GPT-4 is struggling with coding. It doesn’t test its output, causing developers to waste hours fixing mistakes. To stay competitive, OpenAI must integrate GPT into IDEs (VS Code, PyCharm) with real-time analysis and automatic code testing. Otherwise, Copilot will dominate. @OpenAI, please improve AI coding support!