Get ready for OpenAI's GPT-4, a unique multimodal language model that can operate across multiple input types, including video and sound.