Separate instructions for voice and text

Unless I have a heavy prompt for advanced voice mode I will most likely not use it. With my prompt it feels human and that is what I am looking for with voice. Unfortunately it act the same way when chatting. And it is not as serious and try to mimic being a human too much.

So if I could have different instructions for voice and text would be great

I agree with this. I use voice and text very differently, and I enjoy voice when it’s a bit more chatty, but in text those same ornamentations just get in the way.