As of a few hours ago, I’ve been getting network errors about 50% of the time, and it only seems to be getting worse, which is more than annoying. I actually hadn’t been experiencing many network errors since around May, and I’ve never experienced them this frequently ever before.
At first I thought it was because I ask it for long outputs (even though it hasn’t been an issue all these previous months), but even without that I’m still getting errors. It’s especially irritating because you can’t tell the AI to continue from whatever was generated before the error, and although I never even go near the number of outputs allowed every three hours, it’s still a waste of time.
I’ve also noticed that it now cuts off outputs at around 3500 characters (and that the option to continue incomplete outputs was reimplemented) when it used to go beyond 5000, and I’m wondering if that’s related.
Edit: So far it seems to be OK after switching to Wi-Fi, which is weird, because the only reason I even started using ChatGPT on mobile data was because I felt I was getting the network errors more often on Wi-Fi.