49:24] Um, especially if you're like very conflicted about it. [49:29] And one of the things that's really interesting about AI is, as you know, it's quite a good simulator. So if you tell it to simulate a certain personality type, it will... [49:41] it will do a very good job at that. And so for big decisions that I'm kind of wrestling with, what I'll often do is I'll be journaling about it anyway. And then I'll take the journal entries and I'll throw it into ChatGPT. I don't have a custom GPT for this, so I just do a vanilla ChatGPT. And I'm like, given these journal entries, write, [50:01] the next journal entry in the sequence if I make [50:05] if I make the decision X or, and then do it again for decision Y. And like what ChatGPT does is it sort of like picks up on these like little threads or little patterns in how you're already feeling or how you're already writing about it. And then it sort of is able to project them into the future in this way where when you read it, you're like, I mean, it's not telling the future, but you're, you can kind of like recognize yourself in the, in that future state. It's really cool.

[50:35] you can like, [50:36] see and visualize the plan and like see yourself in it in this strange way. [50:42] That reminds me, I think one of the ways I think that initially really surprised me with using this was like, even when I was using it for programming tasks, [50:51] I found that like, [50:52] having it break down and like plan something out for me in the first place. [50:57] um [50:57] was really fascinating. So one thing I've kind of doing [51:00] is [51:01] maybe kind of similar to what you're doing, but instead of doing journal entries, I'm just kind of like taking my, like my daily to-do list, you know, sometimes like two, two, three things, maybe like three, four, just being like, [51:11] "Here's my things for a day, help me plan it." [51:13] And I was like, OK, like in order to, you know, [51:15] Like get ready for the podcast. Like make sure you have a healthy breakfast, blah, blah, blah. [51:19] And it's so interesting because it's like, oh, like maybe that would be a good idea. Or like, oh, I don't really want to do that. Or, oh, but like... [51:27] Having that plan is like, [51:29] pre-injected into my brain instead of like [51:32] me planning it is a very different experience. [51:36] And it's not something that like, [51:37] You know, I don't like-- [51:39] go and like consult people on like how to plan my more like my day like every single day you know like obviously if I'm working on like a large project on a team or something I have like a [51:49] You know, I'm not like... [51:50] planning out like, you know, three hours of my morning with like a friend. So it's, [51:56] Yeah, it's an odd experience to [51:58] have that kind of put in front of you and like, [52:01] I'm also like, am I just like terrible planning and now I'm using any AI for it? Like it's kind of weird.

[52:07] But it, [52:08] it removes some sort of friction and actually makes it much easier to do things, which is [52:13] Shocking to me. I love that. I mean, I have this like gigantic to do list that it only grows basically and there's just stuff on there that's been there for forever. Like I just want a GPT that like. [52:26] auto cleans it you know and just goes through and it's like none of this yeah like you haven't you have this has been on here for four weeks like you're just not going to do it you know um but i also think like uh that's a really good idea like sort of the planning thing what it reminds me of is that i think chat gbt is really good at reminding you of things that are obvious but that you might not think about so like you should have a good breakfast is like obvious um but like [52:56] is sometimes really, really helpful. Yeah. It can kind of like intercept, like, I don't know, like... [53:04] I don't know, maybe it sounds like parenting yourself or, [53:07] nudging yourself in a small direction. I think that's something that excites me about this technology a lot too is like, [53:12] I'm hoping that some of these become more proactive [53:15] where a lot of them are like, I open the app and I kind of have to start the conversation and get it going. Right? Like I open chat GPT and I have to like, [53:23] tell it what I want it to do. [53:25] I'm really curious if we start to see ones that are much more like, I open the app and it's like, hey, good morning. Like, here's all the things on your list, you know. [53:32] We've decided to like, we're not prioritizing this. We're going to do this, this, and this, you know? [53:36] and

[53:37] Maybe you agree with it, maybe you don't, but [53:40] I'm just like, can we like actually put that into like a core loop of like how we're thinking and going about our day? Like we're definitely already so attached to technology, you know, I guess. [53:50] is writing out an actual to-do list on painted paper that much different, but it is. [53:55] like more powerful in a sense that I guess can [53:58] extrapolate and kind of fill in gaps in a way. Yeah. It's sort of like, imagine if you're Jamie Dimon, you're like a really high level CEO, like you wake up in the morning and [54:09] your one of your team of assistants like hands you your like brief for the day. And it's like customized with all the stuff. And then, you know, your personal chef makes you the exact meal that's right for you. You know, like I feel like some of this stuff can [54:23] makes that kind of experience a little more accessible to people who can't hire an entire staff, you know? [54:30] Which I think is really cool. I'm all for it. And I think we'll probably start to see this more... [54:39] as the AI stuff gets built into more of the basic technologies that we use every day, it's still not really in browsers. ARK is doing it to some degree, but it's not really there yet. It's still not really in operating systems. They're doing cool stuff. I love ARK. It's great. Well, I'm an investor, so I have to say that, but it is great. [55:01] But I think that stuff is coming, and it will be much easier for it to be proactive

[55:09] and all that kind of stuff. Yeah, I guess it seems like there's a huge opportunity for, I think, yeah, people to [55:17] Yeah, almost like, yeah, I guess I want to say parent themself. [55:21] um [55:22] And yeah, it's just like, [55:23] kind of like intercept different modes of thinking where it's like, I don't know, if [55:28] in a depressed mood, like, can you use AI to like catch that and like, [55:32] weaken it or subvert it in a way or redirect it towards something more productive. [55:37] I'm really curious if we get these worked into [55:40] some of our loops there. [55:42] Um, [55:43] But yeah, it's crazy stuff. [55:45] Wild times. Well, this was... [55:48] Amazing. I'm really, really glad that we got to chat. Thank you so much for sharing Grimoire, for sharing all the other things we talked about, custom GPTs, your emotion GPT, planning, all that stuff is great. Where should people find you if they want to catch up with your projects and what you're working on? Yeah. So I guess the best place to follow me would be on Twitter, Nick A. Dobos on there. You can also check out my website, Mind Goblin Studios. That's [56:18] publishing all my crazy experiments and seeing where they go. [56:22] I love it. Thanks, Nick. I really appreciate it. [56:24] Yeah, no, thank you. This has been fun. Yeah, great catching up. [56:28] being able to talk about this. It's so much fun to talk about it. Yes. [56:31] Later on.

[56:47] Why? Because this show is the epitome of awesomeness. It's like finding a treasure chest in your backyard, but instead of gold, it's filled with pure, unadulterated knowledge bombs about ChatGPT. Every episode is a roller coaster of emotions, insights, and laughter that will leave you on the edge of your seat. [57:04] craving for more. It's not just a show. It's a journey into the future with Dan Shipper as the captain of the spaceship. [57:12] So, do yourself a favor. Hit like, smash subscribe, and strap in for the ride of your life. [57:18] And now, without any further ado, let me just say, Dan, I'm absolutely hopelessly in love with you.