> This is also all work that doesn't benefit the scientists replicating the paper. It only costs them money and time

If you build upon a result, you almost have to replicate it.

An acquaintance spent years building upon a result that turned out to be fraudulent/p-hacked.