The type parameters of np.number and its abstract subtypes now default to Any.
This means that typecheckers will now infer np.floating as np.floating[Any], instead of rejecting it (when configured in strict mode).

This affects

• np.number
• np.integer
• np.signedinteger
• np.unsignedinteger
• np.inexact
• np.floating
• np.complexfloating (by building upon TYP: Optional 2nd numpy.complexfloating type parameter #27420)

The reason for choosing Any as the default, instead of setting it to the upper bound (numpy.typing.NBitBase), is because these type parameters (TypeVars) are invariant.
If they weren't invariant, then x: int16 = int8(3) would be a valid statement, which we probably don't want. So the invariance means that x: number[NBitBase] = int8(3) is rejected, but x: number[Any] = int8(3) is accepted.

I usually tend to avoid using Any, but because those _NBit type-parameters aren't actually annotating anything, the Any won't affect anything "outside" of the number types, so it's no problem to use it here.