I don't think you should sneer at a triple-digit sample size that's justified with a power analysis based on their previous work.

I agree that you'd want to see replications before making any major decisions--these effects are often finnicky--but this seems to me like a reasonable-sized step forward.