Absolute data points rarely make sense, they are rigid and inflexible, which is not good against an always changing environment.
You should be able to recognize the same scene both with a clear sky and a partly cloudy one, and it’s hardly important that this tiger had a deeper voice, the important part to recognize is the fact that there is a tiger. But even something as trivial as movement coordination is constantly updated by feedback mechanisms, like training with weights and dropping them in the end will temporarily make the movements “weird”.
You should be able to recognize the same scene both with a clear sky and a partly cloudy one, and it’s hardly important that this tiger had a deeper voice, the important part to recognize is the fact that there is a tiger. But even something as trivial as movement coordination is constantly updated by feedback mechanisms, like training with weights and dropping them in the end will temporarily make the movements “weird”.