Voice Mode Interruption Issue: Timing Misalignment During Conversations

Hi everyone,

I’ve noticed an issue with the advanced voice mode that’s affecting the natural flow of conversations. While I understand the system needs to process speech in chunks, the current implementation is creating some usability challenges:

  1. Main Issue: There’s a timing mismatch between user input and system response that leads to frequent interruptions.
  2. Specific Behaviors:
  • The system interrupts during natural speaking pauses
  • When trying to continue speaking, there’s an awkward overlap where both the user and system cut each other off
  • This creates a cycle of starting and stopping that makes it difficult to complete thoughts
  1. Impact:
  • Conversations feel disjointed
  • It’s challenging to maintain a natural speaking rhythm
  • Users often have to restart their sentences multiple times

Has anyone else experienced this? I believe adjusting the pause threshold or implementing a more sophisticated turn-taking mechanism could help resolve this.

Any insights or workarounds would be greatly appreciated.

I agree, there needs to be a control for the AI to notice that you are simply recollecting your thoughts and that the pause doesn’t mean it should now start talking.

This has been terrible for me, the voice mode constantly interrupts, and I am so not finished with my full thought…if I put my keyboard in voice to text, I can speak volumes, including just pausing to collect my thoughts, and then when I submit it, AI will process it and everything’s great. But it’s impossible to get more than a few sentences in on multiple AI is not just chatGPT to be able to say, let’s have a conversation but don’t talk until I say go, or tap the screen or something… It’s utterly useless if I can’t get a thought out before it tries to start responding. It’s just such a weird thing it’s happening on all different platforms, at least for me