Update examples away from rng_seeder

[ricardoV94]

https://github.com/pymc-devs/pymc/pull/5787 will remove the rng_seeder API.

In exchange, it introduces the possibility of passing RandomState or Generator (e.g., created via np.random.default_rng(seed)) to all random sampling functions. This would probably be the easiest way to update the pymc-example notebooks. I think it is already a best practice to create a rng at the beggining no?