Aug 29, 2025 · 31m · cheeky-pint
A conversation with Lovable CEO Anton Osika
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Anton Osika, CEO of Lovable, discusses how his AI-native software creation platform originated from GPT Engineer, scales through organic adoption and enterprise expansion, and empowers creators to build full-stack web applications without writing code.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. John holds 24% of the talking time here. How this is scored →
John as informed peer 4.0 Guest teaching 4.1 Guest disagreement 1.1 John pushing back 1.0
speaking balance: gold is John, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 25:10 Open source is overrated in the AI era
Anton rejects the conventional praise for open source by bluntly calling it overrated since models can synthesize code on demand.
Hardest push from John ▶ 23:30 Is Europe just too comfortable?John challenges Anton's critique of European startups by asking pointedly if Europe is simply too nice and comfortable to produce extreme ambition.
Biggest teaching moment ▶ 22:30 Ambition gap versus talent gapAnton educates John on the real difference between US and European tech ecosystems, arguing the primary limitation is ambition rather than raw talent.
John holds their own ▶ 19:05 Stripe's developer velocity dataJohn brings concrete operational data from Stripe showing a 30% increase in code output to ground the discussion on AI productivity.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | John as informed peer | Guest teaching | Guest disagreement | John pushing back | Why |
|---|---|---|---|---|---|---|
| The Origin of Lovable and GPT Engineer | 3 | 4 | 1 | 1 | John opens by warmly setting context on AI capabilities and asking about the initial spark for Lovable. Anton explains how early 2023 LLM reasoning led him to build GPT Engineer over a weekend, which went viral and proved demand for visual software building. | |
| Identifying Target Users and Democratizing Software Creation | 4 | 4 | 1 | 1 | John draws parallels to Retool when asking how Lovable defines its broad target audience. Anton clarifies that while they initially considered internal tools, adoption surged organically across diverse users ranging from solo founders to large enterprise hackathons. | |
| Growth Metrics and Measuring True User Success | 4 | 4 | 1 | 1 | John asks about SaaS growth benchmarks and inflection points for modern AI companies. Anton explains that their North Star focuses on end-user retention and creation rather than top-line revenue alone, noting consistent organic word-of-mouth growth. | |
| Scaling Upmarket and Building the Team Structure | 4 | 3 | 1 | 1 | John inquires about Lovable's upmarket shift and sales infrastructure. Anton details how their team tier gained traction organically with thousands of paid users while maintaining a lean 30-person headcount without a large sales team. | |
| Product Engineering Culture and Built-in Distribution Tools | 5 | 4 | 1 | 1 | John connects product engineering culture to distribution problems, noting how platforms can assist user growth. Anton outlines their hiring strategy focused on former founders and built-in growth mechanisms like launch features. | |
| Internal AI Adoption and Software Development Productivity | 5 | 4 | 1 | 1 | John shares Stripe internal metrics on shipping 30% more code to contextualize AI engineering impact. Anton shares realistic insights on navigating complex codebases versus building simple SaaS applications entirely through prompting. | |
| Model Strategy vs User Interface and Agentic Systems | 5 | 5 | 2 | 2 | John synthesizes Anton's strategy regarding UI versus foundation models, then pivots to discuss the hurdles of building European startups. Anton pinpoints ambition levels and seasoned talent density as Europe's core bottleneck rather than baseline engineering quality. | |
| Future Predictions for Autonomous AI Agents | 4 | 4 | 2 | 1 | Anton outlines next-generation autonomous agents capable of self-healing and tool integration. In rapid-fire questions, he offers a contrarian take that traditional open source software is overrated in the AI era. | |
| Live Demonstration: Building a Full-Stack Web Application | 2 | 5 | 0 | 0 | Anton conducts an end-to-end live demo creating a Stripe clone with authentication and Supabase backend integration in real time while John observes the workflow. |
Statements from this episode (20)
Collison: Stripe PMs built viable Lovable prototypes in three hours without engineers
“I think I mentioned to you that a few weeks ago, my product management team held a hackathon and within just three hours, they had put together real viable prototypes on products that I would want to fund tomorrow. So with no eng, no design.”
Osika: Software development will transition to visual, conversational AI interfaces
“We're going to see a completely new type of experience and interface to build software products, and it's not going to be humans looking at the code, it's going to be something new, something more like visual, and driven by talking to an AI, and it's going to …”
Osika: Lovable users range from teenagers to funded serial founders
“Today it's used a lot for, as you said, like creating this first version of your startup by, both by teenagers or kids that are like, I want to make money and create something. And serial entrepreneurs that raised like fifty million dollars, but now they were …”
Collison: Traditional SaaS growth benchmarks do not apply to many AI companies
“Obviously that growth curve does not apply to many AI companies.”
Osika: Lovable has reached 120,000 paying users
“Now we have a 120,000 are paying and it's a simple, simplest way to measure that.”
Osika: Multiple Lovable-built apps have reached hundreds of thousands of users
“We have many apps that are like hundreds of thousands of users on their applications as well, or like visitors.”
Osika: Almost all of Lovable's growth is driven by word of mouth
“Almost all of the growth is from word of mouth. So it's quite, it's like, it looks almost scary, like linear or like very predictable growth.”
Osika: Lovable's new team plan acquired 5,000 paying users in one month
“Yeah, so we launched the team plan a month ago, and we've already have more than four, I think, 5000 people who paid more to get on the team's plan where you collaborate, and that's done by a company, right?”
Osika: Lovable employs roughly 30 people, including 17 engineers
“So we're about 30 people now, and we've had to just increase our support side quickly, because they're still with this volume, there's always support the questions. We're 17 engineers, people working in operations, and now increasingly some people who like did…”
Osika: Lovable is embedding distribution and user acquisition tools into its platform
“We don't, we want to help founders and we want to help them build their product, but we also want to help them get distribution on their products, right? So we're kind of building in growth over time into like the offering of building on Lovable. One part of t…”
Osika: Lovable will eventually be entirely built and edited using Lovable itself
“We still have not done this bootstrap where everything on lovable is edited with itself. And we are going to get there.”
Osika: Non-coders are making hundreds of thousands from prompt-built production applications
“That's what we're seeing with people making hundreds of thousands of dollars having built production-ready applications, but just by prompting, not looking at the code ever.”
Collison: Stripe is shipping 30 percent more code thanks to AI tooling
“I think at Stripe we, ah, we're, we found we're shipping 30% more code.”
Osika: Agentic orchestration and UX are the hardest parts of AI products
“The hardest part to get right is this inter interaction between like the AI, the complex agentic systems of like how the models are used and the user experience.”
Osika: The US has far more experienced SaaS talent than Europe
“There is much more people who've built SaaS and like startups before that you can, that have like very, very good expertise that you can hire in the U.S. And much less so in Europe.”
Osika: Low ambition levels are the biggest problem for European tech startups
“The ambition level is the biggest problem, I think. Like, there's fewer people that are, like, super ambitious, like, okay, this is a unique opportunity in the history of mankind, and we're here to build the best way to build software applications with AI. Ins…”
Osika: AI agent failure rates will drop tenfold within a year
“I think the, like, agents, they often get confused and trip over themselves, and we're in a year going to see that happening, like, by an order of magnitude less often.”
Osika: Open source is overrated because AI models can generate code on-demand
“Unfortunately with AI, it's overrated, like the, an AI model can just spit out open source code, so to say.”