Still haven't seen a single reason to own one.
Additionally, my oldest child can put music on for herself with her voice, but can't read or spell enough to do it otherwise.
They use it to play music. Alexa play christmas songs. Alexa pause. Alexa resume.
That's all.
I know about homebrew options. I've tried some of them. The problem is convincing others. They want something that works every time all the time, not something I have to fiddle with every so often.
As for the security, it was my concern from day one, but when laypeople see the convenience and not some hidden boogeyman, you can't convince them otherwise. I lost that argument to my wife and it wasn't a hill I chose to die on. I'm just careful what I say openly, as you should anywhere. Plus we all carry phones around already.
Is the demographic that uses these preferentially si-valley people or is it widely spread in the US?
BTW I'm not particularly against home AI assistants myself. But I'm fairly against streaming my audio to a third party that can transcribe and give everything to any snooping fourth party though :) I did read about Google doing this with their assistant's audio, it ended up at contractors doing manual transcribing.
My anecdotal examples would be mostly middle-class families and college students in the US, not particularly valley-related. In my country these type of devices are on the market but not popular (only seen a couple), but they're priced out of the range of the majority of the population anyway.