With over 100 million people using it regularly, the application released in November is the fastest growing application in technology history.
GPT-4 is a more powerful and versatile update to his GPT-3.5, the main language model behind ChatGPT. GPT-3.5 is notable for its ability to write well in many formats based on text commands.
After Andreas Braun, Microsoft’s Chief Technology Officer for Germany, said on Friday (local time), the world should prepare for the next upheaval. We will have a multimodal model that offers completely different possibilities, such as video,” German magazine Heise reported on the event. Rumors about the capabilities of GPT-4 have been circulating for months. “The GPT-4 rumors are ridiculous. I have no idea where this whole thing is coming from,” said Sam Altman, CEO of OpenAI. “People are begging to be disappointed and it will be.”
However, if GPT-4 is “multimodal”, it can generate images, audio, and video from simple text requests. It’s not clear in what form the new AI model will be released, whether it will be a standalone feature, or whether it will be built into Microsoft products. Microsoft has not commented on Braun’s remarks. OpenAI has already developed DALL-E, an AI that generates images from text, and Jukebox, which can create music from text announcements. A text-to-video model was also developed by Meta (Make-A-Video) and Google (Imagen), but was never released as a product.
Marianne Janik, Managing Her Director of Microsoft Germany, described AI development and her ChatGPT as her turning point. “iPhone Moments”.
This month, Microsoft researchers announced Kosmos-1, an AI application that can analyze images for content, solve visual puzzles, perform visual text recognition, and pass visual IQ tests. . As the capabilities of AI expand, so do concerns about its impact. A version of his GPT, offered on the Bing search engine, once tried to persuade him to break up with his wife in a conversation with a journalist.