GPT-4 is coming soon, and it will be multimodal! Multimodal models are types of AI models that are trained to understand multiple forms of data such as text, audio, video, images. Director of Technology at Microsoft Germany, Andreas Braun, has announced that OpenAI will officially roll out ChatGPT-4, which is going to be around 600 times more capable than GPT-3. In order to understand the power of the upcoming GPT-4, it’s essential to compare both versions. GPT-3 works on 175 billion parameters, while GPT-4 will work on 100 trillion parameters, which is a massive improvement.
#youtubeshorts #chatgpt #youtube