Microsoft’s new MAI-1 AI model likely to challenge OpenAI’s GPT-4 and Google Gemini

Microsoft AI. Representational image from Wallpaper Cave

#Microsoft, #AI, #ChatGPT, #GPT4, #MAI1

Washington/IBNS=CMEDIA: Microsoft is developing a new large-scale AI language model called MAI-1, which could potentially rival state-of-the-art models from OpenAI, Google, and Anthropic, reports The Information.

According to reports, this marks the first time Microsoft is working on an in-house AI model of this magnitude since investing over $10 billion in OpenAI for the rights to reuse the startup’s AI models.

OpenAI’s multimodal large language model (LLM) GPT-4 powers not only ChatGPT but also Microsoft Copilot.

MAI-1’s development is being led by Mustafa Suleyman, the former Google AI leader who recently served as CEO of the AI startup Inflection before Microsoft acquired the majority of the startup’s staff and intellectual property for $650 million in March.

Although MAI-1 may build on techniques brought over by former Inflection staff, it is reportedly an entirely new LLM, Ars Technica reported, quoting two Microsoft employees familiar with the project.

MAI-1, with approximately 500 billion parameters, will be significantly larger than Microsoft’s previous open source models, such as Phi-3, requiring more computing power and training data, according to reports.

Microsoft’s MAI-1’s development suggests a dual approach to AI within the tech giant, focusing on both small locally run language models for mobile devices and larger state-of-the-art models that are powered by the cloud.

Apple is reportedly exploring a similar approach, highlighting the company’s willingness to explore AI development independently from OpenAI, whose technology currently powers Microsoft’s most ambitious generative AI features, including a chatbot baked into Windows, reports said.

However, the exact purpose of MAI-1 has not been determined so far, and its most ideal use will depend on its performance, reports The Information citing its sources.

According to reports, Microsoft has been allocating a large cluster of servers with Nvidia GPUs to train the model and compiling training data from various sources, including text generated by OpenAI’s GPT-4 and public Internet data.

Microsoft is likely to preview MAI-1 as early as its Build developer conference later this month, The Information reported citing its sources.