Many promising ideas never reach experimentation—particularly where resources are scarce. This Opinion advances a conditional, idea-first track for publishing test-ready hypotheses when in-house or collaborative experiments are infeasible or not progressing. “Test-ready” denotes a clearly specified problem, a literature-grounded rationale, explicit falsifiable predictions, at least one feasible third-party test, a brief ethics/dual-use appraisal, an explicit open license, and a good-faith novelty scan. We outline a practical pathway that pairs preprints (for priority, feedback, and versioning) with optional defensive publication to preserve freedom to operate when patenting is impractical. Generative AI can assist, for example, by mapping analogies across fields, probing near-neighbors, stress-testing plausibility, and suggesting alternative test designs that originators may not foresee—provided its use is disclosed, all claims and citations are human-verified, and uncertainty is handled through transparent prompting and aggregation protocols. To respect limited reviewer bandwidth as manuscript throughput rises, we suggest simple, fast triage and purpose-built quality checklists for idea-first submissions—without imposing a single universal metric. Positioned upstream of Registered Reports, test-ready idea papers can shorten the path from insight to evidence while recording provenance and discouraging enclosure. The aim is to move credible, citable hypotheses from overlooked notebooks into the hands of groups able to test them.
Maxwel Adriano Abegg (Tue,) studied this question.