Michael Lemire argues open source drives research

The author argues that open source serves as an effective research method by directly connecting software developers with real-world users, demanding use cases, and rigorous peer review through bug reports and pull requests. Drawing on examples from database and parsing research, the article contends that building software to solve actual bottlenecks leads to widely adopted tools like simdjson, fast_float, and simdutf. Furthermore, the author notes that AI tools appear to accelerate this collaborative open-source research model rather than replace it. Open article →