labs · philosophy
$ Research Principles
Five rules we set before starting any experiment. Not aspirational — operational.
principle_1
$ Fail publicly
We publish negative results. We log failed experiments with the same care as successes. Bad data shared early saves other labs months of duplicated wrong turns. L-009 exists because Argonaut failed on-chain economics — we wrote it up, shipped the dataset, moved on.
operational detail
Every failed experiment gets an internal post-mortem. If the failure is replicable and the dataset is useful, we publish. Our public failure rate (3 this quarter) is a feature, not a bug.