The discussion highlights significant dissatisfaction among users with NumPy's syntax and API, particularly regarding multidimensional arrays, which has prompted users to seek alternatives like Xarray and Julia. Key issues mentioned include: - Complexity in handling arrays with more than two dimensions. - Verbosity and inconsistency in syntax compared to MATLAB and Julia. - The difficulty in abstracting functions due to how NumPy handles dimensions and operations. - Unpredictable behavior of certain operations, leading to confusion and bugs. - Lack of standardization in Python libraries for data science, resulting in extra overhead in data format conversion. Many users appreciate the intuitive nature of alternatives while recognizing that NumPy remains broadly used despite its flaws, suggesting a need for improvement in its API and documentation.