Dalle amounting to GPT-4 use and constant network errors != good time and very low actual GPT limits

The last week or two I constantly get GPT to time out due to network ‘failure to fetch’ this can amount to losing 10-20 generations in the 3 hour limit, adding on top of that that since I use dalle frequently and that contributing to GPT-4 limited use the limit of generations I can currently achieve effectively is 20 due to the previous factors highly reducing the limit/use.