Why we use pointers in the last example?

You're referring to the Increment and Set methods on the Minutes type, right?

type Minutes struct {
    value int
}

func (m *Minutes) Increment() {
    m.value = (m.value + 1) % 60
}

func (m *Minutes) Set(newValue int) {
    m.value = newValue % 60
}

If the m receiver parameter to Increment were not a pointer, it would receive a copy of the Minutes value we want to update. Increment would update the value field on the copy, and then it would discard the copy.

Instead, Increment needs to receive a pointer to a Minutes value. That will allow it to update the field on the original Minutes value, not the copy.

All of this is the same for the Set method as well.

This is exactly like our halve function that needed a pointer parameter in the pointers video, except that in this case we need our receiver parameter to be a pointer.