Issue Summary

When initiating a follow-up question in an existing chat session, the response intermittently fails to appear. The interface shows the 'loading' indicator (flashing circle) next to the response icon indefinitely without returning an error or timeout message. This issue severely degrades the user experience and has been observed more frequently in the past week.

User Environment

Subscription: ChatGPT Plus
Usage: Primarily using the desktop app, web app also similar
Preferred Model: GPT-4o, but also with 4o-mini, o1-preview and o1-mini
Frequency: Happens almost every new chat session

Steps to Reproduce

Open a new chat session
Enter an initial prompt
Receive the initial response
Enter a follow-up prompt (same chat)
Observe that the 'loading' circle appears and persists indefinitely, without a response or timeout

Expected Behavior

The model should provide a response within a reasonable timeframe or present a timeout/error message

Observed Behavior

The 'loading' circle remains active indefinitely
No timeout or error message appears, even after extended waiting periods
In models that provide a narrative (like o1-preview), the narrative never starts, indicating no processing is happening.

Temporary Workaround

Click the 'Stop' button while the circle is flashing.
Click the pencil icon below the follow-up prompt and resubmit the prompt.
The response appears quickly, usually on the first or second resubmission.

Additional Observations

Occurs across multiple models
Consistent across both desktop and web apps
Behavior is the same regardless of the complexity of the follow-up prompt

Impact

High: Disrupts the user experience, requiring manual intervention
Frequency: Increasingly consistent over the past week, affecting nearly every new session

#No response to follow up questions in new or historical chat (no timeout/error message)