In this presentation, we will explore the transformative potential of large language models (LLMs) in reshaping the landscape of open-source software. As LLMs continue to advance, the traditional paradigm of software development is undergoing a significant shift; many software features are now being replaced by conversational interactions with LLMs. Central to this evolution is the realization that software essentially combines personal data with unique workflows—elements that can be efficiently “coded” through natural language prompts. Looking ahead, the future of open-source may resemble a collection of “dinner recipes” rather than rigid scripts, effectively lowering technical barriers and democratizing access to technology.