I have only recently upgraded to the Pro account and I’m using ChatGPT nearly all day long at this point.
To say it has become a key part of my workflow and had a significant impact on my work is a gross understatement.
More than anything, I use the Advanced Voice Mode. Periodically, and with increasing frequency, it will fail to append a response to the chat transcript. Then ANY attempt to get it to fix the issue fails. No amount of trying to get it to either fix it, or even admit that there is an issue, has any impact—other than on my blood pressure. It is much more broken that it tries to fall back to thinking that it can just do things in context, when the ONLY acceptable response when there is an error when it tries to make the API call that had failed, is to acknowledge it. Maybe for the casual general public, having a fallback that glosses over things might be a thought, but certainly not for Plus users and above. It should at least be an advanced setting to set it in strict mode so that it ALWAYS acknowledges issues and stops misleading the user into thinking it can be fixed by just trying harder.