What even is the point of that? Imposing puritan values on customers or some nonsense?

The point is probably, by making the assistant more human-like, you'll grow more attached to it and use it more, depend on it. Perhaps even trust it when it suggests that you should buy this or that, or that it cannot open the airlock, Dave. Also perhaps some focus group showed that you need it to be more human-like so it doesn't scare off tech-adversies.

Or some random dude at google thought it'd be funny. But I put my money on it being used for improving some metric somewhere.

It may me throw all my Google Home pucks in the garbage can and replace them with Echo Dots. When you tell Alexa to cram it she shuts up.

The point is that some programmer (or probably PM) think it’s a better resource to custom program that response rather than to address fixing whatever the device did to make you curse at it.

I think this is an assumption that the device is right and the user is wrong. So chiding them on manners is thought to be the right product move, rather than using it as input for finding better device behavior.

They think giving their digital assistants a “personality” will endear you to them, without consideration of context.