Keep pulling the thread on Lazar Yovanovich.
The traditionally separate roles of engineer, designer, and product manager are converging into a single role due to AI's capabilities.
Lazar Yovanovich predicts that manual coding will become a rare art form like calligraphy, as AI will write the vast majority of code.
To manage an AI agent's limited context window, Lazar Yovanovich creates a series of markdown documents (master plan, implementation plan, design guidelines, user journey) that serve as a persistent source of truth for the project.
Lazar Yovanovich advises aspiring professionals to focus on developing skills like emotional intelligence, good design taste, and high-quality copywriting, as these are less likely to be automated by AI.
As a professional "Vibe Coder" at Lovable, Lazar Yovanovich is responsible for building both internal and external-facing products across all departments, including marketing and sales.
Lazar Yovanovich believes not having a technical background is an advantage when using AI coding tools because it removes preconceived notions of what is impossible to build.
Lazar Yovanovich's development workflow consists of spending 80% of his time on planning and chatting with an AI agent and only 20% on executing the plan.
To achieve clarity when starting a new project, Lazar Yovanovich runs multiple parallel builds using different initial inputs: a voice brain-dump, a structured text prompt, screenshots from Mobbin or Dribbble, and existing code snippets.
Lazar Yovanovich uses a four-step debugging framework: 1) use the tool's auto-fix, 2) add console logs for more context, 3) use an external tool like OpenAI's Codex for analysis, and 4) revert to a previous version and re-prompt.
After resolving a bug, Lazar Yovanovich asks the AI agent how he could have prompted it better to avoid the issue, and then incorporates the feedback into a `rules.md` file to permanently improve the agent's behavior for that project.
Lazar Yovanovich believes elite engineers will be needed more than ever to build and maintain the infrastructure required to support billions of new builders enabled by AI.
At least half of S&P 500 companies have employees using Lovable to some extent for tasks like prototyping.