Looks like a "GPT text output condenser" might be a good project to work on.