No wonder voice models and speech-to-text models have gotten exponentially better.
Daniela is right on the money.