Replication talk is branding until someone loses a paper

Everyone now says replication. Hiring, journals, and grants still pay for a novel claim with a clean story. You can give the talk and keep the old incentives, and most departments do.

Branding is cheap. A result that dies in a second lab is expensive, and almost nobody's tenure file is priced that way.

Whose paper in your unit would you be willing to see fail in public?