You are misunderstanding. He's saying when there's macroscopic phenomenon that are bound to quantum measurements, those can lead to different decisions.

For instance, if you say to yourself, I'm going to step to the left if a quantum measurement observes an entangled particle as spin up. And I'm going to step to the right if a quantum measurement observes the particle as spin down.

Assuming you make good on that, you are going to have two branches of the wave function:

1. Where you observed the particle as spin up, and you stepped to the right. And,

2. where you observed the particle as spin down, and you stepped to the left.

But this does not happen merely as a result of every radioactive decay. If a potassium isotope decays in your abdomen, and the gamma particle interacts with an atom in a skin cell, then the branch of the wave function is described as: There is a branch of the wave function in which that skin cell observed the decay, via the absorption of a photon. That doesn't mean you'll literally have a branch of the wave function as a result wheee you hopped on one foot, picked your nose, sat down, did push ups, ordered Thai food, and every counterfactual permutation. Carroll is NOT saying that.