Ask HN: Is the future of audio compression TTS?

I'm coding an audio streaming service right now and can't help wondering... now that computers can speak in any voice (i.e. tacotron) isn't the best compression: Step 1. Take the audio from the mic and convert it to english. Step 2. Send that english down the wire, it is TINY compared to even opus compressed audio. Step 3. Have computer at other end read that english in the desired voice. (I say english but could be spanish or German or etc.)