How to Find Your Next Big Idea Hiding on the Internet - Ep. 10 with Steph Smith

[00:00] So all I said in this case is really simple prompt. Hey ChatGPT, could you explain what dark matter is in the voice of SpongeBob? It starts literally with brackets, says imitating SpongeBob's enthusiastic and playful voice. Dun-dun-dun, dark matter. By the way, how great is it that they label your chats so cleanly? Wait, I have to stop you. Are you using ChatGPT 3.5? I'm just gonna say I'm disappointed. I know, I'm sorry. It says, use an LLM to de-bias your content. [00:30] his bias is evident in his optimism in his role in the future economy. I have a pretty fundamental question for you. If you don't like the ear flaps, how are you envisioning to be warm? Is this just like a thicker? Look, I'm just the idea guy, you know? I feel like there's probably no search volume for this. Yeah, zero search volume. I'm just like taking it on the chin in this interview. Like my writing is biased and my ideas have no search volume. I feel bad [01:00] Thank you. [01:00] Thank you. [01:12] Steph, welcome to the show. Thanks for having me. This is exciting. Yeah, I'm so excited to have you here. For people who don't know, you are a prolific online creator. You're the host of the A16Z podcast, an amazing podcast. You're the author of Doing Content Right, which is about writing, creating and scaling a blog in 2023. And you are the creator of Internet Pipes, which is, I think, like the most detailed

[01:35] toolkit course I've ever seen for doing research on the internet. I binged it [01:40] like all yesterday and it was just like I was just like like it was amazing I loved it that was a goal I know I was like before I shipped it I was like are people gonna get any value from this but the whole point was just hopefully you go through it and you're just like oh my god I didn't know this existed or like you just end up down some rabbit hole yeah yeah if if you're a person who likes internet rabbit holes it's like the the most amazing thing to to to spend time with so thank you thanks for making it so the place I want to start I want to like start with uh sort of like [02:09] In doing research for the episode, what I found to be what I think is like sort of an underlying theme in the work that you do and how you think about things. And it actually aligns with an essay I wrote recently, which I did not have you in mind when I wrote it. But like, I think it really works pretty well. So I want to read you like just a little bit of that essay. Sure. And then I want to ask you kind of about that. So... [02:29] This is the opening. So it says, time isn't as linear as you think. It has ripples and folds like smooth silk. It doubles back on itself. And if you know where to look, you can catch the future shimmering in the present. This is what people don't understand about visionaries. They don't need to predict the future. They learn to snatch it out of the folds of time and wear it around their bodies like a flowing cloak. [02:51] And like I said, when I wrote that, like I wasn't thinking about you, but I actually think this is very core to your work. And the underlying premise of that passage and I think your work is that the future isn't evenly distributed. It's here and it's on the Internet. And all you have to do is like go and find it. It's there. Yeah. And if you have enough curiosity and enough patience and enough stick to it, like you can find it.

[03:21] is sort of like the place to start. So I want to just talk about that. Tell me about that realization or that thread in your work and how you came to it. Yeah, yeah. The coolest part is that that was always true, right? Time is on the spectrum. And to your point, some people look at visionaries and they're like, they saw the future early. There's this... [03:40] a really cool video from 1964, I think, where Arthur C. Clarke basically talked about this idea of remote work, of course, that [03:48] that that term didn't exist back then but he basically was like you know eventually maybe even in 50 years which is actually kind of funny because that was the time frame that it ended up being people can work the same way in london and bali and tahiti and so that's one example of just again like this is not a new concept that people see the future early but what is a new concept is the [04:12] everyone to kind of get access to that data that also didn't exist back then, right? Like, coming back to your question about when did I kind of wrap my head around this idea that the future is actually like present in these little pockets online is my first job in tech, I ended up working a lot within like the SEO sphere. And it's funny, because that sphere sometimes get some hate because people are like, oh, there's all these black hat tactics and you know, people doing like,

[04:42] things down and you think about what [04:43] search engine optimization is, it's billions of people [04:47] who use this website, Google, the biggest website in the world. And every single day, they go to Google and they tell it, what am I looking for? What do I like? What do I not understand? What are my wants, needs, desires, all baked into these queries? And Google's just one example, right? Many other websites kind of bake this information about billions of people as well, whether it's Reddit or Wikipedia or Twitter or the apps that you have on your phone. [05:17] that help you understand that information and [05:21] And that, you could say, democratization is so... [05:25] nice, because if you think about it, even just like a couple decades ago, if someone wanted to get information about the world, you know, it's the people who were like, [05:34] rich enough to run a, you know, a study on a mass of people or the people confident enough to like walk up to someone and ask them questions. But today it's just all in these data sets online. I think that's really cool. Yeah, I think, I mean, it's, it is absolutely amazing. And you're so good at like any site, you have like a bunch of different tools, like for Reddit, you have like all these different like graphing libraries that you found. And I'm sort of curious, like, how has, [06:04] like turbocharges some portions of that. I'm curious like how and if that has like made its way into your like research workflow. Yeah, I mean, it's limited in my workflow, but the concept is certainly there. And you could say it's the extension of everything I just said, where if you think about keyword research, there are tools like Ahrefs that help you understand the keywords and their volume and, you know, their secondary keywords, which tells you what else someone's interested in.

[06:32] But imagine that turbocharged, that literally is AI, right? That is something like ChatGPT because it's not just scraped Google. It's scraped all the websites we just talked about. And not only has it done that, but it's turned this massive data set into [06:46] into, you know, some people joke that like the world's intelligence is now in a CSV, right? It's like kind of funny, but it's like that [06:53] is the natural extension. And then now we're seeing totally new interfaces where someone maybe who couldn't make sense of a data set [07:02] um now can just query and ask questions right and so that's me i definitely couldn't have made sense of a data set before but even if you think about like the precursor some of the websites you mentioned like i have one of them is this cool map of reddit like to me that's also someone who [07:17] like similar to ChatGPT, created an interface that made this massive data set of millions of subreddits into something where it's like, oh, now I understand that if someone's like on the subreddit Digital Nomad, they also care about like maybe Southeast Asia and they care about like lifestyle design and they care about freelancing. And that's, again, another like [07:36] That's someone who's introducing a user interface to make sense of the world that, again, now we have this data set for. And I think AI is like both the interface, but also just... [07:47] the sheer increase in data also matters there. - Right, right. That makes a lot of sense. I want to get into like how you specifically use ChatGPT in one second, but I have one other question that's just sort of like popping into my head, which is like, I, [07:59] I feel like you have this...

[08:02] Whenever you talk about finding things on the internet, it seems to be about... [08:06] like finding sort of business ideas or opportunities or trends, which, [08:09] I love like it's like it's so interesting to like see all these like little things that are that are starting to trend. You're like, oh, maybe I could make a make an app or make a website or whatever. But then I also think you you seem to also have this like fundamental curiosity about like what's going on on the Internet. And I'm curious, like what that brain space is like. What is it about? [08:27] like doing this kind of deep research that gets you? Yeah, I think it's because the most niche things on the internet are no longer niche. Like, and that in itself is exciting because every single one of us has like, [08:38] the very high level interest that we will talk to someone else about. Maybe it's like your favorite sport or like the like mass book that someone else might find interest in. But there's also all these little things that like, because of the scale of the internet, there are now enough people who might care about something. And the second order effect of that is that I think we actually discover things that like would never have been created before. Like at the very beginning of internet pipes, I talk about like just the weirdest things like someone sending garlic bread to space [09:08] like someone creating this like maze for a squirrel and like, [09:11] That latent interest maybe always existed, but you would never be motivated enough to create something like that because... [09:20] you'd be the only one to enjoy it. And so that, again, the scale to reach enough people who might care about something just actually changes the paradigm of what people are willing to do. [09:30] And I think that's so cool. And so, yes, I think there's, like, a deep appreciation, not just for a trend that someone can make money with, but, like, the creators that are emerging that just do cool stuff for the hell of it. And the internet actually enables that. I love that. There's that, like, urban – I don't know if it's an urban legend or if it's actually real, which is, like, the reason why cats are popular on Reddit is because, like, before Reddit, no one had any place to share their cat. Really? Because cats are, like, antisocial creatures. Yeah, yeah. But, like, finally there was, like, a place to, like, share what your cat was doing.

[10:00] because you can take cat pictures. - Yeah, I know, I love this. - That's so funny. - I talk about these not so niche YouTube channels, and you can apply that to not so niche websites, et cetera, [10:10] And one of them is this [10:13] YouTube channel just about Japanese passenger trains. Can we see it? Yeah, yeah. I need to remind myself of the name, but, like, it's just – [10:23] It's just these Japanese passenger trains and they have only posted, I think, 30 videos [10:29] And they have like a million plus subscribers. That's unbelievable. And I'm just pulling it up now. It's called Travel Alone Idea. [10:38] And... [10:39] They started in 2021. So it's not like one of these phenomena where someone was early to – [10:45] to a channel like YouTube and they got that benefit. [10:49] They have all time views, 240 million all time views, 1.5 million subscribers. And yes, again, only 30 videos with the average monthly views apparently being 700,000. And the best part of this is... [11:02] By the way, this is not a scenario either where someone is just like the best editor and like has really figured out how to nail the algorithm. This person has no narration. There's no sound. The edits are minimal. And all it is is this person walking around these Japanese passenger trains. But again, there's this like... [11:22] fascination I have with the fact that there are enough people who like this, that this channel is insanely successful. I have just decided that I don't like chat GPT anymore. I'm really into trains now. If I can get this many views. I mean, one of the reasons I'm bringing this up is, like I said, I think it kind of defeats a lot of the common mantras around content around like,

[11:52] You have to master thumbnails. I mean, these are not good thumbnails either. It's just like a picture and a red arrow. [11:58] But that's why I actually think some of these tools on the Internet are also very interesting because they help you surface. Like, is there actually demand for this thing or interest in this thing? Right. There's another one that you shared that's like, I think it's called like TV Too Far. Yes. TV Too High, TV Too Low. Oh, TV Too High. And like the most Internet thing ever is that. So TV Too High is the largest one. And I think people saw that subreddit. [12:22] who don't know what it is, tell us what it is. Yeah, so TV2High is a subreddit where I included this in my original note in Internet Pipes because I was like, this is just the most Internet thing I've ever seen. And shout out to fellow creator Pat Walls. He's the one I saw him share on Twitter. But this subreddit has 180,000 [12:41] Um, [12:42] subscribers to it. And it's just people posting pictures of TVs that are too high. So, you know, I have them up on my screen. And again, it's just TVs that are apparently too high. Some of them are obvious and some of them are less obviously too high. But then the nature of the internet, someone saw this and then they created r slash TV too low, r slash TV too far. And then, [13:08] best part is that there is a subreddit tv um just right but it's private and i just thought that was the best that's great i love that um and and for people for people watching or listening who are like why are we talking about this like what is the point of this like you have this thing that you say which is like don't overlook silly yeah um which tell us like what that means and why you think it's actually important to be aware of things like tv too high yeah so i mean in this

[13:38] invent a business idea around this, but I think [13:41] there [13:43] There is a large part of all of our personalities, which is not oriented around like, let me be. [13:49] really serious and figure out an opportunity right now. It's like, let me let me laugh about this thing. Let me again, like watch a Japanese passenger train or like a bunch of other YouTube channels like about someone who's like picking locks or like is like a toe doctor or all these weird things. And like, there's a weird part of all of us. And by the way, if [14:07] there is interest from all of these people, which is proven by like the scale of some of these things, um, [14:12] I just think it's a window into what people really want. And if you can understand what people really want, you know, maybe there is, again, like a direct spinoff in terms of how you can leverage it. But I also think it helps you just create better things in the world. And also, it's almost like the second, third order effect, where if you follow these trails, I think you actually... [14:32] understand your fellow humans a little better and then you know maybe not directly but eventually you will create something a good example is like a creator neil agarwal like he's for me if i think of [14:42] creator that I think epitomizes the internet the best. It is him. And he just creates these websites at neal.fun. I'll pull it up, N-E-A-L.fun. [14:52] And each one of these projects, I bet his first one, [14:56] no one cared about. And a second one maybe got some traction, but he still was making no money. But he's got like 20 or so projects here. And now this is his job. Like now he actually makes money from this. And again, I think it's just this like...

[15:09] Don't. [15:10] So another way to put it is like if you ignore silly, you are so focused on what is immediately actionable and what you can immediately take advantage of. And I think that's just a very short term, like myopic view of the world. I totally agree. I feel like so much of the best stuff is just like. [15:26] wandering through random stuff that like for whatever reason appeals to you and you can't really understand why and then like [15:31] years later, it comes together in this really amazing product or book or whatever. And so having too much of an emphasis on what's practical today, you miss out on making amazing stuff later. And also, it's just more fun. I was going to say, a lot of people, if you think about it, especially our creator space, I think can be really myopic and being like, what's the newsletter that I think has the most demand or something like that? Or how can I create something that goes viral immediately? And then they just have no staying power because... [16:00] they're not having any enjoyment and i think there's like you know neil agarwal is one example but like the the creators i also really respect are the people who you can tell actually enjoy what they're doing and have a deep fascination with it it's not just about like having fun they're just like i actually think this is interesting and that reps off on other people totally totally so i think this is a this is a good segue into chat gbt yeah let's do it so what i want to do

[16:30] how you how you use it we're going to get into like us doing some sort of mutual explorations and all that stuff we have a lot of good stuff planned but yeah tell us at a high level like what do you feel like what do you use it for how does it fit into your life and maybe we can go through some some chats yeah um so i broke this down into what what is this like seven different ways that um i currently use it and what was interesting about this exercise is that like i hadn't really thought about that before currently chat gpt is not my go-to for anything consistently [17:00] as in I don't like wake up and know like, okay, for this kind of problem, I automatically go to chat GPT. But now that I've broken this down, I'm like, okay, I actually have a framework to think about when I reach for it more often. And I want to reach for it more often. But these are the different areas. So for me, one of them is idea generation. And that doesn't mean like just broadly idea generation. But for example, I threw a meetup recently for people at internet pipes, [17:30] CatGPT is really good at kind of helping me extrapolate from a base. Another thing is just helping me understand complex things in simple ways. And that's where you can bring in characters. And, you know, like I once asked, like, help me understand dark matter in the voice of SpongeBob SquarePants. And it like did it really well. And it was enjoyable, right? [17:53] Also practical things. So every time I'm doing anything related to code, I actually find it much better than like Stack Overflow because it actually walks me through the problem or cooking on drinking sweaters, things like that. Another area is debates. And it's not so much that I'm looking to debate ChatGPT. I actually want ChatGPT to give me both sides of a debate for something that I'm, you know, every so often you feel like you have an inclination of like, hmm, I'm not going to debate.

[18:19] I have a strong opinion here, but I also have a feeling I don't know enough about this subject. So I feel like it's really good for that to just be like, here's the other side. Like, help me see this. Cleaning up data. I've only used it for this a few times, but it's actually really nice because... [18:37] You can kind of, unlike Google, actually structure or tell ChatGPT to structure and answer a certain way. I feel like that's one of the most underrated parts of... [18:47] really most of these AI tools is to say, this is the kind of answer I want. And then finally, the last two were, this is mostly for content, but just like the base for titles and intro, just to get something on paper. And then sometimes just for fun, like I've had it right. The other day, I had this deep appreciation for APIs on the internet, like, and just how much of our web is run by API. So I was like, hey, can you like, [19:12] create like a sonnet for apis and it did and it was pretty good right because i mean that's the kind of thing where it's like i would never have spent my own time doing that right um and then the one thing that i have not gotten a chat gbt to crack is assy art um just because i mean it makes sense why it can't quite do that yet but um yeah sometimes just for fun just to be like can can chat gbt do this i love that i love that there's a lot of good stuff in here one of the things that i'm picking up like in the idea generation or understanding complex things or debates it's like [19:42] There's a mind-expanding aspect of the way that you're using it as a theme. I'd love to dive into a couple of those. Maybe we can start with some of the ways that you've used it for idea generation. Do you want to show us a few chats? Yeah. Why don't I just show you? I'll start with the... They called it... By the way, how great is it that they label your chats so cleanly? Wait, I have to stop you. Are you using ChatGPT 3.5?

[20:07] In this case, I will say 90% of these were actually through four. But long story short, I had an enterprise account and then had to switch that for work. And then... I'm just going to say I'm disappointed. I know. I'm sorry. But again, half, at least I would say for sure over 50% were done through four. So you are a four user, but this is not representative of all of your chat GVT usage. This is actually a great example of just like... [20:34] Humans are slow for silly reasons like [20:39] I just think [20:40] If you think about [20:41] probably the reason that people don't use ChachiBD enough, it's just pure friction and habits. And so this is an example of where I've just been too lazy to upgrade this account. And so I have my work account that I'll use for some things. And then I'm just like, [20:56] Yeah, got it. It's okay. You can shave me. I just have to I had professionally obligated to point it out You know what? It's helpful because now that you pointed it out. I won't get as railed in the comments Okay, so tell us about this chat like where did you start? Tell us like what mind space you're in and and how you decided to do it and then what the prompt was yeah, so I [21:20] I was throwing a meetup for people who had bought internet pipes in San Francisco, and I feel like a lot of meetups are [21:26] really bad at [21:28] One being any... [21:30] way shape or form special right like it's just a bunch of people in a room and you it's not memorable in any way um

[21:37] And then two, I feel like people don't. [21:39] like learn anything or meet the right people or really it's just kind of like comes and goes and again maybe that relates to the memorability of it but i said hi chatty by the way do you always greet [21:51] i'm very nice to chat gbt definitely yeah me too me too but not because of the whole like eventual overlord thing i think it's just again a habit oh i'm just neurotic and i don't want it when it takes over i wanted to okay so you are the like looking ahead singularity like i want to be a good standing okay so i said hi chat gbt i'm hosting a meetup for fellow people who love the internet i like to run a few icebreakers what are some some good ideas for splitting the group into [22:21] and internet related, for example, [22:24] What? [22:25] Oh, this is meant to say buy their favorite social media app and screen time usage, which, by the way, it's so nice that ChatGPT can just... [22:31] ignore most of your yeah like typos and stuff yeah um and by the way i didn't even realize i asked this first but [22:38] Another thing about meetups that I feel like often isn't good is just [22:42] You end up in, like, these random groups, so there's no sort of, like, commonality. And so I started with, yeah, how do we break them up? I want to stab you right there. Just on the prompt, I think there are a couple interesting things, like... [22:53] you're giving it enough context. You're telling it a meetup, but like... [22:57] People who love the internet. [22:59] I think it's like a really interesting thing to give to it that someone might not think to do, but it really like changes the output. And then I think you also gave it a couple of like small little pointers, like splitting the group into smaller groups. Like you have a little bit of a vision for what you, for what you wanted to do. And I think all, all that kind of stuff, um, or even examples like, you know, their favorite social media app or screen time usage, like all that kind of stuff is going to get you better results. And it's, it's interesting to see. Yeah. And by the way, I almost, I was like, should I split them by screen time usage? But then I was

[23:29] shaming some people. But yeah, I think you're right that you obviously get better results when you give some guidance. And [23:37] One of the reasons I wanted to share this one is because [23:40] these are some really good ideas. At least I think so. I actually went with the first one or some version of it where it's like, basically I had, [23:48] printed out a bunch of very popular memes that everyone recognizes. And I put them on different tables and was like, which like, what's your favorite meme? And even though that's like a very, very, [24:00] thin slice of like [24:01] people's brains. I just thought it was an interesting way to match people. But I got this from ChatGPT. So they said meme matchup, emoji charades, social media speed dating, tech time capsule, internet trivia challenge, profile picture puzzles, and you know, it keeps going. But like, the point is, it actually came up with like some pretty... [24:24] good ideas for how to actually match people based on internet phenomena. Just like one shot, one prompt, and then you got an idea that you ended up using. Correct. I kept going. [24:35] I always say, can you come up with more? Do you do that? I do that a lot. What I also do, and this is, I actually picked this up from an interview that I did with Linus Lee, who's a researcher at Notion, is I just do, you know, the, like, the redo button? It's like that little, like, circular arrow. I just click that, like, four times, and then it'll just, like, keep going. So you always do the... [24:52] Like refresh, not... [24:54] ask for more. I mean, sometimes if I've done the refresh a few times and it's not giving me like new stuff, I'll do more. But, uh, but I start with a refresh. Yeah. Yeah. In this case, I just said, can you come up with more? But normally I will have, I mean, in this case I was really happy with the result, but sometimes you'll be like, oh, can you make them shorter? Can you like adjust some part of it? But in this case it just came up with more. Um, so again, pretty good. And then I

[25:18] So actually, in this case, if you remember, my original prompt was actually like, can you figure out how to split these people and not like what to do with them? Oh, that's interesting. So it sort of like got it slightly wrong, but it was still a good response. It was like it knew what I actually wanted, but I still wanted to split them up. So I don't know. In this way... [25:38] It got a little confused, but then, yeah, I asked it to... [25:42] follow back up on that and then... [25:44] I think... [25:46] In this case, I said that are more tech related. And then, yeah, I mean, this one's not so crazy. At the end, I did decide to do some trivia as well. And I asked it to come up with some good questions. In this case, actually, I don't think I used any of the questions, but it helped kind of surface questions. [26:01] You know when you see what you don't want? Yeah. And so in this case, I felt like these were like really obvious questions about like, oh, you know. [26:10] when was the hashtag invented or like, what does this like internet acronym mean? And what I wanted is, [26:17] for people to come out of the trivia [26:19] with new... [26:21] not just new concepts, but a new appreciation for the internet to be like, how cool is this? And so I then ended up coming up with my own questions, but like with that realization. I think that's really cool. I think it's, it's such a common experience with chat GBT is like, sometimes it gives you the exact thing that you want. Like it gave you a little idea where you're like, ah, this is great. And then a lot of times it gives you something that's like not actually right, but in not being right, it helps you refine what you actually want, which is itself a valuable

[26:51] the benefit of this like always on, always accessible, like sparring partner is, is, is that kind of thing. Yeah. And the nice thing about it is like, [26:58] you get that same dynamic with humans, but with chat GPT, you can be so explicit and sometimes crude about like, this is not at all what I wanted. And you can't really do that. And chat GPT doesn't get annoyed with you and it's never asleep. I know, I know. There's so many reasons why it's like, yeah, like a 10X experience. I love that. Um, that's really cool. Anything else on the idea generation stuff or should we go on to the next, uh, next thing? Let me see. What else did I put under here? Um, [27:28] Well, sometimes... [27:29] it helps me not so much in this practical way of like, help me think of an activity. But, [27:35] Sometimes [27:36] I just need it to kind of like... [27:39] fill an idea that I don't it's kind of like spinning in my head and so one one example of this is I had this realization a while ago that like [27:49] Our phones are so dynamic in that you think of all the different appliances, like a flashlight or a measuring stick, or I'm trying to think of other examples, a camera, right? Like all of those things are like physical devices. [28:02] appliances or goods that independently existed. And then the phone was like, think of all of the things in there. And I was like, I could only think of a handful. And so I was like, I just want [28:14] a sparring partner to fill in those gaps like chat to what else is in this crazy that's the perfect question for that and it's like how often do you have that question like maybe not that often but when you do it's like it's like oh this saved me so much time right and it and the best to your point chat tpt questions are the ones where you're like i actually don't think this exists

[28:35] explicitly like this on Google. Yeah. It might, I might be wrong in this case, but like it'd be much harder. Yeah. [28:41] to understand totally your question yeah it's like it exists the information exists it's just like no one has collected it and like you get a real-time collector for you yeah yeah exactly so that's another example that's cool [28:54] And then let's talk about understanding complex things. I think this is something that [28:59] People do a lot. Like, I actually was just texting with a friend of mine who said his favorite thing to do in the car these days is he will put ChatGPT on voice mode and then just, like, talk about, like, quantum mechanics with it while he's driving or whatever. Yeah, yeah, yeah. That's awesome. I should start doing that. That's, like, I'm putting that in my back pocket. It's really cool. So, yeah, I'm curious, like, what you're using it for. It looks like, you know, you're maybe using it for, like, you know, explaining things like dark matter or VAT withholding. Like, show us some things you've been learning with ChatGPT. [29:29] matter what. I mean, that was like the best example for me of I had went to I went to this conference and I met this astrophysicist and he gave this talk about it. [29:39] um, [29:40] Well, astrophysics and the most interesting thing from it. [29:44] for me was this point about dark matter and how we just like don't understand it as a species sorry i'm trying to find it it's like me talking to spongebob on dark matter [29:54] Here it is. See, explaining dark matter. [29:58] With SpongeBob. Okay, so all I said in this case is really simple prompt. Hey, ChatGPT, could you explain what dark matter is in the voice of SpongeBob? Like, why SpongeBob? You know, why not any other cartoon character? Like, how did you pick SpongeBob? Yeah, um...

[30:14] I feel like... [30:15] In this case, I don't know if it was, like, super thoughtful other than thinking about, like, what is a cartoon character that... [30:23] I know is like really simplistic and also fun, right? Like, like, how can we introduce like different like conceptual elements? Yeah. And I think something that's underrated here is like, you know, you can, you know, [30:36] the more boring version is you can just say, like, explain this to me like I'm five. Yeah. But if you think about like even just the [30:42] what we talked about earlier the fun in things being helpful in learning as well like if i get the explain it to me like i'm five version i almost feel like like it's like a little condescending or like just i won't remember it but in this case it's like i mean it starts it starts literally with brackets says imitating sponge balls enthusiastic and playful voice and then it says like ahoy me matey like it's it's just like hilarious right and then it's like glad you [31:12] has dark matter. And it's just like, I mean, this is one of those examples where I was like, I will be doing this again. Like, how perfect. And so I won't read through the whole thing. But it basically does go through not just the concept of dark matter in this case, but it was like, so incredible. It just talks about like, you know, the stars and the galaxies first. And like, like, let's first acknowledge what those are, and what we know about them. And then, you know, it goes on to talk about, okay, well, there's this thing that we actually don't know. It's like our [31:42] dark matter. And it was really good at kind of breaking down that concept. That's awesome. And after reading this, like, would you characterize yourself as a dark matter truther? Or do you think that dark matter is like a real thing? Or like, what's your what's your take on dark matter? I mean, I think

[31:58] I could be wrong, but I'm pretty sure the scientific community agrees that this thing exists. We can measure it. We just don't know exactly... [32:08] it's kind of hard. It breaks our brains, which is why I went to ask Spunbob about it, because it's it's a thing that we can like measure and we know exists. But. [32:17] does not follow the laws of many things that we like of normal matter, which is kind of like how our brains operate, right? Like we're not structured to understand dark matter. It's like, that's why we can't see it, hear it, feel it, etc. So yeah, I definitely I guess I'm a dark matter truther. I think dark matter truthers think that dark matter is not real. Oh, interesting. I got to go down that rabbit hole. [32:42] oh by the way i didn't even remember i did this then i said great job just because i was curious and like asked it to do another iconic character and then they chose i said pick your favorite and they chose gandalf gandalf i love that yeah and gandalf's wise and somewhat grave tone you shall not see dark matter yeah exactly like i mean it's pretty incredible like that's why i [33:12] I think an underrated part is almost a game in a way. Totally. I think I would have picked Dexter from Dexter's Laboratory. Oh, okay. Yeah, if I had to pick one. That was one of my favorite cartoons growing up. Yeah. But I think this is great. I think it seems to... [33:27] like replace some amount of like Wikipedia rabbit holing that that people do because you can like get a Wikipedia article that's written specifically for you. Yeah, I mean, Wikipedia is dense. And I think.

[33:40] The real, again, 10X experience is to be able to usher it into the... [33:46] level of complexity that you're at. Because actually, I mean, think about it. I chose SpongeBob, but had it come back with something that was too complex for me to understand, I would just say bring it down further and you can kind of move with the program, which is just fundamentally not what we had before. I mean, you can imagine different versions of humans like that. But even in the case of something like Dark Matter, where even [34:11] our top scientists don't quite understand all facets of it. [34:16] Those people like there's the curse of knowledge, right, where they don't know how to like or that's not the term, is it? Basically, the phenomena where they can't. [34:24] articulate. [34:25] something complex to other people because they don't it's hard there's just so much of a gap yeah [34:30] Yeah. But in this case, they can, this program can take something very complex and usher it back down and also the other way around. Yeah. That's one of the like little rabbit holes I've been down recently in my own thinking about Chachibuti is that it's, it's very good at like one of the, one of the things that has made me see is that there are many different versions of English. [34:50] Um, and you'd think that I'm not talking about dialects. I'm talking about like the way that academics speak versus the way that lay people speak or like the way that product managers speak versus the way that lay people speak or like all that kind of stuff. And there's a lot of inefficiency just in, in our lives, just because those subgroups of people like can't talk to each other. It's so true. Um, and even think about the way that like, you know, different subreddits, like they all have those different, they have different languages for the way that they talk. Think of like the wall street bats.

[35:20] things but it's like it's a community yeah and by the way like so my husband went to princeton and i make fun of all of him and his friends because sometimes there's just like this list of words in their vocabulary where i'm just like no one knows what sisyphusian or sisyphian i don't even know how to say it but it's like they use terms where i'm like [35:37] it's cool for you but like yeah you're like to your point there is a gap there's a gap and i think chat gpt is like really good at like doing like these like subtle translations between different groups of people that wouldn't ordinarily be able to communicate yeah and that's like really really really valuable yeah and to your point it extends also to other languages yeah right there are people i was listening to something yesterday was it on your podcast because i've been where they someone was talking about how [36:04] uh, [36:05] they basically knew people who were not very fluent in English. Um, yeah. And then, right. And then they like skyrocketed, right. There's like full proficiency. That is so cool by the way. Yeah. It's, it's, it's really cool. It's like, it's sort of magical to watch and it'll only get better. Um, it's, yeah, it's really fun. I know this is the worst. Yeah. I love that. Yeah. Cool. So that's great. I love, I love, I love this use case. Um, going back to, uh, going back to your

[36:35] to see other sides of debates. I love that. Tell us about that. Yeah, so let me pull up this. This was... [36:43] So I went to a dinner in Napa with some people. [36:48] with my husband and a friend and it was like a pretty expensive dinner and we got [36:54] not great service. And then we ended up just being kind of like, all like quibbling about how much we should tip. And I, and then we started talking about, you know, like the stripe or sorry, square, um, checkout things that are at cafes. Just this whole thing where I've been seeing it more. There's just an interesting, um, [37:10] backlash against tipping. And I feel that to some degree, but, [37:14] But coming back to what I said before, I was kind of like... [37:17] I have this frustration with some of the tipping changes. Yeah. [37:22] But I also am like... [37:23] What's the other side to this? Like what what would someone else say about, you know, the people who need the tips, etc. So I said, hey, Chachi, could you provide me with a debate between two people arguing about whether tipping should exist in America? [37:37] And then it did. And it's like, I didn't even, by the way, [37:41] ask for this format, but it literally broke down. There's a moderator, and then there's this guy named Alex talking to Jordan, and it just breaks down in really clear bullet points what... [37:52] each side would say. [37:55] And I just thought it was really helpful to... [37:57] I guess, understand both sides. And obviously you can apply this to way more controversial topics. How did it change? If it did at all, how did it change how you think about tipping?

[38:07] Well, I feel like, did I ask any additional questions here? [38:13] I mean, let me look at this real quick because I'm trying to refresh my memory. [38:17] I do know that in going down the rabbit hole around tipping and doing additional research, I learned a lot more about... [38:23] the history of tipping and where it came from. Did you know it actually came from Europe and then was brought to America? And then – [38:29] people in Europe thought it was too aristocratic, so then they actually got rid of it. It's very interesting, and there's slavery that was [38:35] involved in it and it's [38:38] It's an interesting phenomena because it is not universal. Yeah. Every country tips differently. [38:45] That is interesting. It's also one of those things where once you get into a specific regime, like tipping or not tipping, it's very hard to flip. [38:53] So like, um, [38:54] In New York, there's this guy, Danny Meyer, who runs Union Square Hospitality Group, which owns Shake Shack and a bunch of really nice restaurants. [39:03] And they decided to eliminate tipping and to make their menu prices reflect the price. [39:09] a living wage for their staff. Well, that's my argument as well. I just think it should be like [39:14] I don't think people should make less. Yeah. I just think it should be in the price of the meal scheme. Yeah. And then I also think, by the way, if it is just a tax. Yeah. [39:23] which it feels like, right? This whole idea of like, you, you tip like around 20%, then that's okay. But just like, then it should be marketed or branded that way. In my opinion, um, not this like societal, like your, people say that tipping is to reward people for their actions. Well, I feel like if that's true, then there should be a way wider range. That's my argument is like, bake it into the price as much as you can. But then also I would like to reward really great

[39:53] reward the people who don't do a good job or pay or you know give them five right and [39:58] Yeah. No, it's a really interesting, complicated issue. And the thing that they found in Union Square Hospitality Group is they tried to do it, and they did. But I think that they had to stop because when you're the only one that doesn't have tipping, customers feel like your menu is more expensive. [40:28] they can go make more elsewhere. And so it's like really hard to change once it's in a, once it's either a tip, once you have tipping, it's hard to get rid of it. And once you don't have tipping, it's hard to add it. Yeah. It's just like taxes, by the way. Like that, I mean, that is. [40:40] Good. [40:41] What's interesting about, like, all of these, like... [40:44] these things that touch different jurisdictions is that a lot of people often, what they're really arguing over is like what they want to exist, one side, and the other side, I think sometimes... [40:56] is more practical around the idea that like [40:59] there is the reality of people moving to other places or in the case of taxes. Like there's been several studies around, for example, of like – [41:07] I think France added like a billionaire tax and it's like all the billionaires left. Right. So it's like I'm not necessarily commenting on like whether that's good or bad. But I think this like even if we bring this back to like the Internet is just like recognizing the like broad swath of people who all have like individual desires and needs and etc. And the fact that like in the world,

[41:27] Not everywhere is the same. Right. Totally. Totally. Anything else on the debate stuff that you want to show or any other... [41:36] like chat gpt chats that you think would be interesting to talk about um i think just calling out one thing this was like from my um my employer i wasn't a part of this project um but they did ship this thing called sunlight and i just think it relates to the thing that we're talking about where it basically um is it says use an lm to de-bias your content um and what you do is if you go to sunlight um i'm opening it up on github but you can actually open it up on where's [42:06] sort of, um, [42:08] And it doesn't need to be like a news article that maybe has more bias. But you put in a URL. Can you put in one of my articles? Yeah. I'd be curious. Which one? But just if you go to every.to. [42:20] Um... [42:22] So, [42:24] and then scroll down to see where it says chain of thought. Just click the chain of thought icon. Oh, yeah. [42:31] And then... [42:33] click newest [42:36] And let's see, keep going, keep going. [42:41] The Knowledge Economy is Over. That's the one I cited. [42:46] in our uh in the intro of this episode so i'm curious what the bias is well it's funny so obviously this is such a human thing by the way i also the first thing i looked at was one of my own articles or like the same way people look themselves up on the internet um

[42:59] And... [43:00] It was funny because it was my most shared article was the how to be great. [43:04] article and it like it was nice about it but it was accurate too it like ripped me apart it was like this this is like it's just a lot of like you know tropes or like things that are like maybe like not super well researched like not there's like maybe um [43:18] Some simplicity here is, I guess, what it was saying. And I was like, yep, that's guilty. Well, let's see what it does to me. I'm sweating a little bit here now that I know it's not very nice. Well, no, no, no. It does it in a very nice way. But I was just kind of like it definitely surfaced, I think, some accuracy about, you know, how it could be interpreted. Okay. [43:37] So it's given us some factual claims. The author, Dan Shipper, has started using AI tool ChatGPT for summarizing tasks, freeing up his intelligence for directing or editing the summarizing. The author predicts that handing off summarizing to AI will become widespread in the future, impacting the economy. Okay, so like, yeah, it's getting right. Yeah, okay, so it has factual claims. [43:56] scroll down, there is an analysis, which this part is about, I guess, it says techno-optimistic. And so I think different... [44:05] articles would get different flags true like guilty but see it's like you see these things and you're like yeah okay so it says the article the knowledge economy is over welcome to the [44:19] presents a speculative view of the future of work in the age of AI. The author's bias is evident in his optimistic view of AI and his role in the future economy, which he refers to as the allocation economy. He suggests that AI will take over tasks such as summarizing and humans will transition from being makers to managers, allocating tasks to models.

[44:36] Um, [44:37] Okay, the author further reveals his bias in his assertion that AI is an abstraction layer over lower level thinking. This is an oversimplification does not take into account the complexities and limitations of AI. It also dismisses the value of human intelligence and creativity and problem solving. It's filled with speculative statements. This is what I was referring to. I got something similar where it was like, yeah, basically, like, she did not, you know. Wow. Okay. [45:02] Yeah, I feel bad. Visually. [45:07] I would say, like, for the kind of content that we write... [45:09] Obviously, there might be some bias, and maybe it's – [45:14] There are certain statements that aren't necessarily backed by like a scientific study or something like that. But what I think this is more helpful for are things that are like truly biased, whether it's like along the political spectrum or or just I mean, something that if someone happens to chase certain. [45:31] content of a certain style. This... [45:35] could help them recognize that it's like quite extreme in one direction versus I don't think I haven't read this article in depth, but like, I don't think it's extreme in that way. The question, though, is just whether [45:46] um, [45:47] anyone would ever [45:49] actually change their habits based on this or even think to use something like this if they're that far down yeah a bias that that makes it's it's a good question there's so much incentive to like be more biased you know yeah um i had a friend who worked on this thing that um [46:04] It was based on this study that if you're a conservative and you read something in...

[46:11] like Fox News, that is liberal leaning, you're more likely to believe it. Same thing for if you're liberal and you read it in New York Times or whatever. [46:22] And so he built this Chrome extension that would basically, if you were a conservative, it would feed you liberal stories from Fox News. And if you're a liberal, it would feed you conservative stories from The New York Times to, like, try to reduce polarity. It was pretty cool. And it got some traction, but, like, it's like it's not it's always going to be a niche thing, you know? Yeah. I mean, because I think that's a really clever way to do it. I think the other ways I've heard people do it is, like, let me just, like, show all of the different sides in one place. [46:52] does not work because we do gravitate towards certain... [46:55] certain [46:56] you could say like other humans that we resonate with. And I think that's a clever way to do it. The question is just like, I think our brains are so like hardwired to seek out certain things, even with the nuance of like, even though this is from Fox News, let's say if someone is [47:11] like quite far in one direction. [47:14] They will even sniff that out, I think. Like, I think they would be like, oh, this article is, like, not the kind I actually like. And same in the other direction. Yeah, it's a hard problem to solve. [47:25] Okay, so that was disturbing.

[47:31] You know, what's funny is it's kind of like when people Google themselves and they're like, oh, I'm not here. [47:38] There are other more important assessments. [47:43] But yeah, I'd love to keep going. So are there more historical chats that you want to get into that you want to show us? Or do you want to move into the exploration phase of this discussion? Yeah. [47:55] Let's just see... [47:56] I think we've covered most of the interesting ones, maybe one that's [48:00] potentially helpful for people is just the cleaning up data oh yeah show us that one um this was one of my earliest ones and there was a task where i have this website um or i think it's here tags for untranslatable words i have this website called unoya and it's just this directory of untranslatable words and that's so good you can see this is a pretty long thread because i actually got it to do what does you know i mean for me um it's a greek word that basically means like a [48:25] beautiful mind or beautiful thinking, which is [48:28] That's nice. I thought housed the directory pretty well or the ethos of it. But basically, I had this to-do list task where there was – [48:38] I'm always adding to you, Noya, but it's, like, very far and few between because I'll... [48:43] get like 100 new words and then I tag them with these like very broad like is it about like society or like beauty or nature or whatever and you can see all the tags here [48:54] on the screen. But basically, I had all these new words that I wanted to upload, and I just didn't want to tag them. And so sitting in my to do list, and I just said, Hi, Chat Sheep T, I'm building a database of untranslatable words. I already have the words, but each one needs one to three tags associated with it. I'll supply you each word, and then you can select the correct tags from a list that relate to the definition of the question.

[49:19] Like, this is again, where I love that, you know, you don't have to like, [49:22] I think it's a good thing. [49:22] make it super correct for a human, Chachupiti will understand it. And then obviously Chachupiti said, I'd be happy to help. Can you imagine if it was like... No. No, thanks. And then I said, great. So I'm not sure, because this was quite a while ago, why I chose this approach. But what I did is I said, great, let's try one first. The word is Kokoro. And then I said the definition is X. Because it obviously needs the definition... [49:51] to abstract what the tags are. The list of tags is below. Reminder, each word can have one to three tags that fit. And then I sent the tags. And then it said, based on the definition you provided, tags that seem best fit are these ones. [50:05] And in this case, it did a pretty good job. So I said, great job. Let's try another word. And I sent another word and the definition and it did it again. And then I said, great. You seem to understand the assignment. Now, can you please return the responses with just the tags in a formatted table with the columns? And I just said, word, tag one, tag two, tag three. And then I just sent a bunch of the words and the definitions and look what it gave me. That's amazing. It's truly amazing. [50:35] chat gpt so this is many many months ago and i think it tended to hallucinate a little more back then so you might notice actually it added new tags so for example like sensory was not a tag in here slang was not a tag um but pretty good um it got all the work like it

[50:53] I knew the format, it included all the words, and it included three tags for each. And then I guess I was happy enough with that, because I didn't want to be like, can you do this again, but not these tags, I just knew it was going to hallucinate a little bit. So I said, Okay, here's a much longer list of words, as you can see there. [51:09] And then it [51:10] It basically didn't do it. It broke. So then I was like, I basically learned that maybe to your point about like, if you go far enough down the line, like. [51:21] it messes up for whatever reason. So I was like, let me just give it the precise... [51:26] original prompt that seemed to work with the longer set. So I just said, hey, like, this is precisely what I want this table with these columns. And then here are all the words. And I did it. [51:36] You can't [51:37] see there, but again, it did hallucinate. [51:39] a little bit more. So it took a little bit of back and forth, but the [51:44] point is that like you can see one time it hallucinated and didn't do the table and then I said hey can you put it in the table and so but the point is eventually I [51:54] This back and forth seems longer than it actually took. It took me probably like 15 minutes to get a table of like 100 of these words, if not more. And then all I had to do was just scan. [52:05] And we're like, no, that's wrong. Like, let me add this one. And so it took, it definitely reduced it by like a factor or like an order of magnitude. This is great. I mean, this is like exactly the point I'm trying to make in the piece we just looked at that's incredibly biased and techno-optimistic, which is about sort of people moving from doing a lot of individual contributor work to managing models and that they run into the same, when they do that, you get more leverage, but you also run into the same issues

[52:35] managers of today. [52:37] get when they manage people which are for example sometimes when you give a task to someone who you're managing they mess it up like a little bit right and like it's it's it's a huge question for any manager to be like how far into the details do i go and a lot of a lot of managers are like well i don't give people anything because they'll mess it up so i do it all myself but then you realize like you're just going to be doing all the work yourself and like you don't get any leverage that way so then you have to you have to learn how to like which questions to ask how to [53:07] yeah and i think that is a skill itself for chat gpt where a lot of people will be like oh it hallucinated like i could just do this myself and it's like yeah you could just do it yourself um but you'll get a lot more leverage you know you'll end up going faster if you learn the kind of mistakes it's likely to make and learn how to manage it rather than just being like well it's useless and i'm not i'm not going to use it in the same way that managers of today have to learn that same skill exactly and i mean every time you get something incorrect from chat gpt it's like a [53:37] communicating that. And by the way, that's like, I have been that manager that you're describing, who I'm just like, it's just easier to do myself. But at the same time, I think this is also really helpful for people. Like I had a friend do something like this with chat to be tea, where he's like, I, I [53:51] I use ChachiBetit to remove all the duplicates in something. And I was like... [53:55] you can do that in Google Sheets. Like, it's like a button. But there's something really interesting about the fact that, [54:02] This is like a wrapper of all the things that you might be able to do in something like a Google Sheets or Excel or Microsoft Word. Yeah.

[54:10] um, [54:11] But it doesn't require you to know what exists. It just requires you to know what you want, right? And so instead of like, people know that they want to remove duplicates from a list or add tags to something. Yeah. [54:23] Some portion of the population does not know that you can do that very easily and like in [54:28] insert application here. This is, again, like a wrapper that abstracts that. And it's like, you actually don't need to know what functions exist or that we've custom coded into here. And just tell me what you want and we'll work through it. That's so true. It's like the... [54:43] I think one of the downsides of ChatGPT is it's so general that people don't know where to start, which I think is why the show is helpful for people. But the other side of it is like, if you know what you want, you don't have to think about, can it do this? Like, it'll try, you know? Yeah. And I think another takeaway, right, is it doesn't it doesn't have to be binary, like right or wrong. In this case, instead of me doing it myself for six hours... [55:07] I mean, I don't know if it takes six hours. Point is like for many hours would have taken for me to do this right. Or I could get. [55:14] 90% plus there in 15 minutes. Yeah. [55:19] That's much better. And so it's not like right or wrong. Right. So what is your what's your top untranslatable word of the moment? [55:27] Oh, and why? Shoot. I don't know if I have a top one of the moment, but what I will say is my favorite ones are the ones that like... [55:34] reflect something about where it's from in a very concrete way. So if you think about

[55:41] I'm going to get more meta here, but they're code, right? They're code for an experience. [55:48] If you think about why we create... [55:51] code in the form of language. It's because all of us like have all these neurons and things happening in our brain. We're experiencing the world and we want to be able to communicate that to someone else. And if something happens enough, if an experience happens at a threshold where we're like, [56:06] we no longer want to explain this in a paragraph. [56:09] we turn it into a word, right? And so that's why I find it so interesting when words exist only in one place, like these untranslatable words, because I think... [56:19] Often it says something about... [56:21] an experience happening at a threshold high enough only in that place. Yeah. Right. And so, for example, there's some... [56:28] that are quite happy. I'll share a sad one first because I think it like... [56:32] kind of makes this concrete. There's a word in Japan for people who basically die from working too hard. Like literally they like fall on their desk because they've worked too hard. And that [56:45] Word doesn't exist in English because the work culture is different. Now, simultaneously, there's words in parts of Scandinavia that mean work happiness, which also don't exist in English. And, like, again, there's... [56:57] This threshold of experience where someone where something happens enough in one place that they decide, let's no longer say this in a sentence or paragraph. Let's actually encode it into something that we can communicate much more quickly. It's interesting. And so I like when that happens. There's one in Sweden that basically is like waking up to hear the birds sing. And I'm like, why don't we have that? Scandinavians or Northern Europeans or whatever. Amazing. I love it.

[57:27] I mean, this is not comprehensive. There's probably like 700 words on Inouye. Yeah. But it's also interesting at the very least to reflect on which countries have the highest concentration of these untranslatables. And Japan, at least based on my data set, is far and wide the most. And if you've ever been to Japan, you kind of feel that. You feel like the culture is like if you were to map it on some sort of web map. [57:51] Japan feels like it would be more distant... [57:54] in some really excellent [57:56] and also sad ways, like I just said, right? Like they have words for like walking in the forest, like basically like a forest bath, right? And it's like refreshing your brain. So that's like a positive version that doesn't exist in our culture. But, um, [58:09] So yeah, coming back to your question, I don't know if I have a favorite, but I love any that I'm like, wow, that does feel very Japanese, Russian, German, you know, whatever it might be. Yeah, I'm surprised that you think it's Japanese because there's a whole meme that it's like there must be a German word for this, you know? Yeah, but that's because of so that's an interesting thing. There are a bunch of German words in here, but a lot of the German words. [58:27] The way that German works is they'll [58:30] basically... [58:31] piece together words like compound word yeah it's basically a compound word but then they make it they make it a word so like there's one word that a lot of people um where is in the directory but it's like um fear that you know time is running out yeah and i think it's just like time panic [58:47] Wow. Right. So it's like things like that. What do you think it says about German culture that like they have a word for Schadenfreude? [58:54] I don't know. I'm German, by the way. No, are you? Okay, do you speak German? No, but I actually just recently got my German citizenship.

[59:02] Okay. But I don't speak German. Nice EU passport. Yeah. I actually have a British passport, but I got it. [59:09] like within a year of brexit so i was like yeah darn it's okay i didn't vote in it so i didn't have a part of that just sitting pretty with my eu yeah i know i mean it's pretty good access um yeah i don't know i there are words where i'm like i don't i don't know what the key takeaway of this like culture because it doesn't feel as representative of like the country yeah but then there are words where i don't know if i'm just extrapolating more where i'm like [59:33] Hmm. But... [59:35] Schadenfreude, or I don't know how you pronounce it, does feel like one of those words that's like... [59:39] I appreciate that it exists because I think it's more of a reflection of humanity. Definitely. Yeah. Right. When people hear it, they're like, oh, I felt that. Yeah, for sure. Like yesterday. [59:51] Cool. This is great. I love it. So I'm going to move us into the next portion of this discussion, which is the exploration phase where we're going to we're going to what are we going to do together? I don't know. We're going to go down the Internet rapid hole. [1:00:09] to the next portion of our discussion where we're going to go down the internet rabbit hole. We're going to use some of the internet research tools that you use to vet and understand ideas. And maybe we'll use some AI as well. And we're going to see what we do. And specifically, the thing that we sort of came up with is... [1:00:27] One of the things I think you're really good at is if I have a business idea, how do I do the research to understand the market, understand if it's a good idea, help refine it, help differentiate it? So I think it'd be really fun to – I always have lots and lots of random ideas. So what I want to do is I'm going to pitch you some ideas. You say which ideas you think are interesting. We'll pick one of them, and then we'll go down your rabbit hole to understand it more. And I think there are a couple of ways that we can incorporate ChatGPT into this process, and it's going to be really fun.

[1:00:57] that we can start. [1:00:59] Okay, cool. So I'm gonna I'm gonna pitch it's gonna be like a little mini shark tank be as brutal as you want. I've always wanted to be a judge on Shark Tank. [1:01:09] And, and yeah, just like, let's find something that you're excited about. Okay. [1:01:13] So my first idea, it's, [1:01:17] women's walking shoes and i'll tell you about this so my mom has been like walking around her neighborhood for like i don't know 20 years since i was a kid um and she just goes like on one walk a day and it's like a it's like a it's a leisurely stroll but to like keep active basically okay and she used to like i think at the nike store they used to have a section of shoes that was specifically for walking um but they got rid of it and now it's just like running and like very active things [1:01:47] she feels like those shoes aren't really for her and look wise or look wise and just like functionality wise like walking is a different thing from running you know like you have different parameters for what you want when you're when you're on a walk yeah and so i feel like it would be like i've just had in the back of my head that i'm sure if she's feeling that way that there's probably a lot of other people who feel that way where you're like i'm a frequent walker as my form of exercise and i don't really feel like traditional running shoes are like my thing um and so i've

[1:02:17] like [1:02:18] walking shoe brand fur moms, basically. The fur mom's [1:02:22] that yeah would probably work yeah so that's that's number one um [1:02:27] Number two, brand name generic Zoloft or any other kind of like antidepressant or any other kind of medication manufacturer. So like basically if you're taking – there's like tons and tons and tons of prescriptions for Zoloft or any other kind of SRI or SNRI or anything for mental health. And most of them are generic now because they haven't been – Patents expired. Yeah, patents expired. But like if you get a generic prescription, it's sort of like a bummer. [1:02:57] which is great but like i feel like you know like with for example if you're taking psilocybin and you're doing psilocybin therapy like there's this whole like ritual to it where you're like it's like the set and setting and like that's gonna affect how it like how it affects you yeah and i i feel like um antidepressants like have the potential for that um and can be like really transformative for people but they're like purely done in this like kind of clinical way that they make you feel bad that yeah they make you feel bad that you're taking them yeah and so i [1:03:27] to create a brand around generic... [1:03:30] drugs. [1:03:31] So that's another one. I don't know anything about like... [1:03:35] the legalities or whatever. Yeah, there's definitely some considerations. I don't know specifically about antidepressants in this case, but for example, certain drugs you can't brand in certain ways, even just in terms of packaging. I know, for example, in Canada, they have different regulations than America, even though...

[1:03:53] weed is legal in canada all of the different weed that you would get from the government or the government stores just look like these white boxes with like very clinical you know you can imagine like what a drug box would look like yeah which is interesting because all these brands want to you know once it became legalized they wanted to be able to market to different people and attract certain audiences but you can't do that with like a white box with like very limited color differentiation or anything so that's a consideration but there's that's definitely [1:04:23] Okay, so that's one. Another one is a product that we're actually building right now inside of Every that I launched a little demo of it on Twitter that went viral, which I've linked. I'll link it in the show notes so people can look at it. But basically, the idea, it's called Sparkle. And the idea is my desktop and downloads and documents folders are always a freaking mess. And I don't ever want to organize them. [1:04:53] And so it's a little app that just is always running on your computer. It's hooked up to GPT-4. And it basically, when you first run it, it looks at all the files that are on your desktop or on your downloads. It's like... [1:05:05] these are the categories and then it puts them all into folders and then it keeps them organized in that folder structure indefinitely, basically. So everything is always organized, but you don't have to touch it. And I'm excited about it, but like, I don't really know anything about the market, to be honest with you. That's one of those where, I mean, we can, we can go down that rapid hole too, where it's kind of like, yeah, we'd have to figure out the willingness to pay for sure. Yeah. That's, I have a little bit of, a little bit of evidence there where like,

[1:05:35] got a lot of like signups for it. - Okay, nice. - So I have a little bit of validation, but like, yeah, I don't actually know. - By the way, I think that's actually like [1:05:42] That sphere of products I'm very excited about with AI because the whole concept of a second brain I feel like was... [1:05:49] mis-marketed in a way, like not calling out anyone in particular, but the whole idea of [1:05:54] being fully augmented. [1:05:57] was just like the technology wasn't there, but I actually feel like AI changes that paradigm. Totally. Yeah. A hundred percent. It's like all of the things, like I've been a note-taking nerd for so long and like all the things that I've dreamed of are like finally happening. You're like, I don't have to feel overwhelmed by my, you know, notion, Evernote, whatever, because I actually can have something helping me versus... [1:06:16] I feel like it's... [1:06:17] so far for most people, including myself, actually more of a brain drain than a second brain. Totally. Yeah. That's a good one. That's a good line. I like it. I actually just came up with that on the spot, but it definitely resonates. Internet pipes, by the way, would have been created years ago. [1:06:33] If not for just like the fact that I actually didn't have a second brain, I actually didn't use AI very much or really at all in this. But like the point is I was overwhelmed by this like sea of information that was just like sitting. Yeah. So that's why I do think for most people to brain drain. Yeah, that makes sense. Okay, a couple more. So one is I think the next one is it's not a product idea. It's just like I think people pleasing is a big problem for people. I don't know why. Like I definitely never do that.

[1:07:03] for me. And, and so, and I think that there's just like not enough stuff for those kinds of people. So maybe there's like a category, I would just be curious to know, like, what are the communities out there? I know there are a couple, but like, [1:07:18] I'd like to map those communities and see what could possibly be offered to those people. It's kind of an interesting one for me. [1:07:24] My next one is warm baseball caps. I really like wearing baseball caps, but wearing a baseball cap in New York in the winter when it's like literally the freaking surface of the moon temperature out is like the worst, you know, and so then I have to choose between like. [1:07:40] looking the way I want to and being warm. And I would rather just like have a warm baseball cap. [1:07:46] Okay. And I think it's possible to make... [1:07:49] I don't know. Yeah. Okay. I feel like I need to know if anyone actually searches this. That's my first go-to for anything like this, right? Which is coming back to our very earliest conversation, how Google is just this... [1:08:02] It is this data set that shows what people are actually searching for. In this case, at least I use this extension called Keywords Everywhere. It's saying 320 searches per month, which is low. So there's not that many people like me, basically. But this is one of those, like... [1:08:15] I don't know. You come back to the, like, uh, [1:08:18] Fast Horse thing. Yeah. [1:08:19] where I feel like... [1:08:21] Hmm. [1:08:22] It's one of those where I'm like, do people just not know that they want a warm baseball cap? And you might be able to, through secondary keyword research, actually see like, [1:08:29] people are looking for better looking to or better, you know, some version of not this, but insinuating that they actually want.

[1:08:38] a warm version of something stylish. Totally. But I think that's a really interesting move that you just did because, like, I've been thinking about this idea for, like, weeks, and, like, I never... [1:08:46] I never typed in warm baseball cap. [1:08:50] Theoretically, I want this for myself. [1:08:54] And rather than actually researching it, I was like, maybe I could make a whole company to do it. Yeah, I know. Well, that's what's funny. So when I show people these tools, I'm like, this is useful not just to be like, oh, let me go spot a trend, but let me... [1:09:06] figure out if actually no one wants this, that's equally valuable to be like, oh, there's actually not that much search volume for this. [1:09:11] You know, barring what I said before, which is that there could be like secondary versions that insinuate that someone actually wants this. Yeah, I think the thing that's like in the way is you don't want to find out that your idea sucks. [1:09:24] I know, I know. But we all have terrible ideas, right? So I think, well, I mean, I think another version of that is a lot of people, if they see... [1:09:32] research about... [1:09:33] that indicates a lack of demand, they will go down that like fast horse range. [1:09:38] version, they'll almost convince themselves that, oh, well, people don't search for this because they don't know they need it. And obviously that exists. Yeah. But actually the broad majority of things, people... [1:09:48] are searching for if they actually do know what they want. And that's why, like, the beauty of Google or ChatGPT or all these, like, [1:09:55] search engines is that humans actually are telling [1:09:59] that [1:10:00] data set what they want and what they would tell you if you were to pitch the idea [1:10:05] on Shark Tank or something. They'd be like, oh, that's not bad, whatever. But, you know, the search data is more accurate. Right, right. No, that makes a lot of sense. And I think it's like, it's the kind of thing where you have to like literally spend years of your life

[1:10:18] working on a product that no one wants before you this before you start learning how to like do research to like invalidate your idea before you like like pour blood sweat and tears into it and i'm saying this like i've done this a lot and i'm still not doing it but like for something i really took seriously i would i would definitely do the research um but it's just so funny that it's not quick question though um do you see on my screen this one and this one uh-huh because your whole thing was like you want it warm and stylish yeah that for the for the listeners there [1:10:48] caps that have like ear coverings. I don't want the ear coverings. I was going to say, because some like, clearly like someone has created a version of this. I'm going to open this up. Sorry, Amazon. Um, someone just had to pay for that click. When I pull this up, that one little tactic, this is not like, um, [1:11:07] this is not revolutionary, but just going to the ratings, um, [1:11:11] This is where... [1:11:12] you want to pay attention to a few things. Like one of them is for anything that's five stars, are they just like, oh yeah, like good. Like I'm very happy with this. Or are they like... [1:11:22] I've been looking for this, like you're saying. I've been thinking about this forever. I can't believe I didn't find this until now. So that's helpful. But actually, the most helpful things often, especially if you're looking at creating a comparable product or some other version of what you're looking at, is to look at the two- and three-star reviews. Interesting. Because the one-stars are just like the – [1:11:43] my husband likes to say like irrationally angry. They're just like, just they're like super upset about something and they probably would be upset about any product. And similarly, five stars aren't very, um,

[1:11:54] helpful either. But the two and three stars are typically where you'll see someone basically say, you know what, like I was looking for a warm hat. This was warmer. I didn't realize that I actually don't like it. [1:12:04] that the ears do something to my face or like that. That's not really what I want. Um, but you'll see other versions of that where they're like, Oh, well, [1:12:11] I really like this book. [1:12:13] This phone... [1:12:15] carrier case um but i actually need it to be like hipaa compliant or whatever that's just something for like documents that would apply to but you can find little pockets where you're like oh [1:12:25] this is a good product. I could spin it in this way and enough people want that. That's really interesting. What are you, what are you gleaning from this? Like, what's the verdict on this idea? [1:12:33] Well, let's start with what we said. Let's look at the five stars. What are they happy with, if anything? [1:12:38] Winter outdoor hat. [1:12:40] It's funny because also you get kind of some of the stories, like you were saying. I love to wear baseball in the cast. This is me. [1:12:49] And it says, my wife found this one for me, and it is so warm on my bald head. Not me. And the ear covers are awesome. Very dark colors. So, I mean, this does validate this thesis of, like, other people who are specifically looking for something in cold weather, and it's something that they were thinking about. I'm not seeing, like... [1:13:09] super exuberant behavior on this or even like some of this is like gifting or like dad loved it um but and gifting is a negative like you want to find like first person kind of i think [1:13:21] Gifting is not necessarily negative if there's some sort of...

[1:13:25] common gift that people give or sometimes you can find something special like i for example i've been thinking of creating like a hot sauce tasting kit and that feels like a very giftable thing um there might not be search volume for that but it's something that people might talk about right i don't know if this is the kind of thing people would [1:13:41] virally talk about to their friends and family as a gift um so i don't think giving's net bad um [1:13:47] But what you do want is ideally... [1:13:49] The reason gifting could be slanted negative is like you want things that individual people on mass. [1:13:55] realize they need and then search for in some way. I think that's also maybe one of the problems with this idea is like, [1:14:01] it's [1:14:02] supposed to be sort of stealth, that it's warm. [1:14:05] And so it's not obviously something that other people are going to pick up on that you're wearing. Right. It's not a walking advertisement. Yeah. And so that's harder to make that kind of a product work. [1:14:14] So get the three stars. [1:14:17] Great concept, decent execution. [1:14:20] Um, this first one is exactly the solution I was looking for, for a warm hat that looks relatively normal and not an Elmer. [1:14:32] As many, so it seems like some people are unhappy about the size here. Um, yeah. [1:14:37] Ear flaps are pretty thin in this case. [1:14:41] flimsy yeah so this is a case where like i'm not i don't know if there's like a glaring like definitely there's an opportunity here but if you were to pursue this opportunity even just very quickly scanning the three-star reviews it's like oh well the people who wanted this really want it to be warm right and like this is some great middle area where it doesn't really facilitate their needs that makes sense

[1:15:02] Yeah. [1:15:03] Okay, so it seems like we've very quickly invalidated the warm hat idea, which is cool. I actually like that we've crossed it off. Now I don't have to wonder what if. Well, I mean, I should say that was obviously very, very cursory research. [1:15:18] But... [1:15:18] it's also worth doing some of this research to have a sense of, [1:15:22] size. And that's an example where someone actually, and someone has created a small business around this. But it's also helpful to understand as you're looking at these things, it's not just the trend, but it's like how... [1:15:34] you know, [1:15:35] Are you looking to create a multimillion dollar business where it's like you actually just want to solve your own problem and maybe, you know, create like a lifestyle business on the side? Yeah. [1:15:45] Etsy, by the way, is like huge. I bet there's let's look. Yeah, I wonder if there's warm, warm hats on Etsy. Yeah, there's probably like some some guy who just like makes millions of dollars like making warm hats. [1:16:05] That's the biggest thing on Etsy. People will create these puzzles for babies and bridal things. Everything's just personalized, the top stores on here. So warm baseball caps. So we see that same 320. That's from that same extension. Keywords everywhere. [1:16:22] Thank you. [1:16:22] And [1:16:24] So I'm not... Corduroy. Oh, these... Oh, there we go. Everyone has the same idea. 670 reviews. I mean, that's not... [1:16:32] that low um but it's a 43 dollar hat which is a lot i have a pretty fundamental question for you if you don't like the ear flaps yeah

[1:16:39] How were you envisioning to be warm? Is this just like a thicker? Like, I'm just the idea guy, you know? Other people can do the execution here. Because at first when you said it, I was like, is he thinking like heating? Yeah, I'm thinking like it's like lined. It's like lined with... Okay. Because one of the things is wind comes through and that makes it cold. And then, yeah, I feel like if it was lined with something, it would just like retain a little bit more heat. So it's like, it's not supposed to be as warm as a beanie, but it's like... [1:17:08] It's a bit better. Okay. Yeah. So a thick hat. Yeah. We can look that up too. With three C's. That's the domain. Thick hat. [1:17:18] I mean, I feel like... [1:17:20] There's probably no search volume for this, but let's check. [1:17:25] Thank you. [1:17:26] By the way, one of the things that I think is actually really underrated is zero search volume. But obviously, there's probably other permutations, which we can look at here, like the long tail keywords for this. I'm just taking it on the chin in this interview. My writing is biased, and my ideas have no search volume. Well, so something that's helpful, by the way, is I think secondary keywords are really helpful. If you were to think through what they mean, it's like if someone asked this, they might also ask this. [1:17:56] You're like, oh, people are thinking about this. They're just searching it in a slightly different way. And what is that plugin that you're using to find the secondary keywords? It's called Keywords Everywhere. And I'll mention why it's great in a second. Obviously not affiliated. In this case, I'm not seeing any like... [1:18:08] synonym permutations, what I'm actually seeing is like something completely different, which tells me that there's actually just not that much mind space being applied to this. In this case, it's like Google's actually saying like, we don't really see anyone search this

[1:18:20] Are you actually, did you actually mean a baseball cap for thick hair, which I guess is something that, [1:18:24] some small number of people search. But to me, that's actually an indicator where you're like, Google has, again, this is a huge data set. And so anytime Google is like, [1:18:35] Almost like... [1:18:36] trying to take you in a different direction. If you're looking at these secondary keywords, it's probably a signal that it's like, it's like, I haven't seen this before. And to me, that's a pretty strong indicator. But keywords everywhere is great, because it allows you to spot things passively, which I think is one of the most underrated things. [1:18:53] things that you can do on your internet journey. Because think about [1:18:57] all of the different mindsets you take to the internet. Sometimes you're working, sometimes you're checking the weather, sometimes you're looking up some brand or your friend's company, or this is just like endless ways that we use the internet. And typically our minds are unidirectional while we're doing that. It's like, okay, I need to go [1:19:13] grab this thing from Whole Foods, so that's all I'm doing. However... [1:19:18] when you add these extensions and Keywords Everywhere is not the only one, all of a sudden when you're on these extensions, [1:19:24] trips through the internet, you passively notice things like imagine just like, trying to think of a parallel example, but like, imagine if like a typical internet journey is like you biking down the road, but like most people just have like, like a cone around their head, they can only see what's right in front of them. But imagine you like take that cone off, you're like, Oh, there's like a beautiful park over there. And like, Oh, my gosh, there's all these people crowded around this magician, like, what is this magician doing? But these passive examples are, I think, actually underrated, because there's like,

[1:19:53] both the, as we talked about, like the actionable things, I've definitely seen things where I'm like, whoa, like there's 300,000 people who search for 300,000 queries for this every single month. And then [1:20:04] also like, oh my gosh, this trend is exponential. But then there's also the kind of like more fluffy fun part, which is just like, [1:20:11] Also, equally, like there's all these people who care about this thing. Like, that's so cool. I'm not going to go create a business. [1:20:18] Also, if you're searching things like secondary keywords that pop up on the side, it's like, oh, yeah. [1:20:22] I didn't realize that was like a related question. That makes sense. I think what, like another way to summarize what you're saying is like, [1:20:29] If you're... [1:20:30] If you want to be a creative person and if you want to make businesses, if you want to make books, like whatever you want, it's about noticing. Yeah. And like really, really noticing what's going on around you because there's so much that's always going on. And a tool like, you know, Keywords Anywhere is going to like make that noticing sort of passive for you so that you don't necessarily have to be like thinking about it all the time. And maybe that like starts to just like make you think about it more so that you get into that mode more. But yeah, that seems super valuable. I actually really want that. [1:21:00] problem is I use Safari a lot and... Yeah, the Chrome extensions are a lot more like broad. I mean, that's it. You can see my extension bar up here. Like there's quite a few of them. Yeah. For that reason, I just think it's like, [1:21:12] I think of your browser, by the way, as like your access point to the internet. And I think that's actually... [1:21:18] underrated where it's like [1:21:20] If you... [1:21:21] like at least the both of us spend a lot of time online wouldn't you want that vehicle to be

[1:21:27] strong and seamless and like an upgraded version of just like whatever comes out of the box. That makes sense. My other browser is Arc and that has all the Chrome extensions. So I think it's another reason probably to like, I know, I need to switch. You know, there's like tools that [1:21:43] you just have on your to-do list. I've been wanting to do ARK actually for probably around a year now because I can't remember who it was, but someone on Twitter asked this question, like, what is the, I think it was something just along the lines of like, what is the application that you use the most or like that you just think is like so underrated? And the two tools that I saw pop up there were Descript, which I use daily already, and then ARK. And I was like, what? [1:22:06] I got to try it. I love that. Yeah. Cool. It's great. You should definitely check it out. I'm full disclosure. I am a small angel investor, so I'm biased, but yeah, it's great. I mean, can we just quickly talk about the concept I feel like is, is, [1:22:18] amazing just this idea that like [1:22:21] all of our internet journeys, like moving to the cloud, like the idea of a, the same way we were talking about how appliances are like absorbed into this one phone. It's like all of our work online, like actually no longer needs your individualistic experience. [1:22:35] computer eventually like they talk about that like like the internet computer yeah totally they're doing that they have a lot of really cool ai stuff i think if you're like specifically for internet rabbit holes it is like the best organizer of like lots of tabs and all that kind of stuff and one of the features they have that i really love is it like if you leave tabs open overnight it like automatically cleans them the next day and you can go find them i didn't know that it's like so much cleaner okay that's like a thesis i have i think there's there should be

[1:23:03] this whole suite of [1:23:06] products of software products that deal with expiry. Um, yeah. And, [1:23:11] We talk about we're both tech optimists. One of the unfortunate downsides of technology is that there is no cleaning mechanism native. Right. So if you think about like if you live in a house and you hoard, you see all your stuff around you and you like you're like, I can't deal with this anymore. But. [1:23:28] Software allows us to abstract that so things just pile up and then we also like [1:23:32] just accept, oh, I have all these emails in my inbox. I have to get to them eventually. But [1:23:38] I, for example, I don't use it nearly enough now, but I had set up something in Notion that was like a to-do list. And it was an expiring to-do list where I just decided, oh, if I haven't checked this off in 30 days, then it disappears. And I think that's actually like a whole range of products could be designed just around this idea of expiry. I've been on that vibe for a long time. Like that's sort of what this whole Sparkle thing is about, like the automatic cleaning of files and all that kind of stuff. I think it's so important and it's like really magical when it works. Yeah. [1:24:08] I want to just make sure we pick a product and we go down a rabbit hole. We've done a little bit of the baseball cap. We can keep going with that if you're psyched about it. My last pitch to add to you, which I want to say because I think based on having gone through internet pipes you'd be into – [1:24:26] is I was in Thailand recently, um, actually with Hirsch, who's the CTO of, of, of ARC, of the browser company. Um, uh, and, uh, there's this, there's this dish in Chiang Mai, which is a city in the North if you've been, um, called Khao Soi. Yeah. And it is so good. It's so good. It's so good. And you can't get it here. Like you can get it a little bit, but it's like not good. It's not good. Yeah. And, um, and so I've been like looking at a bunch of

[1:24:56] if it's like a D to C cow soy, like ramen type thing, or it's like a cow soy restaurant or whatever, but I just think cow soy is going to happen at some point. And I'm definitely not the right person to start this business, but like, it would be kind of interesting to investigate, you know? Yeah, we could, that one's interesting because I'm like, [1:25:15] I agree with you that... [1:25:18] I mean, this is like a whole other like tangent of just this is going to sound really obvious and dumb to some people. But just like if you've traveled the world enough, I've had this realization when I'm in like Vietnam, for example, is like, oh, the pho here is so much better because the ingredients are local. Like as in like the basil and all of the spices that go into the soup. [1:25:37] are just fresher there. And so there are questions around just as you're like... [1:25:43] spreading food around the world, which is inevitable. [1:25:46] if you can do that well. Is it even possible, yeah, to like to make it good because the ingredients are not going to be as fresh, basically? Yeah, well, I mean, I think it's definitely possible. But it is a question of like... [1:25:57] is it hard? Yeah. How hard? Yeah. How hard? Like we're not going to fly in like lemongrass first class, uh, for, for this experiment. Yeah, exactly. I mean, you could, it was just like, what market are you going for? [1:26:10] Billionaire cow sleigh lovers. Uh, if they exist, I don't know. I mean, we could like start going down that rabbit hole. So, I mean, the first thing, like, [1:26:17] Again, I know this is... Oh, so I said sigh. I mistyped that cow soy. Okay, so the first thing you're doing is you're Googling. I just want to see how... Oh, sorry.

[1:26:29] I'm just curious how many, by the way, 590 people mistype that just like me every month. Not very many. [1:26:37] So I mean, yeah, like just even so just people who are searching like this. [1:26:42] Khao Soi, 110,[redacted address], something that's really important to note as you search these things is Google tells you the intent of the people searching it just by the nature of what is on the page. And so if you think about even like Khao Soi as a query at face value doesn't seem interesting. But it's like, are they searching the definition? Are they searching for pictures of it? Are they searching for recipes? Are they looking for Khao Soi near me? And Google has enough context from all of the 110,000. [1:27:11] thousand searches per month to kind of bake that in. [1:27:15] And in this case, it looks like... [1:27:16] And most people who are searching this are looking for recipes, right? You can see like, this is a recipe page. This is a recipe page. And just like, like in your, in your view, like, is the number of searches like more or less than you expected? And it doesn't seem interesting. [1:27:31] I mean, 110,000 is like [1:27:34] It's not low, right? It's not like the 300 for the more. But it's also, I mean, there's definitely like broader queries that reach too many millions a month. The point of like search volume isn't necessary. [1:27:49] not necessarily to determine like is an idea good or bad from like the raw uh number because we'll go further you know in other dimensions it's more so just to get a sense of like yeah broadly like where does it fit in the like is it like huge is it like medium is it tiny and or is it nothing which is also even more valuable um how is that trending and you'll notice the trending up i'm i'm ahead of the curve here look at you like i think cow soy could be

[1:28:19] think like there's a whole chapter in internet pipes which i think is like very underrated which is [1:28:24] taking things from one part of the world and applying them elsewhere with the right branding or the right like. [1:28:30] uh, [1:28:30] operational strategy, like you said, D to C, right? Making it simple. So I mean, even just at face value, I love just showing people a [1:28:38] SERP, a search engine results page like this, and just being like, what can you learn? And already you can learn like, [1:28:44] Roughly the volume, you can learn that when people search this, they're looking for recipes, not some of the other things I mentioned. We learn that it's growing over time. We can look at the related keywords and you can see it's, you know, not surprising, but like how soy meaning, how straight restaurant, main ingredients, paste, recipes. So there's demand for not just obviously the term, but all these other things. Yeah. [1:29:05] You also really quickly, just because Google's like pretty... [1:29:07] advance, like you can see how many restaurants there are. Oh, sorry, I'm weirdly zooming in here. But there's also, did you know, I discovered this recently, there's food.google.com. No, I did not know that. [1:29:16] It's like a kind of abstracted version of Google. Oh, wow. But you can see... [1:29:20] Like dishes in restaurants? Yeah. [1:29:22] Yeah, so I wonder, actually, if we can super quickly... [1:29:30] And by the way, sometimes it's helpful to do just like comparative searching. So you asked me about like volume. Well... [1:29:36] There's no straight answer as to whether 110,000 is good, but we can also look at something that's a little more popular and common, something like pad thai or green curry to see. So in this case, I don't wonder if it tells you that. Should we just doordash some caps away? Right now.

[1:29:52] I mean, again, some of this is like some people might be thinking, oh, well, not all of this is like. [1:29:57] extremely data-driven but i think what's helpful along the research journey is just picking up on things and one of the things i'm picking up on is like the restaurants that are coming up here like don't really have high reviews and it's like that's like it just a purely anecdotal or like qualitative thing that i'm noticing by high reviews you mean number of reviews no i mean like the actual rating because they're like 4.2 is like in your views so 4.2 is low for you out of 5 [1:30:22] I think so. Okay. [1:30:24] Uh, maybe I'm just picking. I'm just like, I'm legitimately asking, like I would, I would have, I would put 4.2 as like, it's decent, you know, it's not like a 4.9 is like, oh, this is actually, this could be like really good. Yeah. It's decent, but not great. Okay. Yeah. But, but you're, you're saying it's sort of like, eh. I think it depends on the city, right? I know some, I'm not local to New York, so like maybe in New York it's. [1:30:46] people rate lower. There's definitely places in Bali, for example, in Canggu. Everything was so high that we... [1:30:53] some of my friends were like, we can't trust the 4.5. Right. I get 4.5. It's not good. Yeah. Um, [1:30:58] So, yeah, I think I have that with different apps to like DoorDash. I always rate five. [1:31:02] Because I'm kind of rating my Dasher rather than the restaurant. And so I think the ratings are artificially high. But, yeah. [1:31:10] Cool. Um, so coming back here, I mean, one thing we can, my computer is freezing. There we go. Um, [1:31:18] So we can also look at the people also search for thing. And so like, let's just quickly get a sense of like cow soy near me. That's like...

[1:31:26] you know, the classic like location based thing. By the way, are you thinking like, I guess we can figure this out along our journey, but like a restaurant, a kit. One of the things that occurred to me is that a big component of khao soi that makes it difficult to make is you have to make the khao soi like curry paste first. And if you just had the curry paste, it would be a lot easier to like put the rest together. And so maybe it's like a DTC curry [1:31:56] I love to see charts like this, by the way, which are like low volume, but like growing for sure. Because I mean, it's one thing. Let me show you like if we just go to like Google Trends. [1:32:08] Thank you. [1:32:09] I always just like, you know how there's like saved queries. I always open Tiger. Like it's just, it's just there. It's just, it's saved in my bar. Um, [1:32:18] But I just wanted to show you and others like how long some of these trends can go. So like edamame is my favorite because it's something that people are pretty familiar with now. Mm-hmm. [1:32:29] Um, [1:32:30] But you can see it has trickled up since 2003 when they started. [1:32:34] tracking this and it's just like continuously because i think to your point if something is good and like try finding someone who's tried khao soi who [1:32:41] like hates it. There's not many people, but to your point, there's just like operational or logistical constraints. But to me, this is like, if someone can fix that, there's like latent demand for it. Um, [1:32:54] I'm trying to think of like the next... Well, edamame is just like a force. I didn't realize that. Okay. People are going to be like, you're talking about this on another podcast. People love these snackable beans, like lupini beans.

[1:33:07] Have you heard me talk about Lupini beans? No, I haven't. I wonder if this... [1:33:10] I don't even know how this graph is going to look. Oh, there we go. [1:33:14] Lupini beans. [1:33:15] are these beans, I think, originally from Italy. But the reason people love them is because these beans actually taste quite good. They're highly snackable, but they're very healthy. Like the protein and the other fibers in them. [1:33:29] these beans that are native to these parts of the world, like Italy or like [1:33:33] Japan, I think just like are slowly, you know, now you can find edamame in a Costco. Right. Right. Like as like a chip replacement type thing, like the beans are migrating as a chip replacement. Yeah. I think a healthy enough snack. Got it. [1:33:49] Yeah. OK, so. [1:33:51] If we're thinking of CalSway, another few places that I would check, for example, are, I mean, Reddit, of course. And so one... [1:33:59] tool that I always check and it's not always guaranteed to, um, [1:34:05] surface something, but it's just to see what subreddits exist. So tell us what we're looking at here. This is a tool made by this guy. Um, his last name is on Vaca. So that's, you know, why it's like his GitHub and he didn't really name it. Um, but it's a tool where you can see the interconnections on Reddit in any given subreddit. And so we can, I typically start with something broad that I know exists. So for example, r slash Thailand, and then once this does its [1:34:35] the related subreddits. This one doesn't seem as interesting, but you can kind of see like

[1:34:40] What's TFL? That's interesting. Vaz forums? Isn't this the edge, like people who go as teachers to Thailand? But yeah, I mean, one interesting part of this little rabbit hole is purely just... [1:34:54] what other subreddits exist. And by the way, the same person who made that also made this map of Reddit, which is I use both often because they, for whatever reason, don't always function exactly the same. This one tends to give you less subreddits, like less nodes. And [1:35:14] in connectivity, but it [1:35:16] views everything in more of, like, an extended map where you can see... [1:35:20] Yeah. [1:35:20] If I zoom out. [1:35:22] all of the subreddits in this like large. Oh, wow. [1:35:25] You can see it's kind of like a... [1:35:27] a world map, right? Like, so cool. Yeah. So you can see where, like of all things, like our slash Thailand is in this thing, which is kind of in travel land, but also like kind of next to parenting. Interestingly enough, I'm like, yeah, [1:35:41] Yeah. How did they make this? [1:35:43] I think he's just like a machine learning engineer. Because if you look at a lot of his... He's done this in really interesting ways. He's like another one of these like top five creators for me who just does this for fun. But he's got another tool, for example, where he's done the same thing for Google queries where it's like X versus Y. And so you can see for like any given thing... [1:36:01] a product, what are the most obvious close ties for comparison? So, [1:36:09] He has another one, by the way, like another product on his GitHub where...

[1:36:13] Basically, he's like, [1:36:14] taken [1:36:16] maps of places and then just distilled it down into this like really basic street view. And people like print them on... [1:36:23] cut on mugs and posters and he's just really fascinating. But coming back to r slash Thailand, we can also go to Chiang Mai. So you can click any of them and just see like, sometimes it changes and you start to see new connections. But in this case, it doesn't, it seems like that. I can't believe there's not a specific cow soy subreddit. Oh, Thai food, but see there's Thai food. Okay. What can we learn from that? Um, [1:36:47] See, and then now we are actually, see, we're kind of getting into... [1:36:51] a thicker part of the internet, um, where now we see Thai food, but we also see things like Korean food or Japanese food, Vietnamese, Asian eats is a subreddit on its own. Um, you also see, I think I saw something here, cooking videos, recipe inspiration. Um, there's a subreddit called [1:37:10] shitty ramen. [1:37:10] Right. And so, by the way, I think we talked about like not ignoring the silly. I would say when you're going down these Internet rabbit holes, it's equally important to not ignore just like. [1:37:20] What? [1:37:21] you find interesting. As in like, if you're going down specifically to validate this like idea of cow soy, not to, uh, [1:37:28] feel like you can't [1:37:30] explore outside of that, if that makes sense. And then also interesting things like kombucha is up here on the [1:37:36] left hand side sushi abomination, Mexican food gore, by the way, like to me, I'm seeing like something really interesting here, which is like, kind of the flagship, uh,

[1:37:47] foods of some of these places have, um, [1:37:52] you could almost say super fans, who... [1:37:54] really take offense to [1:37:57] like when it's done badly. And so, yeah, [1:38:02] That's interesting to me. [1:38:05] You also see things like recipe gifs, which is... [1:38:09] By the way, do you know this whole rabbit hole? Mm-mm. [1:38:12] I think due to partially the... [1:38:15] the ridiculousness of recipes online people started creating recipe gifs which were just like super and you see this on tiktok as well but that's like a big community um not so much on reddit but i think things like tumblr where people just post recipe gifs instead of the long it's like step one step two but it's like in [1:38:33] Yeah, you just watch it. You just watch the actual thing happening. [1:38:37] Yeah, so I'm seeing a lot of... [1:38:39] here, we can take it in a few different directions. Like we can, for example, one direction you can do is you can go to r slash Thai food and you can use tools like subreddit stats or gummy search to see the top voted posts over time. So you can get a sense of like, [1:38:57] Sometimes it's obvious, like maybe it's all Pad Thai, but I... [1:39:01] Bet you it's not in this case, and you can see, like, what people are actually talking about. If I pull up Gummy Search, for example, you can... What is Gummy Search? Gummy Search? [1:39:09] It is a Reddit tool. Okay. And you can see I've got a few saved audiences when this opens, but... [1:39:17] You can, for example, add an audience. So make a new audience. And I'll just put Thai food.

[1:39:23] Thank you. [1:39:24] And you can add multiple communities here, which can be helpful. Because, for example, if we were actually just trying to [1:39:30] to compile the learnings from what I just noticed about like there's one about like sushi abomination and there's other ones about like other kinds of food gone wrong. Yeah. We could, for example, add food. [1:39:41] all of those. In this case, I'm just going to add r slash Thai food to start. [1:39:46] And this is where we can start to learn. This is a relatively small subreddit. So how many people? 10,000 subs? Like, yeah, it's pretty small. But this is where you can, if you go to, you open up the subreddit here. Again, this is a smaller subreddit. So I don't know if you're going to learn as much, but you can see [1:40:04] keywords that are in here, [1:40:06] posts and like you know specific days or whether it's like [1:40:10] Posts tend to do better with images, etc. You can see the kind of people in here in the subreddit. In this case, again, small subreddit, so we're not seeing too much. But you can also... [1:40:22] The thing that Gummy Search does that's different from something like a subreddit stats is if I go to themes, [1:40:29] And it can be, again, not [1:40:31] always the best for smaller subreddits, but they've basically broken down top content. So you can see automatically, like, what are people talking about? What is the most upvoted? So let's just see browse top posts. [1:40:45] And what they've done is they've applied AI as well, so you can look at patterns. So it'll basically be like, what are the... [1:40:51] What are the themes that we're seeing? [1:40:53] the most on here. So this is just the pure view of just like top voted submissions. I mean, I'm with, I'm with that person. It's so good. I spelled it wrong, but like, it's fine. Let's see here. So for example, 96 submissions,

[1:41:09] Let's just see find patterns what it says. [1:41:16] And [1:41:18] Basically, to me, this is not super interesting, but sometimes something will surface about the patterns. There was one, for example, where it was like Europe Fire, the subreddit, which was significantly bigger. Like financial independence fire. Yeah, exactly. Yeah. [1:41:33] it had all the obvious things like people care about like ETFs and things like that. And then there was one about like Portugal real estate and you're like, Oh, there's a lot of people talking about Portugal real estate. That's interesting. Um, [1:41:43] But for example, like Thai curry, you can click into any of these if they do seem interesting, and then you can pull up. [1:41:49] the specific post that was like highly upvoted. And then what you can also do is like in this section, I guess since there's not enough posts, it won't let me do this, but like you can... [1:41:59] click pain and anger, for example, or specific advice or solution requests. So if there is a big enough community, you can. [1:42:07] one of these other ones, for example, like [1:42:09] like Asian food. I wonder how big this one is. Oh, that's even smaller. [1:42:13] Um... [1:42:14] you can basically actually [1:42:16] surface some very specific solution requests, advice requests, et cetera. Got it. So coming back to the idea we're trying to refine, based on the things you've seen so far, what is it making your brain do, and what do you think [1:42:30] we should do to take the next step in the idea, either validation or refinement or whatever. Yeah, I think so. Maybe the next thing I would open up is something like answer the public. And what I want to get more color around is like,

[1:42:44] What... [1:42:45] people are associating this thing to. So again, some of those secondary keywords to see if there's enough volume for things like, oh, like, I actually just want a better recipe or I'm like, I need something here. Another thing you might want to do is just start looking into competitors or like to see if anyone's actually doing it, doing this. So like, we can look up like, [1:43:05] like cow soy, um, like recipe kit or something like that. And not so much for the volume, but to actually see, um, who's doing this. And, um, you can see a few. So then you would take something like some of these kits and either put them into a tool like similar web to see the, um, [1:43:22] the volume, the number of page views they have a month, or you can also look into, I see some of these on Amazon, for example. And so, [1:43:30] I would be interested in pulling up something like a jungle scout [1:43:33] to see how much they're actually doing. [1:43:36] And on Jungle Scout, you can see, like, [1:43:38] the margins and things like that as well. [1:43:40] So I think one direction that you could take it is actually just trying to see what... [1:43:46] is being sold. So we, we searched cow soy recipe kit. There is no search volume according to that. Um, but you start to see a few different, uh, [1:43:56] Companies that do have, for example, cow soy seasoning mix, a cow soy vegetarian protein hot cup that are selling on Amazon, for example. And then we see on the actual results page a few different companies. [1:44:08] Companies. [1:44:09] that aren't advertising but do sell some sort of kit. And so the next place you could take it is you could take some of these companies that

[1:44:16] cell on the SERP and you can put them into a similar web, for example. So if we take [1:44:22] Actually, this is [1:44:24] This is just one product on TakeoutKit, for example. But actually, yeah, why don't we put TakeoutKit into similar web? I'm just curious. Hmm. [1:44:31] how that does. And then in addition, I pulled up Jungle Scout, which is [1:44:36] an Amazon analytics tool where you can see specific products. Like we saw two of them advertising. So we can actually get a sense of how much volume they do. And of course, I'm getting hit with the captcha. Which I mean, at this point, they should be getting rid of those. Like, yeah, I can do this. [1:44:52] Well, can't AI solve the captcha, right? The robots won. Yeah. [1:44:58] All right. While this is loading, let's look at Jungle Scout. So I searched in their keyword [1:45:06] Thank you. [1:45:06] We'll get to the product. [1:45:07] search as well. Yeah. But where you can get a sense of just like, [1:45:12] how much volume there is. So in this case down here, we have cow soy. There's very limited volume, less than 450 searches per month. [1:45:20] But you can also see, yes, some people are bidding. Easy to rank moderate. This is where you can get additional ideas. Like you can see something, for example, like Asian noodles actually has way more search volume. [1:45:31] but it's actually much easier to rank on. There's a lot of branded terms here, but you can get a sense of what people are actually searching for. And again, like maybe... [1:45:40] Pivot, if you... [1:45:42] If you're not like set on a specific idea. I'm pretty wedded to cows. Okay, so then let's go to product research. We can pivot if you want. Whatever you think is the right move. Let's go to the product database and just see. Let's see how much these companies are making. Interesting. It's so interesting that it's a low competition keyword. Sorry, it's a low volume keyword, but it's still moderate competition. Like that seems like not a great thing to play in, right? Yeah, I mean, all of this is data. So I should say like, you know, some people might be listening and they're just like,

[1:46:11] you know, you don't have a clear answer by the end. And to be fair, some rabbit holes you go down, there is a very clear, like... [1:46:18] No. And then sometimes there's not necessarily like a definitive yes, but you're like, oh, actually, there's. [1:46:23] seems to be like this gap where no one's addressing this like huge keyword search volume and I could create a product in it. But often you're just gaining data, right? You're getting data about the reviews. [1:46:35] the latent demand out there, the products and how they're currently doing in the space. And you're all kind of like bunching that together and seeing if you can create something that's differentiated. So in this case, like there's this like cow soy cooking kit and we can sort all this by the way, with like monthly revenue in jungle scout. [1:46:53] So, [1:46:55] So in this case, it's low, right? Like the top for Khao Soi, at least what's coming up here is like $1,500 versus, oh, but see, I mean, coming back to the review conversation, like really low reviews. Right. And that's like the top one, according to this. Also, something that's interesting, kind of paying attention to what this is surfacing, not just pure numbers. There are T-shirts. [1:47:15] that people are selling that are, like, about cow soy. And, like, I love cow soy. And I don't know. Sometimes you just have to pull in those... [1:47:24] qualitative data points just to be like, okay, it's like, I'm not the only one who really loves [1:47:29] this thing. If we pull up the similar web, if it's going to load, I'm just curious how much... [1:47:34] that website about like, um, [1:47:38] Takeout kits. Takeout kits, yeah.

[1:47:42] And Takeout Kid is bigger than just Khao Soi. Yes, correct. But it kind of gives you a sense of like... [1:47:46] Look at its description here, which is helpful even in itself. It says it's global pantry meal kit. So it's not just like any meal kits. They're specifically orienting around global pantry meal kits, Asian meal kits, Robin meal kits. So it is like... [1:47:59] in the wheelhouse of what you're describing. Um, [1:48:02] So it's pretty low, 10,000 views per month. The nice part about this, a similar web, even their free version, is you can start to see the related sites as well. So this is where you can kind of go down another rabbit hole and you're like, oh, let's try the world. What's house foods? And then in this case, what's also helpful is just to see like, OK, so they basically do no paid search. And sometimes this is helpful to. [1:48:26] um, [1:48:27] You can view it in different ways. Sometimes if you look at a space and you're like, oh, every participant in it is doing organic search, for example. Well, like, what if I actually... [1:48:38] juiced [1:48:39] you know, my product with paid, like, could I outcompete them? Could I like get more attention, build a stronger brand? But there's a second side to this where it's not like binary, where what you actually might be seeing is like, this kind of thing you can only do with organic search, as in, [1:48:53] You can only really be profitable or build something sustainable. [1:48:57] And so... [1:48:58] It may be an indicator, like another data point for when you're assessing these things to be like, [1:49:03] is this the kind of business I want to build? For example, if you go down the rabbit hole of like comparing [1:49:09] different credit cards and points and that's what like what nerd wallet does yeah

[1:49:13] They have so much organic search and like all their competitors are also like organic search professionals. [1:49:17] Kings and like in that particular market if I were to assess it I'd be like oh actually I would only ever do this if I think I have some advantage in organic search right like I'm not going to be able to compete with paid or like social or something else that makes sense. So yeah I mean to me it's everything we've learned so far is like. [1:49:37] One, I probably, if I had more time, I'd be going further down this, like, Reddit rabbit hole and looking for signals in terms of, like, maybe not only on the r slash, like, Thai food, but some of these, like, adjacent subreddits to see if there's just, like... [1:49:49] Latent demand of people who want, like, way better Asian food. Mm-hmm. [1:49:53] in america as an example but then i also would be kind of like bringing every all this together to be like okay clearly there's not like a huge amount of demand do i think i could [1:50:04] create like a differentiated product that probably in this case, because there's not much search demand, I could market really well on social, for example, and like create something like DTC. And like, do I have like good enough Facebook ad chops to do that effectively? And, [1:50:20] You know, again, you're not going to get a definitive answer, but it's probably enough to start working with to be like, can I create some like basic tests where I don't even need to create the product, but I could run some ads and get some data. Yeah, I mean, that makes a lot of sense. It's so valuable. [1:50:33] Because we just went down this list of ideas. We took two of them. And in like five minutes, I'm already like, I'm probably not going to do the warm hat thing.

[1:50:43] If you pursue the warm hat thing after our cursory research, you just want that product. Yeah, I just sort of want it. You want that product. Yeah. [1:50:51] And then I think for the cow soy thing, like the vibe I'm getting is cow soy is going to happen. Cow soy is like it's it's inevitable. Right. And then I have to like I have to do this calculus of like, am I going to be the leader of the cow soy movement? Are you going to be the cow soy king? Obviously not. But like I feel like what I'm seeing, like looking at that chart is. [1:51:12] So probably in five years or 10 years, that chart will have continued that trend and someone will do that. And that's like an interesting thing to pay attention to. And it might be an opportunity later on down the road once it's like a little bit bigger and there's more search volume and there's there's all kind of stuff. [1:51:42] like sometimes that is actually the right move because like you just all you really want to do is like make the best cow soy and in that case like you shouldn't you don't need to do the search engine because you want to be you want to build that movement you want to build a movement of it and sometimes like you want to obviously know is there an actual market here is there a business to be built here and i think what we found is like [1:52:03] There could be, but it's going to take a long time and it probably right now is not the right time. And that's such a valuable thing to know. Yeah. And again, you can go way deeper down the rabbit hole, but I think to your point. [1:52:13] even just these few data points, I think the way you describe the picture is accurate. And people just need to decide...

[1:52:20] from these data points, is this the kind of business I want to build? Like, where are we on the curve? Is it going up or down? How far along are we? Do I want to be the person who's really like pushing the rock up the hill? Or do I want to be someone who comes in at the end and just builds the coolest version of the product once there's demand? By the way, just one more thing that might be worth looking at is if you look up especially something like Khao Soi, which we're unsure of... [1:52:46] in terms of [1:52:47] Like... [1:52:47] If there's enough interest, you can also look at the geographical interest. And so let's do this for the U.S. only. [1:52:59] Something that reminds me, Khao Soi reminds me of Hard Kombucha, which I saw this conversation [1:53:06] Using keywords everywhere, by the way, years ago I was like... [1:53:08] just buying some and I was like, oh, [1:53:11] There's a trend here. Like, I didn't even know that there was such a trend. And if you looked at it geographically back then, it was like only in a few states. And this doesn't surprise me at all. It's on the coast, right? Yeah. But I... [1:53:20] I bet that if you actually... [1:53:24] looked at this a year, three years, certainly five years from now, this would be all blue. It wouldn't all be the same blue, of course. But I mean, even this is interesting in a way, like I wouldn't have necessarily expected Oregon and Washington to be like, maybe there's like a lot of Thai immigrants there or something like that. Maybe it's possible. Or they just have good taste there. It's funny. I was like, you know, do you know what Zen is? [1:53:43] it's like the like um nicotine patches oh yeah i do you know that yeah are you a user no i'm not but there's a there are zen watches which i know um but yeah uh well no the reason i was uh mentioning it is because the exponential curve in this is so big and i was asking people i posted it and i was like why why montana and i think it's because there's a lot of scandinavian

[1:54:05] immigrants there. Okay. That's so funny. Wow. That's, that's fascinating. This is, this was incredible. I feel like I'm like in the presence of genius. Like you're, you're, you're an internet impresario. Well, I mean, I look, what I'm excited about is like, [1:54:19] These are the things that I rely on, and I think there's, like, [1:54:23] a few things going on right now. One of them is that more people are using these tools, they'll get better and they'll improve. The second is that like, there's way more tools that I don't know about. So I want people to surface those and like me to like build out my like internet. [1:54:37] toolkit and then [1:54:40] I just think over time, like this is like we talked about, like, AI is the worst that'll ever be. Yeah. Like our ability to ascertain what people want, need, enjoy. [1:54:49] is only going to improve with time. And some people might think that's like, [1:54:52] privacy concerns, but obviously all of these, this is like aggregate data, right? Like I don't know who actually likes Khao Soi. I just have a sense of that it's desired potentially. And I think ultimately the reason I'm excited is because I hope that as more people use this, we actually get better products, right? Like if people actually understand not just where demand is, but there's other tools we can use to... [1:55:14] more so gauge, like, again, like, not just... [1:55:17] creating a cow's way kit but like if i were to do that how would i do that to make it better totally i love that that that that makes a lot of sense um and it i don't know it makes me really excited for like what these tools are going to look like in five years or something like that um so before we end and i know we're like we're very close to time but we had a bunch of um we had a bunch of twitter questions for you and i would be remiss if i didn't ask like a lightning

[1:55:47] - Okay, cool. So, LaCrypto asks, "Why you stop doing shit you don't learn in school?" It was their favorite podcast. - Oh, that's so nice. It's 'cause of me. So I do it with my husband. [1:55:59] It's funny. That's the thing we get asked about the most. And it just takes more work than I think people realize. You run a podcast, you know, but it's coming back. We're doing it in very disparate seasons. Amazing. I love it. Jessa asked, do you use ChatGPT for therapy or self-help? [1:56:14] No, but I definitely can see the use case. I have a therapist. I'm very for therapy. [1:56:22] couples therapist and that's been amazing that's awesome yeah uh what's something that you worked on andrew asked what's something that you worked on that turned out to be a dud [1:56:31] Oh. [1:56:32] I mean, go to my product hunt. All my first projects are like not... It's interesting because... [1:56:39] They're very much of that ilk of Internet products like I mentioned. [1:56:44] It's like it's a project. [1:56:46] So funny that those two words, by the way, mean different things like product and project. And sometimes they're interchangeable and sometimes they're not. But if you look at a bunch of my early podcasts, [1:56:55] projects like nomad hub was my first one after i learned to code and [1:56:59] It's like... [1:57:01] no one needed it so yeah there's tons of random everyone who makes stuff has that graveyard of stuff they made that was terrible so yeah okay last one what is a habit or or what is a habit or action she's intentionally stopped that has helped her productivity mindset or health from david

[1:57:21] I saw this one and I thought about it and I'm actually not sure. But there's one thing that I probably... My husband's been trying to get me to... [1:57:30] to do this and i think he's right which is i very much am an inbox zero person um but i also do things like i will uh use superhuman to like [1:57:39] take all my newsletters and like, [1:57:41] I read them on Sundays, for example, but talking about this, like weight that the internet can create sometimes, um, [1:57:48] It's like... [1:57:49] requirement to like read all your newsletters or to like get to things. [1:57:53] I think is like not great for me over time. I'm like, why do I feel the need to respond to everyone? Why do I feel the need to read like every version of NumLock every single week? And it's like, I think actually I need to like reverse some aspect of that. It's really interesting. I'm the opposite. I like don't respond to any emails and I feel tremendous guilt and shame about it all the time. And I like want to get better. Well, I don't, I'm not necessarily good at responding to people quickly. I just like, I, I kind of put it away and it's like a weight on your life to be like, [1:58:23] you do have to like get through all these emails versus Cal is more just like, you know, the kind of person that, [1:58:29] He says like 10,000 emails and he's like, [1:58:31] what's the problem? And I'm like, there's a problem. [1:58:35] That's great. Well, I think this is a great place to end. We've covered so much. This is an incredible interview. We had more stuff planned. We were going to go into some more using AI for more research stuff, but we'll have to do maybe a part two. Yeah. After our warm baseball cap project. It's like a year from now. I'll have more business ideas for you. We'll find something that we can validate. Yeah. But this is really great. I really appreciate you coming out.

[1:59:01] This is so fun. Cool. [1:59:03] See you next time. [1:59:05] Music.