You already said the right answer. It's the latency.

Sound is analogue. Both when it's created and when it's played back. Transforming it to digital takes time. It makes sense to avoid that transformation in the signal chain for as long as possible before the "interface", to avoid doing it more than once.

When seeing a band play live, you might be able to appreciate the fact that they're technically forming one electric analogue circuit..

Also appreciate that vinyl records and tape can be made without ever being digital. I think it's pretty wild that we can even take a sound, put it on a record and play it back, thereby transferring data without it being digital. The whole process is much more interesting.

Imagine someone hitting a drum so hard that it makes a microphone membrane move, which makes an electric current, that pushes a needle into a record making a dent deep enough that your record player can feel it on the needle and create an electric signal to move your speaker membrane.

It might take some time to do, but when your ears hear that soundwave its basically the same motion that the drummer did originally. It has not been converted to a digital representation of what happened and back. It is the physical "shadow" of what actually happened.

I think it's kinda cool. I make digital music myself and while it has other cool stuff, I'll never bad mouth analogue.