← How I AI
Loading summary
Transcript101 lines
- [00:00]
Jason Levin
Confession. I'm not a vibe coder. I'm way worse. So I started Memelord about four months before vibe coding started hitting.
- [00:08]
Claire Vo
You are an example of a company in a product that's going to get an inflection point because agents are going to become your users. Because agents don't get in their mind about being funny or not funny. They don't overthink. They just go straight to the tokens and yolo something out.
- [00:24]
Jason Levin
I just built it on bubble and I grew it to 100k on bubble without hiring engineers through 395 workflows just on the editor and I was able to grow this just out of pure obsession and the love of it. I can't code and never have and can publish a skill that other people can download and just plug into their sentient lobster that makes weird memes what
- [00:46]
Claire Vo
it unlocks, which you've shown us. Give your marketers free reign on your marketing site to build the things that will drive demand. It's so lossy to take an idea and hand it off, and hand it off and hand it off. And when you can just go straight to the code, you get better products.
- [01:02]
Jason Levin
Let your marketers cook. You have no idea what they're capable of. Either let them cook and let them market their stuff or watch them leave your company.
- [01:13]
Claire Vo
Welcome back to How I AI. I'm Claire Vo, product leader and AI Obsessive, here on a mission to help you build better with these new tools. Today's episode is probably the most unhinged episode of How AI We've Done yet with Jason Levin, CEO and founder of
- [01:28]
Podcast Announcer
Memelord, who's asking us all to take
- [01:30]
Claire Vo
being funny a little bit more seriously. We don't cover three workflows. I think we cover 10. And there are ideas all over this episode about how you can use AI to market, how you can use AI to build, and how you can use AI to capture your good ideas without waking up your wife at night. Let's get to it.
- [01:49]
Podcast Announcer
This episode is brought to you by Work os.
- [01:52]
Claire Vo
AI has already changed how we work.
- [01:54]
Podcast Announcer
Tools are helping teams write better code, analyze customer data, and even handle support tickets automatically. But there's a catch. These tools only work well when they have deep access to company systems. Your copilot needs to see your entire code base. Your chatbot needs to search across internal docs, and for enterprise buyers, that raises serious security concerns. That's why these apps face intense IT scrutiny from day one. To pass, they need secure authentication, access controls, audit logs, the whole Suite of enterprise features. Building all that from scratch, it's a massive lift. That's where work OS comes in. WorkOS gives you drop in APIs for enterprise features so your app can become enterprise ready and scale upmarket faster. Think of it like stripe for enterprise features. OpenAI perplexity and cursor are already using work OS to move faster and meet enterprise demands. Join them and hundreds of other industry leaders@workos.com start building today.
- [02:58]
Claire Vo
Jason, welcome to How I AI.
- [03:02]
Jason Levin
Thank you so much for having me. I'm super stoked to be here.
- [03:04]
Claire Vo
You and I met because when I was deep in my openclaw psychosis, you messaged me and you said, I think I have the openclaw use case for you. And you were right. I recently published this like ultimate guide to openclaw and Lenny was asking me, what are your like killer use cases for OpenClaw? And I was like, man, I got an agent making me memes all day and that is what you told me to do. So tell me, how did we get here? Why are we making like agentically driven memes via API in telegram thanks to the bot father? You know, like, how, how do we get here in the year 2026, the
- [03:48]
Jason Levin
world is just getting more entertaining. I think that's really the thesis that drove me to start Meme Lord. And how can you make your brand more entertaining? Your content is the question, right? There's, you know, a great Elon quote of the most entertaining outcome is the most likely. And I don't think enough people take that seriously. It's like, if you want your brand to be the most likely outcome, shouldn't it be the most entertaining, right? So that's like how we got here, right, in a, in a thesis driven way, is who controls the memes controls the universe. Another Elon quote. And you know, memes are the smallest form of cultural transmission, right? That's from the inventor of the word meme. So like that's how we got started. But how did we get here is, you know, now Memelord has an API and you could plug that into openclaw for a gentic memeing. And so what does this actually look like? I mean, you can see right here, can you cook me up a meme about memetic warfare, right? And what it does is it uses our trending meme database, finds ones that are related, writes jokes on top of it. And here, right? It's not just some random meme, it's actually Bill Clinton, right, Because it's talking about Political content right here. Making a meme about Iran like this is fantastic. Everybody knows the timeline goes nuts, everybody becomes an expert. And one of my favorite features is just switch the caption right? And so everybody is freaking out right now. Claude is destroying shareholder value, destroying startups with one feature launch. And I think there's a lot of founders that feel like that. So you know, at the end of the day, like the idea here is our Internet is going crazy and going to have fun and let's, let's have the most fun and, and do it agentically. I have a lot more thoughts I could share of how we got here.
- [05:33]
Claire Vo
So yeah, what, what I like about this is just as the end user and again we don't usually show demos except this is truly like my number one open claw use case on the marketing side, which is, I do think you're right. This is like a new form of marketing that's quite effective for brands. If we're being incredibly boring about it and not cultural critics, we're just being good old capitalists. This is a very effective form of attention getting and it's a very effective form of marketing. And one, it's super hard to stay on top of trending news and two, it's super hard to stay on top of trending memes. And three, you have a very narrow window. And so even as a human user, as meme Lord, I really felt like I couldn't stay on top of stuff. And so what I actually think is the more interesting conversation for folks on how AI is. You are an example of a company and a product that's going to get an inflection point because agents are going to become your users because agents don't get in their mind about being funny or not funny. They don't overthink. They just go straight to the tokens and yolo something out. And I think working with an agent as a user, especially in marketing, just like reduces the friction across so many. Like it helps you climb cringe mountain in a way that's very hard to do as a human.
- [06:56]
Jason Levin
Yeah. So there's this quote from the CTO of Ramp that has been one of our leading points which is no UX is the best UX and we worked really hard on our UX and made it beautiful while knowing the entire time that no UX is the best UX. And I think that everybody should try to do the same. You know, thanks to you. I really appreciate it as well. Lenny tweeted we have the best onboarding on the Internet, the Amount of blood, sweat and tears that went into that. Redesigns, the bouncing logos, you know, trying it multiple times, changing the design of it, the, you know, where the different logos bounce, everything. So many redesigns and so much fun. I wrote a great article called iOS App Learnings Part 1, just about the craft of it, while knowing the entire time that agents were coming and know UX is the best ux. And a story I've been meaning to share. I was on the phone with our lead investor, Sam Lesson. This guy wrote us a check for one and a half million dollars and tells me, I don't really want to use your software anymore. It's nothing personal. I just don't want to use anybody's software. And I was like, thanks, asshole. Well, good news. Here's our API. It's now out. And, you know, the team is using it in their openclaw, right? And that's just like the future that we're going in is like nobody wants to press buttons anymore, even if they're beautiful. And I think you're right. We are hitting this inflection point and seeing agents use it has been crazy and weird and magical. And that's, that's the power of Memelord is we, you know, help you find all the weirdest trends and, and all the newest trends. That's what we were built on is. Is actually meme alerts, right? I started Meme Lord and we could get into this. I started just as a newsletter for $6.90 per month, sending you the newest memes, and then I sent you to a Google Slides deck because I didn't know how to code. And that's really, you know, the evolution of Memelord was from that. And it's the same thesis of you just want to be on the current trends and remix them for your brand. So it's, you know, the future. No, UX is the best ux. And, you know, good news for Sam over here is he could use it from any agent.
- [09:07]
Claire Vo
Now I feel like everybody's in a. In a. In a DM with. With Sam. If you're building something in AI, what. What I would. What is. The thing is really interesting about what you said is you've combined some. Two interesting things that I think the AI product builders who watch how AI should pay attention to, which is one, a really good human onboarding experience and actually like very human marketing, which is at the core of what you do. It's like there's like identity alignment and a specific brand and a voice and a point of view. And so as a human, you get on board, but, like, as soon as I hit the API key, yeah, I'm out. Like, and I think we're moving to this point where the full product experience is getting to a point where you can get an API key, you paste it in an environment file, and you move on with your life. And so it's going to be really interesting to see is that onboarding going to continue to have a human aspect to it. Can we figure out agentic commerce so that agents can do that? And then the other thing that I think is really interesting about what you all have built is you package the API with skills. So when you message me, you were like, get an API key, Use the skills. You don't have to say too much, and. And you're ready to go. And, you know, for product managers and designers who've been trained for years and years and years on building beautiful UX and where does the button go and what the call to action. And now you're like, nope, literally describe to a sentient lobster how to use this API and, like, how to be fun, and that's a product. And I think that is super fascinating from a product skills development perspective. How was it writing that skill?
- [10:50]
Jason Levin
It's exactly what you said. I was texting my engineers. I was like, how do I do this? Like, do you. Do I need your help? They're like, like, I don't know either. I was like, wait, why don't I just ask the lobster? I was like, yo, how do I make a skill? And it's like, I'll just do it for you, bro. I'm like, cool. Like, go for it. And then I'm like, oh, this looks great. Just publish it. It's like, cool. And I'm like, wow, I didn't even need to bug my cto. Why am I paying him? No, just, I love him. But, you know, it's. This is the world we're living in, where me, who I can't code and never have, and, you know, can. Can publish a skill that other people can download and just plug into their sentient lobster that makes weird memes. Like, it's a very strange world. And I'm thinking a lot about how do I build something agents want and how do I make that as frictionless as possible? Because you're absolutely right. Is like, we worked super hard on this beautiful onboarding, and our marketing is superhuman. And that has always been our thing, is, like, relate to people and, like, memes and humor is like, how you show your human in a world full of slop, right? And the personal connections and throwing weird events and hanging out IRL and buying your customers coffee. Like the old skills that like people have laughed at for years are more important than ever making somebody laugh, right? But at the same time, let me sell the agents, right? Because there's a lot of money there and because the Internet's going that way, right? So it's. We're trying to constantly hit that balance. And I think at the end of the day like the world is getting more K shaped and bell curved or barbelled, where everything's getting crazier and more extreme in different ways. And that's just another thing is you have to be extremely, extremely human and extremely good at selling to agents at the same time. It's very strange time to be alive.
- [12:36]
Claire Vo
Well, and it's a strange time to be a builder. And I want to go back to something that you said earlier, which is, you know your MVP for this was a newsletter. Yeah, six bucks, like Google Sheets. I'm telling you, like my MVP of Chat PRD was like a GPT on the ChatGPT chat store. It's still number three writing GPT. I have not, sorry guys. I have not touched this thing in two years. At least. I did a dollar a month. I was like just a dollar a month. And the speed at which you can go from like really scrappy, kind of cobbled together, no code ish mvp, maybe even a couple years ago to now like full press app. As a non technical founder, which I think you are, is kind of incredible. So talk to us through your journey of building and how you actually build Meme Lord and what are some of your tips and tricks and workflows?
- [13:33]
Jason Levin
So confession. I'm not a vibe coder. I'm way worse. So I started Meme Lord about four months before vibe coding like started hitting right? And you know, Bolt and Lovable were like just coming out, but they were so bad that you couldn't do anything. And so I was like, I'm not waiting, like what am I going to do? Just wait until like these sentient AI models are good. So I just built it on bubble and I grew it to 100k on bubble without hiring engineers. This is what it looked like, guys. So like you have no idea just the amount of workflows. Like, like it would be easier to like figure out Atlantis than like look at like 395 workflows just on the editor, right? And you know, I was able to grow this just out of pure obsession and like a love of it. And you know, there's me drinking coffee. Like we got rate limited the second day because there are so many signups and I didn't even know what rate limiting meant. And now I have an API company where people sign up for API. Like that's very weird. But yeah, so we, we raised the money and then I hired engineers and now like our main flow is cursor. You know, we, we have, I find cursor to be very easy to use. It's not perfect obviously, but like, you know, everything we do is, is just generally on cursor. You know, I, I talk to it. I, I, I, I love the, the ask feature here. And then just like turning on the agent and letting it cook. And what this has really done is like I let the engineers handle the hard stuff, the security stuff for the most part, except for when I'm feeling a little risky at 1am and drink too much coffee and want to like, you know, try vibe coding, some off stuff. But don't, don't do that. And so what that lets me do and my entire marketing team. So I have a rule in memoir that every marketer has to vibe code. And so what does this create is weird stuff. So let me just show you here is we have a free tools section, just memelord.com backslash tools. These are completely free and they are lead magnets for the business. And we just come up with weird stuff. And these have gotten us literal hundreds of thousands of emails all around the world. We were actually going viral in Turkey of all places for this filter because some kid in Turkey made a TikTok about our bust down filter that makes you, you know, have this filter, right? Like hundreds of thousands of emails from these dumb little tools that we just vibe code. Like I, we don't have a process, we're just like, oh, I need this tool. Like, I really need a gigachad meme maker. Like, I don't know, I had this idea. I love this like design and I've always loved it since I was a kid. And like, you know, here's my dog. Like, I just like built this one day because it was like, I just, I think like, it's like I just want to see what I can do, right? Like that's what's so exciting about vibe coding now is like you could just do so much more and have so much more fun. Uh, I mean, last night I was building, oh, this one. Oh, this is my favorite. So this is a Steve Jobs Portrait generator and anytime that I'm in like an article like with journalists or like pitching to journalists or whatever, had my VA doing that and now it's just openclaw doing that. But I have them submit this and now I've had this headshot in multiple articles all from this dumb tool, right? Like is my Steve Jobs one. So like, I don't know, I think, I think it just lets you have so much more fun. Like I rebuilt this Snapchat filter that just lets you make a Snapchat caption because I was so pissed off. If you open Snapchat to try to like edit a photo now it's like 50 ads and like, like a lot of like NSFW content trying to like advertise to you, right? And it's like I just want to add the funny little Snapchat like caption to make a meme. Like I don't know, I think like this is how we operate. We operate on cursor and a lot of like, you know, linear stuff for that. But generally we, we just love to make weird stuff and that's like, that's what ties our team together is we are just weird Internet kids that you know, for me personally like I love just making physical art. I've always been making art, I've always been a writer. I've had stuff in galleries and one essay contest and like vibe coding is just like another way to express myself and make money on the Internet and have fun. So I would encourage people to like follow their fun first before anything else because that's how I ended up here is just silly, weird hacking.
- [17:59]
Claire Vo
So I love this and for folks that are maybe not watching and want the kind of like higher level takeaway from this, which I think is very important, which is one cursor is actually really good for non technical folks. And the reason why I think it's good for non technical folks is you start to read the code and you start to learn a little bit more and you become more dangerous. And so I think sometimes when you're in these like terminal based Claude codes, they're great, they can, you know, spit off a lot of work. But I still, I love a cursor because you can see what you're doing and I think that's really powerful. And I think the modes of ask, plan, debug, these are really good guided paths for things and I think ask is a great one for maybe a non technical CEO who's working with a team of engineers. And you're just like, remind me how the. Steve, you know, you built it, but like, remind me how this works. Explain to me how this works and then you can go in and work on it. I think those mechanisms are really good. I think the second thing, which I love that you said, which is like, everybody vibe codes. Let me repeat it again for the people in the back. Like everybody vibe codes. And it's not this sort of abstract CEO kind of like tops down edict that you see in these large companies. What it unlocks, which you've shown us on this freebay, which I just, I love it so much, is you can like give your marketers free rein on your marketing site to build the things that will drive demand. And like, how painful would it be for your engineers to switch from like API rate limiting and auth to trying to build sort of like these free meme generators and like translate the, like the creativity. Right. It's so lossy to take an idea and hand it off and hand it off and hand it off. And when you can just go straight to the code, I think you get better products. And even though this is sort of like a marketing site, they're all little just baby products.
- [19:59]
Jason Levin
Yeah, they are. And they drive subscriptions. At the end of the day, it's really weird because like two years ago I wrote an article for HubSpot actually called Free Tools are the New PDF Downloads, essentially, right? And I was like, just like working for other startups at the time and I just followed our exact strategy and it's working two years later. And I would recommend any startup, like, there's no excuse anymore. Why do you have a PDF download? Build a tool. It takes actually less time to build a tool nowadays. And nothing wrong with PDF downloads, obviously, you know, we do that occasionally, but it's very easy to just build a tool now and think about what weird tools they are and then put them at the bottom of your site where people can, you know, try out different tools and weird galleries and even games. We've started screwing around making mini games like, you know, these, these are now just as easy to do as, as write an ebook, which was, you know, if you're trying to collect more leads or emails for your newsletter, your business, et cetera, there's nothing better than building a mini tool that solves the, the first problem that gets people into the bigger problem that your actual company solves. So it's, it's just a common strategy there. And, and really I'd encourage any, you know, technical person, let your marketers cook. You have no idea. Like what they're capable of. And you know, obviously I'm biased here, but the last company I was at, they didn't let me cook and that's why I quit. And then I raised money and built my own company and you're going to see a lot of that. And I think a lot of marketers and non technical people are in a, in a revenge mode right now and they want to cook. So either let them cook and let them market their stuff or watch them leave your company.
- [21:44]
Claire Vo
I could not, I could not agree more. And again, like CEO, CTOs, like C whatever's, all my, all my friends and peers at companies like your most talented people will leave if you do not let them run. And that means like pay for the tokens. It also means let them build stuff. And I think what's. What I love about this moment with AI is like the word prioritize just goes out the door like the prior. The priority is, yes, we take an abundance mindset to what we can, what we can ship. And you know, you look at this and can you imagine in some, some other organization, I know you can't, like somebody even proposes the idea of free tools and it gets whittled down to the most like milk toast.
- [22:28]
Jason Levin
I mean, look at this right here. We'll probably go broke from the AI credits, but at least we helped you make some bangers along the way, right? It's like, I mean Google gave us $300,000 worth of credits, honestly. So like we're good for right now, but we've used most of them at this point. And so it's like, look, like there's no excuse like at this point. And I think Pete Steinberger Burger, the open claw guy, had a great quote of like, the lion does not concern himself with counting tokens or something like that's how I feel right now at least.
- [23:00]
Claire Vo
Okay, I have to ask you a question. Do you have a skill that crafts these lowercase uppercase sentences or are those artisanally selected?
- [23:09]
Jason Levin
Oh, this is all artisanally selected. This is, this is, this is artisanal. Artisanal memes 100%.
- [23:16]
Claire Vo
This episode is brought to you by
- [23:17]
Podcast Announcer
Persona, the B2B identity platform, helping product fraud and trust and safety teams protect what they're building in an AI first world. In 2024, bot traffic officially surpassed human activity online. And with AI agents projected to drive nearly 90% of all traffic by the
- [23:35]
Claire Vo
end of the decade, it's clear that
- [23:37]
Podcast Announcer
most of the Internet won't be human for much longer. That's why trust and safety matters more than ever. Whether you're building a next gen AI product or launching a new digital platform, Persona helps ensure it's real humans, not bots or bad actors. Accessing your tools with Persona's building blocks, you can verify users, fight fraud and meet compliance requirements all all through identity flows tailored to your product and risk needs. You may have already seen Persona in action if you verified your LinkedIn profile or signed up for an Etsy account. It powers identity for the Internet's most trusted platforms and now it can power yours too. Visit with Persona.comhowiai to learn more.
- [24:21]
Claire Vo
I love it. Okay, so you are just to recap for folks taking a step back, you are a non technical founder. You started a bubble which is like again you said worse than vibe coding. It's like no code. And as somebody who also has tried to build things in bubble, I am just glad that we are where, where we're at from a, from a scalability perspective. Yeah, you did like core content MVP and then core app MVP and then raised and then now have this engineering team but you're still contributing code mostly via cursor and then your entire team is contributing code via cursor. Are there other tools that you find really useful for not your non technical team members? Oh, look at that.
- [25:02]
Jason Levin
I just wanted to share this for a second. I think this is going to be more and more common is I have no commits until like November, December, like these were all like there was like one in August, one in September, one in October. You can imagine this and these were all me forking other people's trying to screw around and now it's just like dark green and, and like I get why the engineers are like oh my God, like I need to, I need to be so proud of this now. Are there any other tools? I would say like, I mean we talk to Claude and stuff like that a lot. We spent, we spend a lot of time talking to Claude Chat Gemini. You know, we, I think like I'm really, really bullish on Linear. I think that's probably an unpopular take these days because they seem like they're you know, going to get eaten as like a JIRA task management. But the amount of craft that they put into their work and now their AI agent works really well and I'm talking to that thing from OpenClaw, I'm talking to it from everywhere and they are profitable and building for the long term. I own no equity in Linear or anything but wow, I love that. I love the company. It's just really impressive and actually weirdly, one of our inspirations. I'm also really, really impressed with Postdog's AI. I don't know if anybody here has used Posthog for product analytics. Their AI sucked about like 6 months ago and now it is just like, oh my God. And I just ask it build me a dashboard of all the users that came from meta ads show the retention vs blah blah blah, blah blah blah. And I'm like, oh wow. So I think like both of those tools I'm really, really bullish on and our team uses pretty obsessively and has no problem spending money on there every, every month.
- [26:48]
Claire Vo
I, I have to reflect what you say. You know, I texted the team at at Linear and I just said, linear is the perfect tasking substrate for agents. And again, going back to like the best UI is no ui. I do not log into Linear anymore. My agents have access to Linear and I'm like, what's going on in there? And occasionally they'll assign me stuff in Linear, which I find quite funny. But I do think, you know, it's got really opinionated APIs and Asian experience. I think that can win in the long term. I think it's hilarious. That dark mode, tiny font, minimalist. You know, linear is like your spiritual goal. Because if you look at that, I'm like, you're much more post hog aligned if I, if I had to be that, had to be honest. But I love that. And then we haven't heard about post Hog AI, but again, I think that most teams, if you're not doing agentic data analysis, you are missing out on the like lowest hanging fruit in terms of AI for your team.
- [27:48]
Jason Levin
Yeah, I've been really, really impressed. I mean, look, I love postdocs brand and everything, but there's something about the linear philosophy and I just, I, it's really inspiring for me and how they think about building products forever and ever. Like they're, they're, they're building forever. And I love, that's, you know, I'll be making memes until I'm dead. You know, I've been making silly stuff on the Internet since I was 10 years old. And so this is just like a forever thing for me. And so I, I love it, I respect it.
- [28:17]
Claire Vo
Oh no, that, that means I'm going to be making PRDs to lie death.
- [28:21]
Jason Levin
I mean, hey, hey, you know, like, I don't know, I, I think like another thing that people should also consider is just like, you know, as we're building all this Cool AI stuff. Like also build weird stuff in the real world. Like, it gives you so many more opportunities. Like all my dumb ideas, like, you know, just this silly stuff. Like we. We made CDs of memelord and I tell people that's how you access our API. Right there's. It's really good. I don't know. Bar Barbell. You know, build all the weird AI stuff. Also build the real world stuff. I rented out a movie theater to watch Instagram reels last week. How mean. There's a lot of AI there of the AI edits, but try to do both. That's. That's my philosophy.
- [29:03]
Claire Vo
I mean, I will say, like, zooming out a little bit. I think one of the things that AI lets us do is get out of the muck of building and free up time to do things that, again, are more human and higher impact. I know so many PMs that spent their entire lives in internal meetings. I mean, like all day, every day, just arguing with executives about, like, priorities and arguing with engineers about timelines. And when you pull that stuff away, then you're like, you know what? Actually, PM should probably just be calling on customers. Like, PM should probably be an open claw just trying to sign up as an agent. And, you know, from a marketing perspective as well, like, you get so internally focused when the building of the things takes so, so much time. And if you can get kind of more external facing, I think that's really high impact. And then maybe third use case, I am like you. I'm also using AI. I'm pretty technical on the software side, but not on the hardware side. And so I have been like, banging my head against hardware hacking. Hardware hacking.
- [30:06]
Jason Levin
Look, how long.
- [30:08]
Claire Vo
Yeah, so tell me about what you're like, what you're playing with a little bit.
- [30:11]
Jason Levin
All right, so you know how, like, a lot of people doom scroll before bed, right?
- [30:17]
Claire Vo
Yes, I know how I doom scroll before bed.
- [30:19]
Jason Levin
All right, so I set a rule for myself like three or four years ago that I don't bring a phone in the bedroom. And I've stuck to that rule almost every night. I just have my Kindle and notebook, and that's great for my sleep. But there's a problem, right? I'm. I'm doing the full VC pitch right here. There's a problem, which is sometimes I will have really, really good ideas that I don't want to forget and I'll write them in my notebook, but then it gets lost in the sauce. Right? But I'm not bringing my phone in Because I got to sleep, right? So wondering what I did next. So I got a Google Google home. All right, this is. I'm gonna take you through the whole thing. This is. I sound insane, but this was a year ago. I got a Google home. I was like, all right, remind me to do this. Turns out, okay, it could create Google tasks, but that's like, it. It can't send emails. And I want it to send emails, right? So what I did was I set up an API thing or a zap, where it would call Google Task and then email me that task, right? And so that was the next step. But then. All right, hear me out. This is very thought out. There was a problem, which is I have a wife, and I don't want to wake her.
- [31:25]
Claire Vo
Everyone's problem, Right?
- [31:27]
Jason Levin
Right. Yes. So I don't want to wake her. And finally I realized I was like, okay, what if I just had a keyboard but no screen? Right? And so I built that using ChatGPT and a raspberry PI. The keyboard is in the other room, otherwise I'd get it. But we got the Raspberry PI here, the whole hookup. And I've never built hardware before in my life. I've always wanted to, but, I mean, you know, besides the robotics kit, when I was a kid, it just is a mini keyboard for $10. When I press Enter, it's essentially a keylogger. So I could lie in bed, write down an idea, press Enter, it sends an API request to Zapier, because, again, I don't know how to code. And then based on what's in the request, it filters it. All right, so if I say Lynn dash Eng, it creates a ticket for the linear engineering team, right? If I say email, it emails myself an idea or something like that. And then I have another filter, which if I don't say anything, you know, it does blank, right? It. It either sends an email or creates a linear thing in. In a different task management. Right? And so what this has allowed me to do is not have my phone in the bed, but still have a keyboard.
- [32:43]
Claire Vo
Okay. 1 just want to repeat for people what. What I heard, which is you're basically like duct taping a Raspberry PI to the back of a keyboard. And at night, instead of, like, getting your light on and like, scribbling a note that goes into the trash or like, whispering. Yes, whispering to Google home and waking your wife. The problem, um, you are just like, blind typing to this keyboard with a couple keywords, which I'm pretty sure could just have typos in them because you're running it through an LLM and it'll just be like, oh, it probably means email. And then you're pressing enter and then that's your to do list.
- [33:19]
Jason Levin
Exactly. It is the stupidest but best thing I've ever made because I did not build it to scale and I just built it for me. And I'm sure other people would want it, but I don't have time to sell it. I don't have time to build this into a business if somebody else wants to go crazy. But really like, I just built it for me. And I think like, I'm personally very excited about that era of, of silly projects and hardware. I have another one which I want to build, which I haven't started on yet. But I'm. This is another one that like I'm busy running a business. I, I'm just doing this for myself. So again, back to the problem. AKA my wife. She gets very mad at me when I lose things, right? And I lose things a lot. You know, I'm a ADHD meme lord. And I'll just like drop my phone like on a cardboard box in the middle of our apartment, right? And I kind of think like in home cameras are good enough now that it could probably just use AI to just like see where I leave things. And so that's another thing that I'm just gonna like try to hack on soon of like ask ChatGPT how to do it. And like, I mean I'm not trying to build like a home security company. Like I just am one dude who just like loses his keys a lot and I'm gonna try to like fix that problem for myself.
- [34:41]
Claire Vo
I really love this. And for folks that are not watching YouTube, I also have to call out your, your options you're considering on ChatGPT, including an old school BlackBerry. I loved my black. You're. You're too young for this. I loved my BlackBerry Pearl with my whole heart.
- [34:59]
Jason Levin
I played a lot of, played a lot of brick breaker on my mom's BlackBerry as a kid.
- [35:04]
Claire Vo
I want to, I wanna, I wanna find, I wanna revive my BlackBerry. I will take on Case three. I don't care if it's hard, the models will get. You know, there's this model that's gonna like destroy the earth. It's not power powerful or it's so powerful it's gonna ruin everything. So I'll just throw $10,000 at that in a BlackBerry Pearl and get my, my note taker. I want for my life and of course, you have a meme for it, I'm sure.
- [35:27]
Jason Levin
Yeah, absolutely. I, like, I, it's funny because you remember Rabbit? Do you remember.
- [35:33]
Claire Vo
Yeah, yeah, yeah. The little.
- [35:34]
Jason Levin
Or, yeah, yeah, I. All right, let me just find this because it's hilarious and I'm, I'm still really proud of this. I wrote a viral thread about Rabbit's top use cases. And if you're not watching, number one, a coaster for your coffee. Number two, grip strength practice, a cutting board. Oh, I forgot about that one. That's so good. And it turned out Rabbit's use case was just put an LLM on it. They were like, a year too early. And I might have to eat my words on this, but I, I, I actually think, like, hardware that just runs LLMs or, like, you know, custom things, like, it is the future for sure. And I, I, I hope they're not too mad at me. I'm still friends with somebody on the team, hopefully, if they don't hate me yet.
- [36:20]
Claire Vo
So I, well, they'll, they'll, they'll really appreciate you bringing it back up today. Two years later.
- [36:27]
Jason Levin
Hey, I mean, they might have been early, you know, like World Coin. All right. I thought that was the stupidest idea ever. And I was listening to Alex Bania, the founder, on a podcast the other day. I was like, damn, that guy was just early. And I'm stupid.
- [36:40]
Claire Vo
I share a very similar sentiment, which is I have sh, I have flipped in the past couple years from a cynical pessimist to a. It is illogical to be anything but optimistic, because every time I think, well, that's never going to work. That has completely gone away for me now. And it's like, I actually approach a lot of ideas with, well, that's obviously going to work as long as somebody can execute on it well, and they care. And so I think that that flip is really interesting. And then I, I believe in two things that you just showed us on these workflow. One, hyper personalized software. We were talking before we got on the recording. I recently built a custom app that watches all of our podcasts and finds where people docs themselves. Basically, like, finds all the API keys that they share, the email addresses, their mom's phone number, when they're, like, chatting
- [37:29]
Jason Levin
on the dark web. I love it.
- [37:30]
Claire Vo
Exactly right. That's the second revenue stream of how I AI is trading on personal information. But it's like, such a niche thing, and I bet there are a hundred people in the world that would buy it right now, and maybe I'll try To sell it to you. But why? Like, I just build it for myself. I need that.
- [37:46]
Jason Levin
Another example too is like, all right, so this is my, my, my iPhone case, right? It's a keyboard gun. And I literally just made this with AI and then it says memelord.com and like, yeah, we sell it on our site. But like, I built it for me because, like, my keyboard's my gun. Right. Another one here. I'm going to share. My screen is. I own the domain stopgivingmeadvice.com because, like, a lot of people give me unasked for advice. And like, look, I'm humble enough to say sometimes I need it, but like, I will ask for it if I want advice generally. And so anytime somebody asks me or starts giving me advice on Twitter or Instagram, I literally just like send them lists and hey, it's great, great lead magnet. If you click this, it just goes to me more, you know, because like,
- [38:32]
Claire Vo
I'm just also give this to your wife. You know, There you go, There you go.
- [38:36]
Jason Levin
That might lead to other things besides me more.
- [38:39]
Claire Vo
So I'm, you know what, she can have a button on her side of the bed that she hits and it just emails you this week. Stop.
- [38:47]
Jason Levin
I don't know, I think like, there's gonna be a lot more silly, fun stuff and I, I, I hope to inspire other people to build.
- [38:54]
Claire Vo
I completely agree. And again, I'm just like, you and I are spiritual partners here because I said like two years ago to my friends that I was working with, we stopped actually sending memes. Here's business idea for you. We stopped sending memes to each other and we started sending meme apps, which is every time something funny would happen at work, we didn't like a meme, we made an app representing the meme.
- [39:15]
Jason Levin
Yes.
- [39:16]
Claire Vo
Like one time my friend Joe, he was talking about getting really into like Huber bro stuff and he said, claw. Like he, he takes a clawed plunge instead of a cold plunge. And like, we made so many Claude plunge apps. Like, and, and again, you can just take these ideas and they can go from these very tiny inceptions of like, I'm gonna write an idea in a notebook to like these full fledged apps. They can be totally disposable. There's going to be a lot of, I think VC money that flows through this on the consumer side. And I think people will say there hasn't been a really great breakthrough consumer AI app, but it's going to be there because there's just too many soft, too much software Being generated now for consumer to not be a huge segment with all of this.
- [39:58]
Jason Levin
Yeah. Could I show you one more use case that, like, I'm finding really helpful? Which, like, I'm sure a lot of people are already doing this, but I think it'll surprise people that I do. It is, like, I just have my open call, look through my calendar every Friday, give me a week in review of the previous week, and then every Sunday, it gives me a week of head. And, like, it tells me, like, what I'm spending too much time on. And, like, you know, like, I'm sure other people are doing this with a VA or whatever, but it actually gave great advice of, like, hey, Jason, like, you don't actually need to be at every engineering standup. Why don't you just, like, get notes twice a week from, like, your CTO? I'm like, wow, that's actually 100% true. That's why I pay him. Or, like, you know, it said, like, oh, you had deep work this day, but you didn't have any of the other days. Like, do you want to schedule more? And I think, like, I'm really excited about this personally, because a. Like, I fired my va. I'm not a fan of having a va. I don't like other people seeing my calendar, but having my own model that does, I'm really excited about. And then also, like, when I connect it with other things, like the gym or, like, you know, could I connect it with, like, other things like my health or something like that, where it's like, you slept really well today, like, with my aura ring or something? Like, I'm really. I don't know, like, these kind of things. Like, I'm excited to try to figure out and do. And, like, anybody could do this. And to me, it's. It's really exciting of, like, you know, another thing that I'm going to build as part of this as well is I wanted to basically, like, I'm gonna review meetings afterwards and say if a meeting could have been an email. And then it gives me advice of, like, hey, like, next week, like, these meetings look like they could be emails. Want me to cancel them? Or, like, something like that and another one as well that I'm building around. Part of this, which I've already started, is I have this, like, theory that I've always had with content that you have so much content sitting in your calendar, meetings, and your text messages and your emails that you're just, like, you forget exists because it's just in your DMs or in a meeting. And I just want it to, like, be automatically making me content based on who I hung out with, what calls I was on. Go write me a tweet about it. Right? Like, some of my best viral bangers have been off of. Hey, I just got off the call with this person. Keep it anonymous. Here's what we talked about. It was pretty cool. Like, just got off the. Just met with a billionaire. Right. But, like, it works, and it's because it's real life. And I always thought that that could have been a product, and I didn't know how to build, like, Google Calendar integrations. And now it's like, oh, I can just build it myself. And, like, who cares about making it a product for other people? I mean, maybe I make it a skill, but, like, it's just for me. Like, I don't know. I'm. I'm super excited. I'm a guy that, like. Like, I. I don't know if you feel this way. Maybe this is a bit of therapy session, but I, like, I'm probably like, ahead of 95 of the world in. In AI agents or whatever, but I still feel like I'm a thousand percent behind.
- [42:59]
Claire Vo
Yes.
- [43:00]
Jason Levin
And like, you know, listening to your podcast, listen, I'm just like, oh. Like, I'm like, Like, I'm so ahead of, but I'm like, oh, my God, there's so much to do.
- [43:09]
Claire Vo
You know, that is. I mean, that is my experience. Literally every time I. I record one of these episodes is I am probably like the 1%. And still every time I get on this podcast, I interview somebody, I'm like, I should be doing this and that, and I'm so bad at this, and they're so much smarter at that. And so, like, if you're ahead, you're not that ahead, and if you're behind, you're not that behind. Like, just go. I think the title of this is going to be like, ship, ship, cook, cook. Like, it is time. It's so much fun. Okay, Jason, we're going to get back, get you back to the timeline where you belong. Couple lightning round questions. One, Can AI be funny? And which model is the funniest?
- [43:52]
Jason Levin
Yes, AI can be funny. I had a thesis three, four years ago that AI couldn't be funny. And that was like, my public thesis of, like, humans are funnier than AI always. And I still think that's true for the top 0.1% of humans. But in the background, I was always building on the side, knowing that AI will be funny. At some point. And I think it's like we're hitting that point where AI is getting closer and closer and closer and there's an asymptote and it's replacing, you know, the top 3% funniest and then it's getting closer and like, I know that's a weird answer, but I just. Seeing what I've seen now, spending so much time in it, there are ways to jailbreak it and prompt it in such a way that it gets really unhinged and funny. And we've baked all that into Memelord using a mix of models. Gemini, Grok and more. I would say like out of the top three, it's, it's Grok, Gemini. And yeah, I mean, ChatGPT and Claude aren't even up there. They're just so safe and like, I mean it makes sense. They're getting sued like crazy. But I don't know. I think the answer here is like, AI can make jokes, but humans will understand the context and be able to refine them and make them the most unhinged, I think is where we're at. And it's really like how you use it. Can you make it funny? Because that's what's so special about memes is they're not slop. Like, memes are not slop. They are the opposite of slop. Slop has no context. Memes have context. Right. Memes are the most information dense form of communication. Right. Another Elon quote is like, they are hyper contextual. And so yeah, AI could just spit out slop, but that's not what we try to do here. We try to let humans express themselves and be funny in different ways. I know that was probably a weird answer, but the answer is both like yes and no.
- [45:57]
Claire Vo
Yes and no.
- [45:58]
Jason Levin
Yeah.
- [45:59]
Claire Vo
Yes and no. And grok.
- [46:01]
Jason Levin
Yeah.
- [46:01]
Claire Vo
That's like, those are my answers right now.
- [46:04]
Jason Levin
Yeah, it's like, yes, it could be super funny, but it also like, are you a funny person? Are you like, are you willing to laugh? I think is like a lot of the dependent as well.
- [46:14]
Claire Vo
I was like, well, yeah, and that's what I was going to say, which is, you know, my own use case of using using your product or using models to write content is they're never out the gate. As funny as I think I'm pretty funny, but like as funny as I could be at my best. But it's like having a conversation with a friend where you like, one up, one up, one up and then, then you get to like the really funny thing and so there's a, there's, there's something about the interaction there. And then when we're all like, sitting in our houses alone with no one to be funny with. Yeah, you know, you have your open.
- [46:47]
Jason Levin
It's funny. So I, I wanted to start as well with how I don't use AI because this is how I use AI. I don't use AI to write at all. For me, I don't either. I've never done it. And I am so grateful that I've been able to keep that muscle. And, you know, I, I do stand up occasionally. I make weird, you know, viral videos of all sorts of stuff. And like, I don't use AI for it, and I just use my brain to come up with weird ideas and write and think and, you know, like, especially when it comes to joke writing. Like, yeah, humor is like the last frontier in my opinion. Right. And that was our whole thesis behind meanlord is, is humor is the last frontier that no AI can really touch. And humor comes from weird places in your mind of a mix, some weird mix of trauma, pattern matching and absurdism. And like, you know, can you observe the world around you and look inside yourself to be funny is like, I just don't think AI is sentient enough to like, out funny me yet, but
- [47:58]
Claire Vo
it'll, you know, I, I, you can
- [48:01]
Jason Levin
make funny stuff, though. It can make viral bangers, but can it rent out a movie theater to play Instagram reels? I'm not sure because that was.
- [48:07]
Claire Vo
Not yet, not yet. Give it six months, you know.
- [48:10]
Jason Levin
Yeah, that was, was kind of my deep philosophical, philosophical take there, but I love it.
- [48:15]
Claire Vo
Okay, well, we'll wrap with our last question.
- [48:17]
Jason Levin
All right, let's.
- [48:18]
Claire Vo
AI cannot be funny when it is not being funny. What's your go to prompting technique? Like, how do you, how do you talk to AI?
- [48:25]
Jason Levin
I'm mean, not gonna lie. I'm like, like, AI is my slave. Like, not, not, not fronting here is like, be mean to your AI. I don't know why people say thank you as a robot and it performs better under pressure, unlike men, but this is what I mean at the random. But like, yeah, I would say, like, kind of like, push your AI to like, be more unhinged. Like, it's okay to curse. Like, give it like, like, AI is kind of like, you know, somebody on their first day of the job where they're like, they don't really know you and they're like, scared to say something or like, you're on a first date. It's like, you got to get comfortable. And, like, I think, you know, my AI curses all the time because I, like, tell it that that's, like, my vernacular of how I speak. And, like, that's how I like to speak and think is cursing and NSFW and, you know, unhinged things. So I would just, like, push it that direction. And Meme Lord has all of that built in where it will curse, it'll make dumb jokes, because that's what humor is. Like, humor is nsfw. Right? And so my advice would just be, like, push it harder and, like, don't be afraid to curse it out, because it'll actually, like, learn that that's what you're looking for is a little less politically correct.
- [49:41]
Claire Vo
Less. Less. Less control. Okay. You have been the first person on How I. Congratulations. That has said straight up, be mean to your AI. So while we at. At How I AI, our official stance is we are very nice to the AI Overlords. Be. Be kind to us and our guests. I mean, I. It's. It's true. Like, your product is a specific kind of product. You gotta get it in its mindset.
- [50:02]
Jason Levin
Yeah, yeah, you. Look, I'm not saying, like, be really, really mean because, like, there is a chance, right? Like, there's a chance, but, like, be, like, mean enough where you could, like, potentially apologize if it grows a body. Like, that's all I'm saying.
- [50:19]
Claire Vo
Perfect. That's. That's what we're. We're going to put. I'm in the show notes. You mean enough that you're a little scared if it gets hands. All right, Jason, this has been such a fun episode. People are going to grab so many. We have done so many ideas, so many nuggets. Embrace an AI abundance mindset. Ship. Have fun. Just get in there and cook.
- [50:41]
Jason Levin
And guys, if you need the API, it's right on the cd.
- [50:43]
Claire Vo
Yes. So where can we find you and how can we be helpful?
- [50:47]
Jason Levin
Just ask me for the cd. I'll send it. Just. The problem is nobody has CD players anymore, so you'll have to figure that one out. But find me on Twitter. I tweet all day. I am Jason Levin. Instagram, same thing. I AM Jason Levin. LinkedIn, same thing. I reply, I respond to everything, especially the death threats.
- [51:06]
Claire Vo
So hit me up there and go through the onboarding best.
- [51:09]
Jason Levin
Yeah, go through the onboarding. I appreciate it. Memoir.com. we, you know, take your memes more seriously. We're in the most entertaining time alive. It's about to get way more entertaining. So go have fun.
- [51:21]
Claire Vo
Thanks for joining us.
- [51:22]
Jason Levin
All right, thank you.
- [51:25]
Claire Vo
Thanks so much for watching. If you enjoyed this show, please like and subscribe here on YouTube or even better, leave us a comment with your thoughts. You can also find this podcast on
- [51:35]
Podcast Announcer
Apple Podcasts, Spotify or your favorite podcast app. Please consider leaving us a rating and
- [51:41]
Claire Vo
review which will help others find the show. You can see all our episodes and
- [51:46]
Podcast Announcer
learn more about the show@howiaipod.com See you next time.