1. Microsoft is reportedly developing an in-house large language model called MAI-1, a departure from its previous investment in OpenAI.
2. The move is expected to give Microsoft more control over the technology, following the hiring of Mustafa Suleyman, a former Google AI chief.
3. The decision to develop MAI-1 alongside using OpenAI models like GPT-4 for Copilot could give Microsoft a competitive edge in the AI landscape.
Microsoft is reportedly working on developing its in-house large language model called MAI-1, a departure from its previous focus on investing in OpenAI and using models like GPT-4. This move will give Microsoft more control over the technology, following the hiring of Mustafa Suleyman, a former Google AI chief. While Suleyman’s team has joined Microsoft, MAI-1 is expected to be an entirely new model, according to reports from The Information. This follows Microsoft’s significant investment in OpenAI for AI rights, potentially affecting the use of GPT-4 in Microsoft’s consumer-facing chatbot, Copilot.
The development of MAI-1 suggests that Microsoft may be building two large language models simultaneously, hedging its bets on the future of AI technology. While chatbot interfaces are popular, they come with high running costs due to data transfer and privacy concerns of sending queries to external servers for processing. Others, like Apple, are focusing on processing data on-device, but the complexity of large language models poses challenges for this approach. By developing both a smaller-scale LLM and a more ambitious one, Microsoft aims to offer a balanced solution that leverages the strengths of both models.
Overall, Microsoft’s investment in developing its large language model and its partnership with OpenAI signal its commitment to advancing AI technology and enhancing user experiences. With a strategic focus on building multiple LLMs, Microsoft aims to navigate the complexities of AI technology and provide innovative solutions to users in the future.