Getting paid to vibe code: Inside the new AI-era job | Lazar Jovanovic (Professional Vibe Coder)
Summary
The episode profiles Lazar Jovanovic, a "professional vibe coder" who builds production-quality features and internal tools primarily by orchestrating AI agents rather than writing traditional code. Key themes include the value of non-technical perspectives, the centrality of planning and "chat mode" over raw prompting, and practical workflows for prototyping, debugging, and keeping agents aligned via persistent markdown artifacts (PRDs, masterplan.md, tasks.md). Lazar emphasizes working in parallel—spinning up multiple prototypes (brain dump, typed prompt, design references, code snippets)—to quickly converge on direction and conserve model context. The conversation closes by arguing that as AI commoditizes baseline output, human skills like taste, judgment, and design will become the primary competitive differentiators while engineering roles evolve toward maintenance, scale, and training.
Key Takeaways
- 1Vibe coding — building with AI without a coding background — is a viable full-time approach.
- 2Spend the majority of effort on planning and 'chat mode' rather than on execution prompting.
- 3Mitigate LLM context limits with a persistent markdown source-of-truth and modular agent files.
- 4Run several parallel prototypes to clarify thinking and pick a winner fast.
- 5Design judgment, taste, and human decision-making are the new differentiators.
- 6Have a structured debugging workflow when AI-generated features break.
Notable Quotes
"AI will just produce garbage faster if you don't know what you're doing."
"I spent 80% of my time in planning and chatting and only 20% in executing the plan."
"I'm optimizing 100% of my time today on good judgment, clarity, quality, taste."
"The gap is now this because everybody produces good enough with AI... so now learning and optimizing for how do I produce world class and magic is the key lesson to take away today."
"If you don't know what you're doing, you're just going to produce garbage faster."
"We will need elite engineering more than ever... who's doing the maintenance? Training code bases, scaling code bases, maintaining projects."
"When the problem gets fixed, I go into the chat mode and I ask Lolo: 'How can you help me learn how to prompt you better so that next time we do it in one go?'"
"The ceiling on the AI isn't the model intelligence. It's what the model sees before it acts."
Episode questions
What is vibe coding and what does a professional vibe coder do day to day?
Vibe coding is using AI tools (e.g., Lovable, Cursor, Cloud Code) to build internal tools and public product features without heavy manual coding. Day‑to‑day work includes prototyping multiple parallel builds, creating PRDs and markdown task files, feeding agents references, and reading agent outputs to iterate.
How does Lazar handle LLM context window limitations?
He mitigates context limits by creating and providing persistent documentation (masterplan.md, implementation plan, tasks.md, design_guidelines.md) and rules/agent files so the agent can read a compact source of truth and execute one task at a time. He also runs multiple parallel prototypes to pick a winner before committing token budget to deeper iterations.
What practical workflow increases clarity when building with generative tools?
Start with a brain dump (voice/whisper), then a typed version, then bring in design references (Mobbin/Dribbble), and finally attach code snippets/templates — run these in parallel to converge. After choosing a direction, spend a day creating PRDs and tasks.md to drive agent work.
Which human skills will become more valuable as AI takes over coding tasks?
Judgment, taste, clarity in requirements, UX/design sensibility, copy and typography — the ability to decide what to build and how it should feel will be rewarded more than raw output speed.