Sam started this. The comparisons would have come up anyway, but it's a lot harder to dismiss the claims from users when your CEO didn't tweet "her" before the release. I don't myself think the voice in the demos sounded exactly like her, just closer in seamlessness and attitude, which is a problem itself down the road for easily convinced users.

AI companions will be both great and dangerous for those with issues. Wow, it's another AI safety path that apparently no company is bothering exploring.