
Is open-source software still a sustainable strategy in the age of AI?In this episode of Databased, Convex co-founders Jamie Turner and James Cowling examine how AI coding tools, large language models, and increasingly capable software agents are changing the economics of open-source software.For developers, open source has historically been one of the best ways to learn how production systems work, explore software architecture, contribute to meaningful projects, and build new applications on top of shared infrastructure. However, AI models can now analyze large codebases, explain complex implementation decisions, generate competing software, and potentially recreate products without directly copying the original source code.Jamie and James discuss whether this shift makes copyright, software licenses, patents, trade secrets, proprietary test suites, and operational expertise more important than they were before. They also explore why many successful open-source projects are funded by commercial companies, what happens when larger organizations use AI to replicate startup innovations, and why the team behind a codebase may still be more valuable than the code itself.The conversation also covers Convex’s approach to source-available software, the Functional Source License, closed-source testing infrastructure, deterministic simulation testing, network effects, operational excellence, and the challenges of protecting a technical advantage while still contributing to the developer community.Developers interested in open source, AI-assisted coding, backend infrastructure, databases, software licensing, startup strategy, full-stack development, distributed systems, and the future of software engineering will find a nuanced discussion about where the industry may be heading.Timestamps00:00 The Future of Open Source in the AI Era08:32 Copyright, Patents, and Protecting Small Companies16:57 How AI Makes Large Codebases Easier to Understand22:25 Can Large Companies Use AI to Clone Startups?28:24 Why Trade Secrets Are Becoming More Valuable34:10 Open Models, Distillation, and the Future of Copyright38:57 Convex, Source-Available Software, and the FSL45:24 Why There May Be Less Open Source in the Future