Microsoft has long been a major player in the field of artificial intelligence (AI), and with its early support for OpenAI, it has emerged as one of the biggest beneficiaries of the generative AI boom. To further solidify its position, the tech giant has been proactive in diversifying its AI portfolio. In line with this strategy, Microsoft has recently announced that it will be incorporating Meta Platforms’ rival open-source AI model, Llama 2, into Azure AI Studio. This move aims to offer customers a convenient and hassle-free way to access AI models through a model-as-a-service (MaaS) framework.
Similar to the concept of software-as-a-service (SaaS), MaaS allows users to leverage AI models like Llama 2 on-demand over the web. This eliminates the need for complex installations on private clouds, hybrid clouds, or other locations. Even with the assistance of Microsoft experts and trained IT personnel, deploying AI models can be a challenging task for enterprises. Recognizing this, John Montgomery, Microsoft Corporate Vice President of Program Management for its AI Platform efforts, stated, “Not every customer wants to think about operating infrastructure, which is why at Ignite we introduced Models as a Service.”
Azure AI has previously enabled users to deploy models on their own infrastructure effortlessly. However, the introduction of MaaS provides a significantly easier alternative. Model-as-a-service operates through API endpoints, where users can simply call the Azure OpenAI Service. Now, Microsoft is expanding its MaaS offerings by making Meta’s Llama 2 available through Azure AI in public preview. This includes a range of Llama 2 models, such as Text Generation and Chat Completion, offering customers a diverse set of options when it comes to generative AI.
Microsoft’s decision to integrate Llama 2 into Azure AI Studio is a strategic move. It allows Azure cloud storage and service customers to access a wider range of AI models at a lower cost compared to offerings from OpenAI, a close partner of Microsoft. Llama 2 has emerged as a popular choice among users and enterprises as an open-source model for generative AI. This move by Microsoft not only provides more options to customers but also demonstrates its commitment to diversifying its investments in AI.
However, Microsoft’s collaboration with OpenAI is still strong. In addition to introducing Llama 2, the company has also made OpenAI’s latest publicly accessible AI model, GPT-4 Turbo with Vision, available to Azure customers through Azure AI Studio. This model enhances AI capabilities by enabling analysis and description of photos and visual material. The availability of GPT-4 Turbo with Vision on Azure has already attracted customers’ attention, with companies like Instacart and WPP utilizing its features.
Azure AI Studio also offers tools for fine-tuning all the models provided. This means that customers can optimize the models according to their specific needs and requirements. As Microsoft continues to compete in the AI cloud wars, the expansion of Azure AI Studio becomes a critical factor in its success. The question now arises: how many more models will Microsoft be able to add to its AI studio? Could Mistral or Deci be next in line?
Microsoft Empowers AI Research
Aside from the integration of Llama 2 and GPT-4 Turbo with Vision, Microsoft recently released its own small language model (SML), Phi-2. It is important to note that Phi-2 is solely intended for research purposes and cannot be utilized for commercial applications. Nonetheless, the availability of Llama 2 and GPT-4 Turbo with Vision provides a powerful foundation for users seeking to explore the potential of AI.
Microsoft’s introduction of Llama 2 to Azure AI Studio marks a significant milestone in the accessibility and availability of AI models. This move expands Azure’s offerings, providing customers with a broader range of options and lower-cost alternatives. With the inclusion of OpenAI’s GPT-4 Turbo with Vision, Microsoft further strengthens its partnership and commitment to empowering AI research and development. As the competition in AI continues to intensify, Microsoft’s ability to expand its AI model library becomes crucial for its success in the ever-evolving AI landscape.
Leave a Reply