Microsoft announced plans to develop a large-scale, cutting-edge AI model next year to build its own internal solution to replace OpenAI and Anthropic's most powerful AI tools. Microsoft AI CEO Mustafa Suleyman stated in an interview, "We must deliver absolutely cutting-edge technology. By 2027, the goal is to have models that are truly state-of-the-art in text, image, and audio generation and response." Microsoft's AI division unveiled a speech transcription model on Thursday, which reportedly outperformed competitors in 11 of the 25 most commonly used languages. However, similar to the division's previous speech and image generation models, this model is a highly efficient, specialized tool, requiring less training data than general-purpose models such as Claude 3 Opus or OpenAI's GPT-4. Suleyman stated that Microsoft is consolidating computing power to build models with broader capabilities, and the company began using a set of Nvidia GB200 chips to expand its computing resources last October. He said, "Over the next 12 to 18 months, we will gradually upgrade to cutting-edge computing power." (Jinshi)