Microsoft has launched its own artificial intelligence models: MAI-Voice-1 and MAI-1-preview, marking a departure from reliance on external partners like OpenAI. The MAI-Voice-1 is a groundbreaking speech model that generates one minute of audio in under a second using a single GPU. It powers features like Copilot Daily, which narrates news stories and creates podcast discussions. Microsoft emphasizes that voice will be crucial for future AI interactions, offering high-fidelity audio for single and multi-speaker scenarios.
The MAI-1-preview, trained on 15,000 Nvidia H100 GPUs, is designed for real-time user assistance and will be integrated into Copilot for text-based tasks. Microsoft aims to deliver specialized AI models tailored for consumer needs, focusing on enhancing user experience. With ambitious plans for future developments, Microsoft AI Chief Mustafa Suleyman highlights the aim to optimize models for consumer applications rather than enterprise-centric solutions.
Stay tuned for updates on Microsoft’s AI advancements in the consumer space!
Source link
