As already explained, there is still an important distinction. One would first have to be the subject of some kind of criminal investigation, and then your house could be bugged from that point in time.

When data is requested from a device like a smart speaker, suddenly everything that has been said, even before a potential investigation, would be accessible.

They can bug my house, but they can still only listen in on what is being said after the bug was installed.

Smart speakers transcribe everything they hear, meaning you lose all tone and nuances of speech patterns, creating a very dangerous cocktail when allowing law enforcement to do text searches or even worse, let AI sift through it.

Imagine a speaker that is “always on” while watching black hawk down, or Air Force one, transcribing every line, or worse, a line here or there.

I have the same opinion of wide surveillance as DNA registers. It allows for law enforcement to be lazy, and does not guarantee the right person is convicted. In fact it moves the burden of evidence from the prosecutor to the defendant. Suddenly you have to prove your innocence instead of law enforcement having to prove your guilt.