It's a foot gun, and the behavior is often surprising until you get used to it and sometimes after, but Powershell would be very painful to write without it. Basically all functions return a list in Powershell, including all user-defined functions, so you'd need to unwrap singleton lists all over the place.
But if they COULD be lists of >1 then by unwrapping it you have two scenarios to handle instead of one. How is this an improvement? And if it is ALWAYS a list of one why return a list?
All Powershell functions are basically python generator functions. It could have been designed to have more traditional functions, but I find generators fit well into a language based on piping streams of values between functions.
Usually, the distinction between a single item, a flat list, and a list of lists isn't important in Powershell, because commands are written with Powershell's behavior in mind. It's extra painful when it does matter, but it's a trade off against being more verbose every time it doesn't matter. Ultimately, it's personal preference; for an interactive shell, I like the tradeoff Powershell made, and I think it's a cool design space to explore, but I do think there's a lot of room to improve the implentation of the concept to make it less surprising and less painful to choose the alternate behavior when needed.
