#2048 token output limitation renders production of code useless

3 messages · Page 1 of 1 (latest)

trail pine
#

Developing code I need to grow code as I'm introducing changes. First off, it's very difficult to have the AI change code as it is not clear where and what gets replaced. Probably for that reason, it frequently reposts the entire code, which takes - even with the plus subscription ridiculously long.
Worse, however, it the fact that the output of that code often gets interrupted, output stops in the middle of anything (be that a variable, a function, etc.). Trying to resume the production of code and see the rest of that code proves very difficult if not impossible. It then either simply starts from the top again, or resumes from somewhere else - sometimes leaving out sections of code.
Producing code this way is simply not productive and renders the entire notion of having a code helper at your hand totally pointless. You can't create code if the output is limited to 2048 tokens.
What can you, what can I do to change this or better remove this limit?

meager sun
#

Is this why I keep getting "Hmm...something seems to have gone wrong. Maybe try me again in a little bit." after long conversations?

#

Having the option to delete responses from ChatGPT would clear up wasted space, at least in my case. I don't use it for coding though.