I understand why people use it but as a recipient I really hate voice messages. Here in Spain they are becoming more common too. The problem is as the article mentions. Difficulty searching back, not being able to use Google translate (I'm still learning Spanish), and the time it takes to listen to them. Sending a voice message is faster than typing for the sender but much slower than reading for the recipient which…
It would seem obvious for the transcription and the sound file to be available to both sender and receiver of the message.
I think so far the sticking point is that message transcription is still hard to do with any accuracy on-device, and if you use a server then you can't claim e2e encryption.