Disparity in comparison between different sized arrays from cpython numpy

[hgomersall]

In the following code snippet, cpython numpy and pypy numpy evaluate the comparison differently:

cpython:

>>> import numpy as np
>>> a = np.array([1, 2, 3, 4])
>>> a != a[:2]
True

pypy:

>>>> import numpy as np
>>>> a = np.array([1, 2, 3, 4])
>>>> a != a[:2]
False

It seems to me that the cpython way is more intuitive, irrespective of consistency. The arrays aren't equal.