From the wave-0 evidence ledger
Recorded claims
Each claim states what the inspected source says, at the recorded location—bounded by its scope and graded by its confidence. Nothing here is a synthesis across works.
Across 246 real tasks performed by 16 experienced open-source developers in repositories they knew well, allowing early-2025 AI tools increased completion time by 19% on average.
Scope: Experienced maintainers in mature open-source repositories; small sample and specific tooling.
Participants nevertheless predicted that AI would make them about 20% faster, showing a large perception-versus-measured-performance gap in this setting and a reason to instrument adoption rather than rely only on user belief.
Scope: Same small expert sample; does not establish a general user-perception bias magnitude.
Boundaries
Limitations & independence
Recorded at coding time, carried with the work forever. A claim without its limits is not evidence.
Recorded limitations
- 16 experienced developers
- 246 tasks in mature open-source repositories
- Developers had about five years' repository familiarity on average
- Early-2025 tools and interfaces
- Preprint
Source independence
Nonprofit research organization study; not authored as a software-vendor case in the inspected record.