Hi OpenAI Team,
I hope you’ll allow me to share some honest feedback. I’ve been using ChatGPT on Android quite a lot, and I really appreciated how natural the old voice-to-text flow felt in daily life.
Before, it worked in a way that gave me a bit of breathing room:
I could record my message at my own pace,
Review the transcription afterward to catch any slips,
And decide whether to send it, or fix it if needed.
That little pause helped a lot, especially because I often use this while managing real life around me — I have kids, there’s noise, and like anyone, I sometimes need to gather my thoughts before sending a message. Honestly, it felt very human and respectful of my pace.
But with the recent changes:
The send button feels smaller and is awkwardly placed next to cancel, making it tricky to use confidently.
Sometimes my messages are sent automatically before I’m ready, or worse, if I speak for too long, they just disappear.
And the biggest loss for me is the automatic “read the next message aloud” feature.
When it worked, it made my life so much easier. Especially when driving or when I couldn’t keep my eyes glued to the screen, it let me stay engaged and focused, without constantly reaching for my phone.
I also want to be open about something else — I’ve found that Advanced Voice Mode adds a bit of stress for me. With background noise or when I pause to think, it feels like the system is unsure whether I’m done speaking, and that unpredictability makes me tense up. It’s a bit hard to explain, but it even causes physical discomfort at times. The old way let me breathe, think, and feel calm.
I really miss the balance and smooth flow the old setup gave me. It wasn’t about fancy features — it just felt natural, and it fit into the rhythm of my daily life.
I’m not writing this to make demands, or claim my experience is the only one that matters. I just wanted to share this in case it helps give a fuller picture of how these changes affect users like me. Maybe there’s a way to bring back some of that calm, manual control, or at least consider users who rely on it.
Thank you for taking the time to read this. I appreciate all the work you’re doing, and I hope this kind of feedback is helpful.
Warm regards,
Melos