The Information newspaper reported that Microsoft has developed a new large language model for artificial intelligence within the company to compete with Google and OpenAI.
The report mentioned that Mustafa Suleiman, a prominent figure in the field of artificial intelligence and co-founder of DeepMind, as well as the former CEO of Inflection, is responsible for the new model known internally as MAI-1.
Microsoft has not yet clarified the exact purpose of the model, as it will be determined based on its performance quality. The report suggests that Microsoft may unveil the new model during the Build developers conference later this month.
According to the report, MAI-1 is much larger than the previous small open-source models Microsoft had trained earlier. This means that its cost is higher due to the need for significant computational power and training resources due to the complexity of the training process.
Microsoft launched a small artificial intelligence model last month known as Phi-3-mini, aiming to attract a broad customer base through a variety of cost-effective offerings.
The company has invested billions of dollars in OpenAI and integrated its technology into its suite of production software, allowing it to take the lead in early competition in the field of generative artificial intelligence.
Microsoft has allocated a large set of servers containing NVIDIA GPUs, along with large amounts of data, to enhance the MAI-1 model.
The report mentions that MAI-1 contains approximately 500 billion parameters, while OpenAI’s GPT-4 contains a trillion parameters.
In March, the giant software company appointed Suleiman as the head of the new Consumer Artificial Intelligence unit, along with several Inflection employees. Suleiman brings extensive experience to this project.
Microsoft acquired Inflection and its intellectual property rights for $650 million, though the new large language model MAI-1 is an independent initiative, regardless of its potential reliance on training data from the startup company.
MAI-1 is part of Microsoft’s strategy to enhance its presence in the field of artificial intelligence.
Microsoft’s commitment to artificial intelligence is evident through its increased capital expenditures, with the company’s spending rising by 79 percent to reach $14 billion, with a significant portion of these expenses allocated to artificial intelligence infrastructure.
Microsoft faces challenges in meeting the growing demand for intelligent computing power, due to a lack of infrastructure in data centers, affecting its ability to expand and utilize resources.
Artificial intelligence services are an essential part of Microsoft’s Azure cloud computing platform.
Azure’s revenues saw a 31% increase, with artificial intelligence playing a role in this growth.
Microsoft plans to increase its investments in cloud infrastructure and artificial intelligence. The company aims to increase capital spending to meet the growing demand, particularly for cloud infrastructure.
