It is very likely (p > 0.75) that handling a small ball and an apple are more distinct use cases than simple hierarchical specializations, despite appearances.
The regular geometry of a ball greatly simplifies grasping; the probability that the same sensor is sufficient is high (about 90%).
However, an apple presents an intrinsic variability in shape and fragility, making the task of not damaging it much more complex, with a success probability with the same basic configuration being low (p < 0.3).
For example, for harvesting in an orchard, pressure sensors and adapted algorithms are needed, which are not necessary for a rigid ball, implying different capabilities.
It is not just a matter of degree but of task nature, with basic technical requirements that diverge significantly.