Agentic Coding and the Economics of Open Source

In this episode of the Practical AI Podcast, host Chris Benson welcomes Dr. Miklos Shkoren, an economist from Central European University, to discuss his provocative paper on the economic implications of 'vibe coding'—a term describing the rapid, AI-driven creation of software with minimal human intervention. Shkoren argues that while AI dramatically lowers the cost of writing and sharing code, it simultaneously undermines the core incentive structure of open source: human attention. As AI agents increasingly select and install dependencies without human interaction, the visibility and engagement that developers rely on for recognition and career advancement decline. This creates a systemic imbalance where open source projects face reduced user attention despite higher production volume, threatening their long-term sustainability. The episode explores empirical evidence from a controlled experiment showing that AI-generated websites lead to surges in package downloads but stagnant or declining GitHub stars, confirming the shift of attention from humans to machines. Shkoren concludes with a hopeful but cautious outlook: while AI automates code generation, the irreplaceable human role lies in problem framing, user understanding, and strategic thinking—areas where comparative advantage remains firmly human. The discussion extends to broader implications for software development, labor markets, and the future of knowledge work. Shkoren emphasizes that AI functions not as a replacement but as a powerful co-worker, augmenting productivity when paired with strong foundational thinking and structure. He warns against the risk of 'throwaway code' and highlights the importance of maintaining core open source libraries even as AI reduces dependency on them. Looking ahead, he speculates on a future where AI intelligence becomes locally accessible, potentially decentralizing the digital economy and challenging the dominance of centralized platforms. The episode concludes with a call to reframe how we teach programming and value developer skills in an age where English is the new programming language.