← The My Wife Quit Her Job Podcast With Steve Chou
The My Wife Quit Her Job Podcast With Steve Chou
631: Sabrina Ramonov Shares the AI Playbook That Got Her 500M Views in a Year
00:46:28
In this episode, I sit down with Sabrina Ramanov, a CS major turned social media entrepreneur who walked away from the traditional career track and built an audience of over a million followers across multiple platforms. -
Loading summary
Transcript132 lines
- [00:00]
Steve Chou
Welcome back to the podcast, the show where I cover all the latest strategies and current events related to E commerce and online business. In this episode I sit down with Sabrina Romanov, a CS major turned social media entrepreneur who walked away from the traditional career track and built an audience of over a million followers across multiple platforms. We get into her exact strategy for growing from zero, how each platform is different and where you should actually be spending your time, and why she thinks most people are completely wrong about what it takes to build an audience. And then we go deep on vibe coding, how the average person can actually use it effectively, and how I personally built a Shopify app in a weekend that saved me 300 bucks a month. But before we begin, I want to let you know that tickets for Seller Summit 2026 are now on sale over@sellersummit.com and if you sell physical products online, this is the event that you should be at. Unlike most events that are filled with high level fluff and inspirational stories, Seller Summit is all about tactical step by step strategies you can actually use in
- [00:56]
your business right away.
- [00:58]
Every speaker I invite is in the trenches. People who are running their own e commerce stores, managing inventory, dealing with suppliers
- [01:04]
and scaling real businesses.
- [01:06]
No corporate execs and no consultants. Also, I hate large events so I intentionally keep it small and intimate. We cap attendance at around 200 people so you can actually have real conversations and connect with everyone in the room. We've sold out every single year for the past nine years and I expect this year to be no different. It's happening April 21st to 23rd in Fort Lauderdale, Florida and if you're doing over 250k or $1 million in revenue, we also offer a private mastermind for high level sellers. Right now tickets are the cheapest they're ever going to be. So if you want in, go over to sellersummit.com and grab your ticket.
- [01:40]
Now on to the show. Welcome to the My Wife Quitter Job podcast. Today I'm excited to have Sabrina Romanov on the show. So Sabrina jumped on my radar screen in just the past year where all of a sudden my Entire feed on TikTok was suddenly flooded with her a with her content, both AI and non AI. And you can tell that she's using her AI avatar because she's wearing this purple cap. Sabrina is a serial entrepreneur when she where she started and sold her AI company Curious several years ago and since then she's been on a mission to teach 1 million people about AI for free. I'm pretty sure she's exceeded 1 million people at. But she also started this new AI company called Blowtayto that helps scale your organic social media, which I am actually happily using right now to handle all of my social media posts. And it's been a game changer for me. In fact, the only negative thing that I had to say about this woman is that she graduated from Cal, but we'll let that go for today. But in this episode, we're going to talk about vibe coding. We're going to be talking about scaling out your social media with the power of AI and with that. Welcome to the show.
- [02:55]
Sabrina Romanov
Yeah, thanks so much for having me. Yeah, go Bears.
- [03:00]
Steve Chou
Go Cardinal. Okay, so Sabrina, your posts are all over my feed. And I saw today I saw a post from you about skipping out on college to do an apprenticeship or just to create like 12 different projects. I'm afraid that I'm going to have to revoke your Asian membership.
- [03:20]
Sabrina Romanov
I knew that would be a little spicy.
- [03:24]
Steve Chou
Well, I, I know that you went into debt, right? Going to college. I mean, do you really? I, I kind of feel the way you do, but I'm not allowed to say that publicly.
- [03:33]
Sabrina Romanov
And the thing is, I do make an exception for the top colleges. Like Berkeley was considered top college, obviously, whatever one you went to, I'm just kidding. Stanford, like absolutely insane network. Just incredible people you're surrounded by and opportunities that are incredible. And so there is kind of like a soft spot in my of the top universities, but there are many other universities that are charging those price, those same prices, let's say 200. I graduated for everybody who knows, $250,000 in student debt. And it was really interesting for me actually when I made the Forbes 30 under 30 list, there was like a private survey of demographics of like who else had graduated with student debt. And it was actually a very tiny, tiny, tiny percentage. And I was like, oh, maybe there's a correlation between who people who make these lists. And there's a socioeconomic correlation with like how easy some of this stuff is attainable to you if you come from a background of where you're financially well off in the first place. And I really love Berkeley because it had that socioeconomic diversity for me that was like huge, huge. So my co founder for my first company and he's now my husband George, he transferred into UC Berkeley from a community college in California. And so that was, yeah, that was super cool. Very different worlds we come from. But yeah, I knew that post would be spicy. But I also really believe college, especially today in the age of AI is not really adapting to all of these things you can do now with AI. Most curricula still don't teach kids how to use AI and not college kids either. And I do think there's a lot you could learn just by taking initiative and like building these projects yourself or apprenticing under somebody that you think is really cool. Right. And for a college kid that could have all, it could be content creation, influencing being a YouTuber, could be building apps. Whatever it is, it's just like go apprentice for that type of person and like really see how they work, what it takes to succeed in that environment, and like try to do it yourself. So I'm a big advocate of just doing that as well, not just studying theoretical stuff in the classroom.
- [05:39]
Steve Chou
I am very pro AI, so I actually have a daughter who's applying to college right now and it seems like the schools are almost against it in a way. Like they have anti AI tool. I mean, this is going to be a part of everyone's lives. So I've taken a proactive AI approach with my kids as well.
- [05:55]
Sabrina Romanov
But yeah, and like I made another post on this. I think it's a shame that most schools aren't teaching kids how to use AI. So kids end up using AI in a cheap way, like just copy pasting and plagiarizing. And then when they graduate, they're expected to know how to use AI for everything because they're young. So why don't they know how to use AI for everything? But they weren't like trained or educated with this during college. So I think it's a huge. I think it's a really big disservice, to be honest.
- [06:25]
Steve Chou
I agree with you. But Sabrina, so it sounds like you had a decent exit from your last company. Why did you decide to create one in the social media space? And what made you want to even reach like a million followers and teach AI in the first place?
- [06:41]
Sabrina Romanov
Well, so when I moved outside of a tech bubble, so I was in Silicon Valley for over a decade. And then when I moved outside of the tech bubble, it became glaringly obvious to me that nobody knows about AI. And the things they do know are often driven by like fear and anxiety. And that really bothered me. When I opened TikTok a year and a half ago and I started creating content, it was like, you're missing out. Also, buy these prompt packs for like a thousand dollars. It was a combination of like FOMO and ripping people off. And I really did not like that. That did not resonate with me at all. We were just talking about kind of that pay it forward mentality that's prevalent in Silicon Valley, and I felt like this was my opportunity to pay it forward. I had been in AI since 2013. My first company was also an AI company. We did speech recognition, natural language processing, and I just felt like I was in a position and also enjoyed teaching other people. Like, hey, here's all the cool, positive, productive things you can do with AI.
- [07:40]
Steve Chou
It's just funny because I'm an ee, I have a master's in E, and I believe you have a CS major. Social media is not something that I would. Wouldn't be my first instinct, you know, to go into. So that's why I asked.
- [07:52]
Sabrina Romanov
Oh, yeah. So I, you know, I've listened to Gary Vee and Hormozi for years, and now I understand what they say. You know, most people are going to listen to this and never take action. Like, there's this element of, like, fear and, honestly, insecurity. When I started, especially being female on these platforms, like, I don't wear makeup, I don't doll up, I don't dress up or anything, but really, you are facing a lot of your own insecurities in your mind. That first 100 days of posting content, and even for my tech friends, like, they all looked down on TikTok initially. It wasn't until I started growing on Instagram, they were like, oh, that's cool.
- [08:29]
Steve Chou
Interesting. Okay.
- [08:31]
Sabrina Romanov
Yeah, so I dealt with all of that, but I just feel like my position is I'm an educator. What is the most efficient way to reach many, many people with exactly what I want to say without having to go through some gatekeep. And so when I thought about it in that way, I was like, well, social media makes the most sense. I can talk about what I want to talk about. The algorithms will try to find my audience for me. Like, people who actually appreciate the stuff that I'm saying. And it's free. Like, anyone can get started. You don't have to, like, pay for pr. You don't have to pay for anything to start posting on social media and teaching folks what you want to teach.
- [09:05]
Steve Chou
Yeah, totally. I actually, I get a lot of flack from my Stanford buddies. Like, they make fun of me nonstop. But, hey, you know, I found my tribe, you know, who wants to listen to me, so. So, yeah, so what I wanted to do today, since you've been very good at growing your audience, primarily this podcast serves, like, business owners, and I want to get your strategy for Building up a social media account from complete scratch for, let's say like an e commerce company or a services based company using your strategies.
- [09:40]
Have you ever wondered how much your
- [09:41]
business is actually worth?
- [09:43]
Now, I sold one of my businesses through Quiet Light and honestly, just getting that initial valuation changed everything for me. Not because of that number itself, but because of what came with it. My advisor walked me through exactly what buyers would be looking for, how I needed to restructure my accounting, what documentation I was missing, the gaps in my financials that might kill a deal before it even starts, and stuff that I really had no idea that mattered when
- [10:07]
it came to selling a business.
- [10:08]
And here's the thing, I wasn't even ready to sell yet. But knowing what I needed to fix meant that I could actually start preparing. And I now had a roadmap. Everyone at Quiet Light has built or sold businesses themselves. So my advisor told me what needed to change. It was actually coming from real experience, sitting across from buyers. And by the time I was ready, everything was positioned right and we attracted serious buyers. So if you've been thinking about selling someday, even if that day feels way far off, just getting a free valuation from Quiet Light will make a huge difference. You'll learn what you need to fix right now so you're not scrambling later. And if you're interested, go over to quietlight.com
- [10:48]
Sabrina Romanov
yeah, so I can recap at least what's worked for me. And then, you know, you can probably take pieces of it for me. I start with my personal brand and thought leadership. So giving away AI education as highest quality as possible and just giving it all away for free. Free, that builds up my personal brand. And then I sneak in like, hey, check out Blotato here and there. So I'll make two TikToks a week, right? Taco Belloto in my newsletter. At the bottom it'll be, it'll say, hey, I built Blotato for this particular purpose so that organic social media for my brand then funnels a small percentage of my audience who are a good fit to blow Tato. They trip funnel them to the website. And I think just the Blotato website just this year has gotten 1.4 million visitors already just through, yeah, just through that funnel. So like my personal brand is probably around 500 million views. Okay, so but that's like broad AI education content. And then it's just a matter of like, you know, inserting it here and there in subtle ways that aren't pushy and aren't for me. It's important like not to be too salesy because I don't resonate with that. And so I just like insert it in different ways here and there and that just funnels that initial massive audience at the top. It funnels of people or a good fit directly to the website.
- [12:02]
Steve Chou
I mean it works. I mean I fell for that funnel. Let's say you're selling. So do you recommend that people start with their personal brand first?
- [12:12]
Sabrina Romanov
I think it's very powerful. I know that a lot of people aren't willing to do that. I also don't think it's necessary for anything that's consumer facing. I see a lot of, for example like B2C consumer mobile apps on TikTok. Most of those I'm seeing succeeding do not have any face or personal brand at the helm. But because you want to find some repeatable viral format that you can then hire your UGC army of 100 creators to then replicate. So if it's only attached to you, then it's not necessarily as scalable as having a different type of video that any creator you hire can replicate. So for B2C I don't think it's necessary at all for B2B though, which is what Blotato is. So Blotato sells to agency owners, people building their brand, their businesses building their brand. For B2B, it's incredibly powerful. The inbound opportunities you get get by putting your face out there, your thought leadership out there are pretty incredible irrespective of whether it generates leads for your business. I think like both are incredible. So if you are willing to put yourself out there, I would definitely vote for that for anything B2B context.
- [13:21]
Steve Chou
So you're pretty well spoken the videos that we watch now. Were you always this eloquent? Did it take practice? Do you have some earlier videos that were just horror horrific?
- [13:32]
Sabrina Romanov
Oh yeah. Like I couldn't watch my earlier videos. Like I cringed. I was so critical of myself. I. I always, I also felt I was trying different formats. So in the beginning, like you're trying to find and create your own voice as well and that's a very awkward process. So I'd like try to copy someone else's style and I couldn't watch it again because I knew it wasn't me. Like I wasn't authentic. Maybe nobody else picks up on these things, but I could pick up on it articulate. I don't, I don't, I don't know. I just try to. I trying to say something as clearly as possible using simple language. It's in my mind. It's much easier to complicate stuff and sound smart than it is to try to actually simplify something, especially knowing that these are very nuanced topics. But to teach somebody, you need to like really distill it to the core essence of the material.
- [14:25]
Steve Chou
So can I ask how many pieces of content you're putting out a day?
- [14:30]
Sabrina Romanov
Yeah, yeah. So I actually have my entire social playbook published. So I only make content once a week and then it starts with a YouTube video in the morning, so that's my long form. And then I repurpose that into a newsletter and a couple short form pieces just kind of announcing hey, like here's my latest YouTube tutorial. And then I take a lunch break and then I film 20 TikTok videos and then when each TikTok video is posted I have an automation that uses Platato to repurpose it to other platforms.
- [14:59]
Steve Chou
I just wanted to take a moment to tell you about a free resource that I offer on my website that you may not be aware of. If you are interested in starting your own online store, I put together a comprehensive six day mini course on how to get started in ecommerce that you should all check out. It contains both video and text based tutorials that go over the entire process of finding products to sell all the way to getting your first sales online. Now this course is free and can be attained@mywifequitterjob.com free. Just sign up right there on the front page via email and I'll send you the course right away. Once again, that's mywifequitterjob.com free.
- [15:35]
Now back to the show. So is that 20 pieces a day is your.
- [15:40]
Sabrina Romanov
Oh no. So that's per week. So let's say oh per week. Okay, so let's say 21 original pieces of content per week and then that is just repurposed everywhere else.
- [15:49]
Steve Chou
Okay, so three pieces a day approximately. Would you recommend someone starting out have that frequency?
- [15:58]
Sabrina Romanov
Ideally yes, but I understand like most people don't have the time so it's more a function of like what can you fit in to the time you have? I do think consistency is the most important thing. So if you can only do one a day, like just go for it and just stick to it for as long as you can. It didn't really become a habit for me until about like 90 to 100 days where I truly felt like it was wrong to not post something. I just felt like it was part of my identity now. Like I just got to post something today, even if it's just a tweet, one line on LinkedIn and that will count, you know, in my head. But yeah, I noticed most beginners kind of just fall off track. They never actually make it to 100 posts published.
- [16:42]
Steve Chou
Yeah. And to come up with that many pieces of content, I imagine you use AI to augment that. Are, are you eating your own dog food and for ideas and whatnot.
- [16:54]
Sabrina Romanov
Yes. But I think one of the things people don't realize is I also source. I use AI to source ideas from stuff I'm already doing. So when I'm coding for blotato or answering support tickets for Blotato, for example, once a week my AI bot goes in there and tries to like look for interesting trends. And that can be a content idea. For example, like a lot of people are asking for particular feature, like create carousels with blotato and that's. That was the origin of that particular feature and that could be a content idea. I also look at things like meetings I'm already having. So I don't have many meetings, but my blotato weekly office hours has an AI meeting notes taker and that is also repurposed into content ideas. So I like. My suggestion for most people is to like not change what you're already doing, but just look for opportunities where you can kind of plug in AI to kind of listen, extract ideas organically from the stuff you're already doing. For most people, it's probably combination of meetings or maybe emails are kind of the two most obvious places to start and just have an AI flow that's like, hey, these are like 10 pretty interesting ideas from the stuff you discussed today. And the benefit is like you were authentic and engaged. Like you weren't trying to force these ideas. They came up naturally in conversation and that's what makes them like really valuable and interesting thing.
- [18:15]
Steve Chou
I can tell you how I'm using Blotato. So anytime I'm on TikTok or in my emails or seeing an interesting article, I immediately just shove it in a blow tato. And then when it times to create the content, I just put my own spin on whatever that article was. And that has made life just so much easier for me because I don't have to stare at this blank screen
- [18:34]
Sabrina Romanov
and exactly come with the post. Yeah, because like we get inspired by content ideas all the time. Like I'll see a LinkedIn post, I'll have a reaction to it and so I'll plug the post in blotato and then I'LL use my voice dictation app. So a lot of people use whisper flow. I use voicing. And then I'll be like, well, this is what I think and it's slightly different. And then blotato kind of mashes them together. Right. Do final edits and then publish it out. And that's like, exactly the core use case it was built for. I know a lot of people ask me for full automation, but in my opinion, most beginner creators should not start with full automation. They should, like, take interesting ideas that they already have or that they see online, add their unique voice, and then publish it, rather than trying to automate everything from ground zero.
- [19:19]
Steve Chou
I'm actually not a fan of full automation at all because I don't feel like what's generated reflects. It's never exact. I always have to edit. So I'm actually curious. I've noticed you've used an AI avatar. Do you continue to use the AI avatar? And does that flow work for you?
- [19:36]
Sabrina Romanov
Yeah, I actually only use it when I'm on vacation or out sick and so. And always wears a hat. So, like my audience knows. And it's transparent and it's honestly safe. Like, there was two weeks when I lost my voice and it honestly saved me during that two week period. But yeah, there's still like, massaging. Like, I really care about the script that it's saying. It's also difficult to get the avatar's like, enunciation correct and sounding correct. So I've sat there for four hours tweaking 11 labs parameters until I could find something that was 80% of the time acceptable. Right, 80% of the time. That was like the best I could do.
- [20:15]
Steve Chou
Okay. Yeah, yeah. You know, I started replacing my outros in the podcast with 11 labs just so I could play around with it. It's gotten a lot better actually, recently.
- [20:25]
Sabrina Romanov
Oh yeah? Yeah, the professional voice clone definitely has gotten better. Yeah.
- [20:29]
Steve Chou
But for you though, I was under the impression that everything was kind of automated. So you're actually going through and tweaking the output before it gets posted.
- [20:38]
Sabrina Romanov
Yeah. So I think what I've kind of learned. So when I initially started making automation tutorials for blotato users, I would have them fully automated because that was like the simplest version that I could do. Because once you have a human in the loop step, it's like, okay, you write it to airtable, wait for somebody to approve it, then send it back. And I. I didn't want to introduce that level of complexity in a tutorial. What I've Learned though on this topic is like, people really download these templates and just expect it to work out of the box and like go correct. And that I did not anticipate. And so kind of my learning from that moving forward is to have a human in the loop QA step in every single tutorial moving forward, because I truly believe that's like the approach 90% of people should probably take. I'm not a fan of using AI to distance yourself from the creative process. And that's what a lot of people want it to do. Like they don't deal with the edits and the thinking and finding your voice. But I believe the complete opposite. Like the people who are succeeding using AI are using it to go deeper into the creative process. Like, find my blind spots, give me five different perspectives that challenge me, you know, and that's the way to use AI successfully.
- [21:56]
Steve Chou
Okay, I'm so happy you said that. That being said, though, we kind of glossed over your flow. So from what I understand, you have AI just kind of browse all the different things that you're doing and does it put it in airtable or.
- [22:08]
Sabrina Romanov
Yeah, exactly. Yeah. So when it comes to ideation. Yeah. Bucket number one is what am I already doing today? So it's looking at meetings, it's looking at emails, and it's looking at my GitHub projects, which is where I manage like blotato tickets, like feature tickets and bug tickets. Number two is a news. Right? So I use an app called RSS app where you can pay attention to different Twitter accounts, TikTok accounts, Instagram accounts, and just like compile what is the most interesting news for my particular audience. That tends to be tools that are free or resources that are free. Like free free. Deep learning AI just released a free course on AI governance for agents. Like, that's. That's interesting. I really like that. I'm going to share that. And also new tools that come out, like the ChatGPT subreddit sometimes has funny stuff. So I'll include that in there. And then the third idea, the third source of ideas is actually scraping consumer content and just thinking about the different hooks. So just niches completely unrelated to AI. So let's say dating or relationships. But there I'm looking for just inspiration around like very interesting hooks that I could repurpose for an AI use case.
- [23:16]
Steve Chou
So these scrapers, are they code or are you using something like make or nadin?
- [23:21]
Sabrina Romanov
Yeah, I'm using N8N, but you can definitely code it today with coding tool. So I personally Use cloud code for coding. Platato.
- [23:31]
Steve Chou
Yeah. Okay, so N8N. So you have all these things now then I imagine that's like a whole bunch of stuff that you have to sift through. And then do you pick one and then feed that into AI to write your script? I imagine you have like a prompt that you use that sounds like you. Or what's your process there?
- [23:48]
Sabrina Romanov
Yeah, exactly. So it will put it in airtable. So I have an AI agent where I've given it examples of previous content that has gone viral. And so that agent scores, like, let's say it has 200 ideas. It'll like, score it based on what it thinks will do well. And then obviously I apply my own judgment as well. Sometimes there are just topics I want to talk about that I think are important to talk about. Like, a lot of people overlooked Brave's initial research on AI browsers. That was like, a while ago. I posted about it on TikTok a while ago, but it really didn't explode until OpenAI released Atlas last, you know, last week or two weeks ago. And people were like, wait, this is not safe. But, like, that's a topic that my AI system didn't find. And I just, I randomly found it just reading through social media. I was like, oh, I should talk about this. So there's all. I mean, there is, like, there's room for human judgment. Like, there's certain topics I feel are important. And then. Yeah. So basically I'll take whatever the research was for that topic and then plug it into Blotato. I have a TikTok default script that I like to use, and then it just spits it out. Sometimes if I'm like, get a little low on views the past week, I'll spend extra time on the hook. So I have another thing in blotato called a viral AI coach where you can up, you first record your video and then upload it. So I'll record like 10 seconds, the hook, including a visual hook, and then upload it to that viral AI coach to get an additional layer of feedback.
- [25:17]
Steve Chou
I didn't even know that feature existed.
- [25:19]
Sabrina Romanov
Yeah, a lot of people don't. But once you find it, it's really. It's really awesome, especially for beginners. So you can upload and you can upload an entire video, but it's really focused on analyzing the first 20 seconds. So it'll transcribe it, it'll analyze it visually for like, the visual appeal of the hooks, and then it'll give you a scorecard of like, how, how is the first hook? How was the second hook, et cetera.
- [25:43]
Steve Chou
Okay. And so you record it first and then you use that.
- [25:47]
Sabrina Romanov
That's. If I'm feeling like I need to work on the, like I want more views this week, I really need to work on the hooks.
- [25:55]
Steve Chou
Yeah, my ego is like that too. If I'm, if I'm low on views, my ego needs that fix. So. Okay, so for filming, do you just kind of ad lib based off notes or do you have a teleprompter? What do you use?
- [26:11]
Sabrina Romanov
So I only use a teleprompter if I'm literally reading a list, like one, otherwise I can't remember the list. I don't normally use a teleprompter. So literally what I'll do, I don't have it here, but I just have like a selfie stick that holds my phone up. And then my computer with the notes from Blotato will just be like right there. So it'll just be like my phone notes. I can just glance at it and wing it, honestly. And then that's it, that's the video.
- [26:36]
Steve Chou
Okay. And then your routine is, do you do three a day or do you just batch all 20 for like the week?
- [26:42]
Sabrina Romanov
I batch everything in one day of content creation.
- [26:45]
Steve Chou
Okay, okay. And then the rest of the time, I imagine, is working on Blotato.
- [26:50]
Sabrina Romanov
Yeah, people think I create. People think it's the opposite. Like I must spend all my time creating content. I specifically built Blotato so that I don't have to spend all my time creating content. Right. But yeah, most of my week is like fixing bugs, building new features and answering support tickets like that. That's pretty much it for the week.
- [27:10]
Steve Chou
And then assuming you follow this process of posting three times a week, or, sorry, three times a day, when did you start feeling, seeing traction?
- [27:20]
Sabrina Romanov
Yeah, that's a good question. So in the very beginning, I only posted one LinkedIn post a day. That was for a month. I actually had one LinkedIn post go viral and it was about like the way of building AI startups that, that, that had a lot of interesting discussion. But I didn't start posting TikTok until maybe like June or July. And then it took like three months to get out of 200 view jail. Like my next milestone was 1000 views regularly. And to do that from, I distinctly remember from 200 to 1000, what made all the difference was using hashtags, just using hashtag AI, hashtag chatgpt, GPT. And that was it. Because I apparently I Wasn't doing that before. I didn't realize that was a thing.
- [28:06]
Steve Chou
Interesting.
- [28:07]
Sabrina Romanov
Yeah. And then my very first video that did well, it was because I use creator search insights on TikTok where it was like, oh, if folks aren't familiar with it, basically TikTok tells you what people are searching for and you can even find content gaps. So like maybe people are searching for voice, AI apps and there's a content gap for it. So you can make a video for, for that. And that was my first video ever that did more than 1,000 views. Yeah, but this was three months in already. Right. And so if I had given up earlier, then like none of what's happened would have happened.
- [28:41]
Steve Chou
I mean, most people don't make it to three months. And that, that's why I'm asking that question, like, you got to stick through. In my opinion, you should stick with it for at least a year.
- [28:49]
Sabrina Romanov
But dude, I, in my mind I was like, this is my five year commitment. Like, right, five years. I just think of in terms of years. I don't even think in terms of weeks or months, you know?
- [29:00]
Steve Chou
Yeah, no, I'm the same way. Three to five years is my thing, so. And that and blotato is like a long term thing for you too, I would imagine. Right? So let's switch gears a little bit and talk about Vibe coding, because I've noticed you talked a lot about it. I'm an EE and I used to design microprocessors for living. You're a CS major. I hear you talk a lot about it. So do you think the average Joe can do it effectively?
- [29:24]
Sabrina Romanov
Oh yeah. I mean, my 22 year old cousin just quit his full time job, moved in to our house. We call it a hacker house now. And he's on it. He just vibe coded his second app last night, which is. I don't, I don't know if he's listening, but you basically take like a picture of your Instagram post and you get like a virality prediction score for it. Because he was marketing his first app, which was like a study tool for college students. He was marketing it on like TikTok and Instagram. He's like, oh, it'd be really cool if I also had this tool and it's fully functional, you can sign up for it. I forgot the name of his new one, but yeah, he's 22 years old with absolutely zero background in technology, so I think vibe coding can work really well for simpler apps. But what I always tell people is like the goal of Vibe coding is to hire a full time engineer. That's the goal. The goal is to like validate your mvp. Get paying users, get it to the point where like people are like, why is this keep breaking? But that means they want to use it, you know, and then you have that validation and confidence and conviction to go hire like a real tech person to build the rest out. Maintain the code base. Probably refactor 95% of the code base. That's what I did. Blotato. I Vibe coded the first version with cursor AI. I filmed myself doing it. So it's on YouTube. Oh yeah, yeah. Well, the MVP, like the, the version that got my initial set of 100 paying customers. It didn't even have publishing, by the way. Like that's how MVP it was. Like it was a social media tool. You couldn't actually publish it right to social media. But like the very core feature was just that remixing, that repurposing. Right. Like put in an article, get posts out in your voice or in a viral voice. But yeah, you couldn't even publish to social media at that point.
- [31:13]
Steve Chou
Well, I know I am loving, I call it Vibe coding, I guess, but because I feel like I'm not really a developer and it sounds like you don't portray yourself as a developer either, I don't think. But like we know way, way more to be dangerous. But right now I feel like invincible. Like I feel like I can create anything. And I literally been going down like the Shopify app store and just running a lot of these apps literally in a weekend because they're so simple.
- [31:38]
Sabrina Romanov
Right.
- [31:39]
Steve Chou
But then again, I also run a class and some of these people, it's. It's hard for them. Right. Because it spits out a bunch of stuff. It doesn't work. And then.
- [31:48]
Sabrina Romanov
Yeah, you have an engineering mindset. Yeah, yeah.
- [31:51]
Steve Chou
And you can't really look at it. So what I wanted to get out of you today is let's say you are someone who's not tech savvy. Like, was your cousin or your.
- [31:59]
Sabrina Romanov
Yeah, cousin.
- [32:00]
Steve Chou
Yeah, your younger cousin. Right. So where should they start?
- [32:04]
Sabrina Romanov
Yeah, I mean there are a lot of good Vibe coding tutorials online. So for example, this past newsletter this weekend, I walked through how to use lovable cloud to Vibe code a lead magnet. And I think a lead magnet is actually a pretty good starting point because you don't have like a complex backend. You don't need authentication, you don't need payment provider on top. Right. But this was still an interactive little app, meaning like, you answered a bunch of questions, it spit out an ROI calculator and a personalized report. And you can even put in your email address so you can capture emails. In other words, it's, it's a true lead magnet, where people can put in their emails and stuff. And like, things like that are a really good starting point. Again, not a very complex backend. It was really just storing your emails and responses that people put into your lead magnet. No authentication, no payment system either. Another example I Vibe coded was my email.
- [32:56]
Steve Chou
Does that email go straight to a specific provider or is it just putting on like a sheet or something?
- [33:00]
Sabrina Romanov
So, Lovable Cloud, as long as you connect to your Supabase account, it will spin up the database, create the database schema, and then it will just insert the emails into Supabase. And then you can, you can just click one button, you can grab everybody's emails right there.
- [33:14]
Steve Chou
And so for everyone listening, Supabase is just like a database. And from what I'm understanding, there's no hooks into popular email providers. It just puts in a database that you can extract and add it to your provider, right?
- [33:25]
Sabrina Romanov
Yeah, exactly. And you could hook it up with something like Resend. So Lovable has a pretty streamlined integration with Resend if you actually want to send out emails from the app itself. But again, I would probably consider that like step two or mvp. The other example I was going to say I just posted a video on, this is my AI agents and automations directory. It had 37,000 visitors in the past three months. It was vibe coded in just a few hours. And obviously I'll put like a couple links there to potato. So it helps drive.
- [33:54]
Steve Chou
Yeah, I'm sorry, what did the app do?
- [33:56]
Sabrina Romanov
So it's called, it's, it's just a directory of like Automations. So let's say you're looking for marketing automations in N8N or make dot com. You can filter like marketing N8N is your platform and then it'll just show you a bunch. You click on one, you can see like who made it, you can contact them for help and you can download the template. It's literally like a single page directory and the data is just in Supabase. But that's another example of like, you know, start there. Start with like something like how come
- [34:24]
Steve Chou
no matter what I type, it always points me to blow? Tato, That's cool. Yeah, I can totally see something simple like that. Absolutely. I didn't even know Lovable had a database attached.
- [34:38]
Sabrina Romanov
Oh yeah, it's pretty New.
- [34:39]
Steve Chou
Okay.
- [34:40]
Sabrina Romanov
Yeah. It was basically one of the biggest complaints. But it's called Lovable Cloud and basically it'll manage your Supabase for you. Set everything up because it was really annoying the way you had to do it before.
- [34:51]
Steve Chou
Yeah, yeah. I thought it was only a front end before, so I didn't know. Is that. So is Lovable what you would recommend to someone just completely green to this?
- [34:59]
Sabrina Romanov
I think so. You know, the other ones are strong as well. So I've been really impressed with Emergent Sh lately.
- [35:06]
Steve Chou
I haven't used that one.
- [35:07]
Sabrina Romanov
Yeah, everybody I've shown it to. So it's similar to Lovable but it can also build mobile apps and instead of Supabase, I think, I believe it's using Mongo to the object storage. But it's really neat. Everyone I've shown it to has kind of been blown away by the level of polish that you get. With that said, though, my husband and I, on our last livestream, we actually tried building a mobile app live and submitting it to the app Store and we almost got there. However, when we actually downloaded the code base and tried to build it locally before submitting it, there were so many issues with the emergent code base. Out of date libraries, things were missing. And so I would say it's not completely streamlined for mobile development yet, but it's a brand new player and it's incredibly promising. So I would probably recommend lovable.devemgent sh or I know a lot of people who also still love Bolt New and that can also build mobile apps.
- [36:04]
Steve Chou
So how do you distinguish? Because it's kind of dizzying. There's like a new platform every day. So I've been trying to train my son to use replit recently. Or do you use. Would you use blotato? Use cursor for blotato.
- [36:19]
Sabrina Romanov
For blotato, yeah. I mean, a lot of these did not exist at the time. Like Lovable I had not even heard of. I don't think it had launched publicly at the time. I was filming me vibe coding, so I was using cursor AI.
- [36:31]
Steve Chou
Okay. And then, I mean, cursor by itself is just. I mean, you were just generating code, right. And you had your own platform for. Okay. Because you're a CS major.
- [36:41]
Sabrina Romanov
Yeah.
- [36:42]
Steve Chou
Did your husband tech also?
- [36:44]
Sabrina Romanov
Oh, yeah. So electrical engineering? Computer science. Yeah. Yeah. We actually met in our data structures class in undergrad and we had like one of the heart, the hardest professor, everybody at least from Berkeley knows him. And the beginning of the class was like 150 people. By the end it was like 20.
- [37:02]
Steve Chou
So I can't remember if that was a weed out class. I don't remember. But yeah, for CS majors. All right, so if you're new, it sounds like lovable. You can make simple apps. Lead Magnets is a good use case. I think for you, it sounds like you did you end up hiring developers then and you said you had to refactor all your code. Are you maintaining it?
- [37:24]
Sabrina Romanov
No, I'm maintaining it. Yeah. And so I don't use cursor AI anymore. So I definitely switch to cloud code, which I highly, highly recommend if you have a complex existing code base. But yeah, it's. It's a lot of work. I will say I think I forgot how much work a startup is. So I'm building my brand solo. I'm also building potato solo. For me, it's more of like an intellectual challenge. How far can I go solo? I'm not at all saying I believe that's the optimal approach. If my goal was only to optimize revenue. I can definitely see, like it would be really nice to have another developer. For example, if I'm doing marketing, then suddenly like product development just stops because, like, my focus is on marketing. Like, I'm traveling to an event next week and like I'm not going to be able to publish new features. I'll be able to barely maintain the code base, maybe fix a few bugs. But yeah, it's been a lot doing it all solo.
- [38:22]
Steve Chou
Isn't that your husband's job?
- [38:24]
Sabrina Romanov
No, no. He has a totally different set of ambitions, so he wants. No I gaming company.
- [38:31]
Steve Chou
What you're doing is actually my dream. What I fear is actually the support. And I know I've asked you a couple questions on Blotato and I. I never wanted to press anything just because I know like, you're busy doing other things. But that has been my greatest fear because you always underestimate someone's technical ability or overestimate, I should say, for any tools. What has been the biggest challenge for you with Blotato?
- [38:55]
Sabrina Romanov
Actually, I think it's exactly that because, like, I'm in it every day and I don't see what other people see. Like when I. When the remix screen to me made sense, I showed it to some other creator friends. They're like, yeah, I get it. But the typical person who doesn't have a strong background in content creation, they're just like, I don't know what to do at this step. And so my biggest challenge has just been like, like Almost being less technical or like simplifying it or streamlining the flows for beginners to kind of get up and running. Even to this day, I don't necessarily recommend blotato to absolute beginners. Like, it's, it's really helpful if you have been trying to create content on your own for a little bit so you can kind of understand like, oh, repurposing it. Oh, here's I just add my own voice. Okay, here's my content calendar. I can schedule it out, but yeah, that's, that's been a big challenge. Support is also definitely a big challenge to scale. It's tough. Like I get 100 to 200 support tickets a day and yes, I have an AI support bot that does it and does a pretty good job. But you still have to like update the knowledge base all the time. Like I just pushed a new feature where you can now schedule through the API, schedule a post to your next free slot. I have to update the documentation, update the AI bot so it refreshes its knowledge base. Look for anything else in the documentation that might have outdated knowledge. And I use cloud code to help me with that, like identify things that are outdated. But it's a lot of work. Like it really, really does add up. So.
- [40:27]
Steve Chou
So you don't have a dedicated support person. Is it a one man show right now?
- [40:31]
Sabrina Romanov
Literally, yeah. People are really surprised when I just hop in the support chat and I'm like, okay, amazing. Okay, so that is one way to reach me, I guess, if you really want me to reply.
- [40:42]
Steve Chou
Yeah. So it sounds like support is really the pretty much the only manual thing that you have to take care of outside of the coding, which is for me it'd be the fun part.
- [40:53]
Sabrina Romanov
Well, yeah, it's mostly fun I think, when I get to work on new features. But it's honestly a lot of plumbing. Like the social APIs change or the documentation's horrendous. Like if anyone here has tried to actually integrate with the social APIs, it's pretty like tedious. I'll say is a nice way of saying it. But yeah, like that stuff is not that fun for me, honestly. But I have to do it and it's probably like 60% of the work I do is, is what I consider plumbing, like not really that fun at all from a developer perspective.
- [41:25]
Steve Chou
I'm using this tool called Repurpose. Actually I was considering just writing my own version of Blowtato, but then you're right, I walk through like all the, the documentations and then there's these tokens that expire and then all of a sudden you have to manage. I'm like, screw it, I'm just gonna. Okay, so to code something like Blotato sounds like it's much more involved for, for someone, but.
- [41:51]
Sabrina Romanov
Well, for like I will say I think you could build the same MVP I built in probably a few hours now with how vibe coding tools have evolved. But I always like to caveat people that when it comes to building product, launching it and getting users is the start of your product development journey. So like to build Blotato what it is today would take significantly more work than just spinning it up and emergent. But you could definitely build the initial first version v2, v3 in lovable or emergent today for sure.
- [42:22]
Steve Chou
Well, it's all fun and games when you're the only one using them. Right? As soon as it gets the masses off, hell breaks loose. Okay, there's one more thing I wanted to talk to you about which I'm drawing a blank now. But Sabrina, where in terms of your marketing, what has been your most effective way to just get users actually on the blotato?
- [42:46]
Sabrina Romanov
I would say tutorials on YouTube. Okay. Because it's long form or long. Yeah, long form. Because I think it's attracting people who are very much already looking for a solution and already pretty educated on the problem. Especially for the API. I would say probably the. Yeah, YouTube long form tutorials with like automation templates has been the primary driver of API users. For the non API users, Instagram has actually been pretty good as well. So posting videos as well as carousels are just different features that I have. I've had a few viral videos on TikTok specifically for, for the faceless video creation in Blotato. So that's been really helpful as well. But I can tell it like I had another marketing experiment where just did a faceless channel. It got hundreds of millions of views, had a blotato watermark, but a very, very low conversions. So like that was, you know, I was really excited about the views and also just proving that Blotato can work for this use case around like face story videos. But terrible, terrible conversions. Like just because you put a watermark there does not mean it's targeted to your audience. They don't know what that watermark references like why should they go there?
- [43:56]
Steve Chou
So yeah, so, so it seems like you're using short form for awareness and then the long form YouTube videos to actually convert them into customers.
- [44:05]
Sabrina Romanov
Yeah, exactly. And that's, that's how my entire Kind of brand funnel is laid out like I kind of assume short form content. It's just not as authority building as long form content. So I want all of my social short forms to link Back to my YouTuber newsletter.
- [44:21]
Steve Chou
Okay. And then your newsletter, is it once they sign up or. No, that's right. You have an AI newsletter.
- [44:28]
Sabrina Romanov
Yeah, yeah, that's separate.
- [44:29]
Steve Chou
Okay.
- [44:30]
Sabrina Romanov
So yeah, my goal for social is like everybody on my email list, it's by far highest roi. And I had always heard that as a creator, but I did. I don't think I like believed it until I started saw it myself. But it's really nice having control of your own distribution list and not being, you know, at whims of the algorithm every other week.
- [44:50]
Steve Chou
Yeah. And I know you have a Discord group also. Right. Is that one of your primary legions also? Like, no. No, it's not.
- [45:01]
Sabrina Romanov
Yeah. So the discord. So I just created a, like a Discord community for women building an AI just because honestly, it can get lonely. And I just wanted to. To meet more other women builders in AI. But it's a pretty small group, like a couple hundred women. And we're testing it out in school as well. We're trying to figure out like should we stay in Discord or should we move over to school?
- [45:21]
Steve Chou
I love Discord because of the coding aspects. I said you too, right? It's like infinitely powerful.
- [45:27]
Sabrina Romanov
Yeah.
- [45:28]
Steve Chou
Yeah.
- [45:29]
Sabrina Romanov
I grew up playing video. I have three brothers, so I grew up playing video games and I've used Discord for as long as I've known.
- [45:35]
Steve Chou
So I can tell just by your chairs, but Sabrina, hey, it's been great talking to you. If people want to sign up for your newsletter or check out your tool, where can they go? What's the best place?
- [45:49]
Sabrina Romanov
Yeah, so for my newsletter go to sabrina.dev and that's honestly. Yeah, best place.
- [45:55]
Steve Chou
Okay, awesome. Well, thank you so much for your time. I really appreciate it.
- [45:58]
Sabrina Romanov
Yeah, thanks so much for having me.
- [46:00]
Steve Chou
Hope you enjoyed this episode. If you're not using Claude or Blotato at this point, make sure you check out these tools because they will change your life. For more information and resources, go over to mywifequitterjob.com episode631 and once again, tickets to Seller Summit 2026 are now on sale over@sellersummit.com and we are almost sold
- [46:19]
out at this point.
- [46:19]
If you want to hang out in person in a small intimate setting, develop real relationships with like minded entrepreneurs and learn a ton, then come to my event, go to sellersummit.com.