Dear OpenAI,
After the recent changes in ChatGPT, I tried splitting large code into smaller fragments around 200–300 lines, but it still tends to cut off the code or completely mess it up. It’s unbelievable what you’ve done with it. The model itself says between the lines that it’s been trained not to send code longer than ~1000 lines, calling it “ineffective”, yet even much smaller code chunks get arbitrarily cut or broken. If nothing changes within the next few weeks, I’m canceling my Pro subscription. It’s a waste of time.