Hi. Pandas is broken with numpy master :)

Reproducing code example:

In [1]: import numpy as np
In [2]: out = np.empty(3, dtype="object")
In [3]: out[:] = [np.int64, np.int64, np.int64]

Error message:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-3-d20bd885ed9b> in <module>
----> 1 out[:] = [np.int64, np.int64, np.int64]

ValueError: invalid __array_struct__

Numpy/Python version information:

Installed from a wheel with pip install -i https://pypi.anaconda.org/scipy-wheels-nightly/simple numpy

1.20.0.dev0+2d12d0c 3.7.6 (default, Dec 30 2019, 19:38:28)
[Clang 11.0.0 (clang-1100.0.33.16)]

I don't immediately see what commit could have caused this. #16850 mentions object-dtype scalars, but doesn't look quite right.