On Tuesday, OpenAI announced GPT-4, a large multimodal model that can accept text and image inputs while returning text output that "exhibits human-level performance on various professional and ...