![Never wrote a line of code, now a $6.6B unicorn : the vibe coder - Lazar Jovanovic [Lovable]](/_next/image?url=https%3A%2F%2Fimg.youtube.com%2Fvi%2FWqZzC8Wragw%2Fhqdefault.jpg&w=1080&q=75)
Never wrote a line of code, now a $6.6B unicorn : the vibe coder - Lazar Jovanovic [Lovable]
AI Summary
Lazar Jovanovic, the world's first official professional vibe coding engineer at Lovable, a startup that achieved $100 million in ARR in just eight months, shared his insights on the future of software development, AI, and the concept of "vibe coding." Despite never having written a single line of code manually, Lazar ships production apps daily, often faster than senior engineers. His approach leverages AI, clarity, and "taste," which he defines as a crucial element in design and product development.
Lazar believes that non-technical people have a unique advantage in this new era of AI-powered development, citing Kurt Cobain's naiveté in music theory as an analogy. Those unfamiliar with traditional rules are more likely to innovate and create new ones. He argues that even the creators of AI models don't fully understand their capabilities, pointing to examples like Claude code being used to vibe code Claude co-work in just seven days, and OpenAI's latest model being largely written by its predecessor. This suggests that if models can write models, individuals can also build anything they envision.
He clarifies that being an engineer is not a disadvantage, especially for experienced 10x engineers who embrace these new tools. However, for non-engineers, the advantage lies in quickly immersing themselves and leapfrogging traditional development cycles. While some knowledge of software architecture is necessary, the key is to experiment and iterate rapidly by prompting AI tools without getting bogged down in technical nuances.
Regarding software maintenance, a common concern with AI-generated code, Lazar acknowledges it's a current bottleneck but believes it will diminish as models improve. He proposes a framework where AI tools are treated as technical co-founders, requiring thorough documentation both before and after building features. This involves creating Product Requirement Documents (PRDs) before development and backing up knowledge about how the feature was built and its dependencies afterwards. This systematic documentation helps in maintaining software and directing AI agents to fix issues.
Lazar emphasizes that the underlying AI models are constantly improving, with expanding context memory windows. He highlights Anthropic's Sonnet 4.6 as an example of a massive model with an amazing context window and reasoning capabilities for coding. He believes that the bigger problems for "vibe coders" are not technical but rather conceptual: determining *why* something is being built and *how* to distribute it. He echoes the sentiment that if coding were the sole problem, every developer would be a millionaire, underscoring that attention and getting a product in front of users remain the biggest challenges. He predicts that within 6 to 12 months, many of the current technical problems will no longer be issues.
Lovable's strategy is to empower anyone to build anything, from games for children to businesses and enterprise-grade internal tools. Lazar highlights the company's mission to serve the 99% who don't know how to code, while also acknowledging that engineers can use Lovable for tasks like building user interfaces, data visualization, or feature adoption tools, even if the tech stack (React framework) might not align with their company's primary stack.
Lazar draws a parallel between what bootstrapping did for venture capital and what vibe coding will do for traditional software development. He believes that just as bootstrapping disrupted the norm and became mainstream, vibe coding will normalize rapid, AI-assisted development. He expects this shift to happen much faster than the bootstrapping movement, possibly within two years, due to the rapid advancement of AI models and increased attention spans.
He encourages listeners to embrace this shift, emphasizing that the fast pace means anyone can jump in and quickly become proficient, as current knowledge becomes irrelevant within six months. He highlights his "28 days of Lovable" series as an example of how quickly one can learn to build.
Lazar views AI as a massive opportunity unlock, not a "doom and gloom" scenario for jobs. He believes that in a world increasingly filled with "fake" content, human-to-human interaction, trust, and taste will become even more valuable currencies. He notes that while some predict AI will take jobs, people are hired for more than just efficiency; human nuances like reading a room or understanding context from past interactions remain invaluable.
He points out that unlike the sudden shift during COVID-19, the adoption of AI is a choice, and many people are still on the sidelines. He advises anyone watching to take action, even small steps, to start using AI in their daily lives, as this will prepare them for what's coming.
Addressing the "SaaS-pocalypse" and the idea that interfaces might become obsolete, Lazar believes that extreme predictions are unlikely to be entirely true. He envisions a synergy between interfaces and chatbots, where software becomes more personalized and adaptable. He describes building internal tools at Lovable with an interface for initial interaction and a chatbot layer for automation, demonstrating a "not one or the other, but both" approach.
Lovable's internal obsession is the value generated for end-users, not just ARR. They track metrics like daily active apps and user monetization. Lazar highlights initiatives like "She Builds" to empower women founders and features success stories of users building businesses, even non-digital ones like a baker making $150,000 a year using Lovable for online distribution. He stresses that the best entrepreneurs often come from non-technical backgrounds, citing a landscaper and a radiology technician as examples of individuals with great ideas for Lovable.
Regarding potential churn from B2C users and the failure rate of businesses, Lazar confirms Lovable is exploring partnerships and scaling its go-to-market strategy, particularly for enterprise clients. He notes that internal champions often drive adoption within organizations, starting with personal use and then advocating for Lovable at a corporate level. Lovable actively seeks partners to aid distribution and focuses on driving internal adoption within companies, not just securing contracts.
Lazar explains the "Aladdin and the Genie" analogy to illustrate the challenges of building with AI. The "genie" (AI) has a finite "wishes" (tokens) limit, which is the machine problem. The "human problem" is the user's lack of specificity in making wishes. He advises users to be precise, ask for only a few related changes at a time, and provide clear documentation to guide the AI agent, reducing its "research" effort and maximizing "execution."
He prefers using Lovable for prompting because it retains full context of the project, codebase, and past messages, unlike generic LLMs like ChatGPT. While he uses other LLMs for educational purposes (e.g., understanding technical concepts), he leverages Lovable's planning mode and project knowledge feature for actual development. He encourages users to ask the AI *how* they should have prompted it to avoid bugs, acknowledging that errors are often the human's fault. He also uses AI design tools to improve his "taste" and then imports those concepts into Lovable for building.
Looking ahead, Lazar is excited about the future of agentic workflows, where AI agents can perform complex tasks autonomously, though he acknowledges security concerns. He believes in eliminating oneself from the equation by building systems where tools prompt themselves. He is eager to see self-intelligent and self-correcting AI implemented in daily life.
Finally, Lazar offers advice on how to get hired as a "vibe coder": don't wait for a job description. Instead, identify problems within your company, build solutions using AI tools, demonstrate the time and money saved, and then present these results to management. This proactive approach allows individuals to define their roles and responsibilities, creating a unique position with zero competition. He encourages everyone to "just build" and not wait for permission.