The zero-cost fallacy: Open source software in the agentic era

Open source software is widely treated as free, but that perception papers over the real human cost of maintenance. This ThoughtWorks piece argues that the asymmetry between value extracted and value returned to maintainers is a structural failure — one that generative AI is now intensifying.

Maintainers of load-bearing open source projects report financial destitution, burnout, and pressure from corporate consumers who demand support without reciprocation. The economics were already strained; AI adds new stressors. Low-quality, AI-generated pull requests and issues are flooding repositories, turning maintainers into unpaid code reviewers sifting through plausible-sounding nonsense. Meanwhile, traditional trust signals like GitHub stars have become meaningless as AI tools mass-produce repos, stars, and marketing that mimic genuine community engagement.

The authors trace the root problem to permissive licensing, which enabled the vast open source ecosystem but also created conditions for exploitation. Restrictive licenses offer a partial fix but force maintainers into enforcement roles and trigger procurement gatekeeping. The result is an asymmetrical tragedy of the commons: critical digital infrastructure rests on a handful of uncompensated people, one resignation away from cascading failure.

An emerging thesis offers a longer-term escape: if coding agents can reimplement functionality from behavioral specifications rather than importing entire libraries, the economic calculus around dependencies changes. But this shift — from shared maintenance to disposable, spec-driven reimplementation — would also deny maintainers the recognition that is often their primary compensation.

The piece closes with three concrete recommendations: treat dependencies as active responsibilities rather than passive gifts; invest in rigid supply chain auditing that goes beyond superficial trust signals; and formalize contribution budgets as basic risk mitigation, not charity.