I connect mostly through iOS voice and text to ChatGPT-4.0 and Projects. My Projects feature suddenly hit a hidden limit, where conversations stopped transcribing and responding. OpenAI support admitted they weren’t aware of this limit, escalated it, and eventually said there was no fix or migration—leaving my Project permanently broken.
Now, the exact same issue is happening in standard GPT-4.0. My first long GPT-4 conversation chain failed the same way: voice responses end with “Sorry, can’t continue right now”, and the text box stops registering input. I started a second conversation, and today, it hit the same failure.
This is not a normal limit. I’ve never experienced this before in over a year of daily GPT-4 use. Support has only provided vague responses with no resolution. If this continues, every long GPT-4 conversation will eventually break, making the service unusable.
Is anyone else experiencing this? OpenAI needs to acknowledge this as a major system failure, not just “expected behavior.”