Microsoft is building LLM more powerful than Google, OpenAI’s, report says – Technology News

Microsoft is reportedly working on an in-house large language model (LLM) that may be more powerful than Google and OpenAI’s. Recently the new model, named as MAI-1, is being overseen by Mustafa Suleyman, the Google DeepMind co-founder and former CEO of AI startup Inflection. Reports suggest that MAI is citing two Microsoft employees with knowledge of the effort.

Is MAI-I better than OpenAI?

Reuters reported that the MAI-1 will have roughly 500 billion parameters while OpenAI’s GPT-4 is reported to have one trillion parameters and Phi-3 mini measures 3.8 billion parameters. The new MAI-I is expected to drain out more data as it will be using high computing powers. Critics argue that in comparison with OpenAI this new model could be more expensive.

“MAI-1 will be “far larger” than the previous smaller, open-source models Microsoft had previously trained which means it will be more expensive,” Reuters reported.

A few days back Charlie Bell, Executive Vice President, Microsoft Security explained that Microsoft will be expanding its security structure. It launched the Secure Future Initiative (SFI) to prepare for the increasing scale and high stakes of cyberattacks. Experts suggest with MAI-I these structures can be developed with the intervention of AI.

What’s next?

Microsoft has reportedly invested about billions of dollars in OpenAI and deployed the ChatGPT maker’s technology across its suite of productivity software. This is expected to help Microsoft to take an early lead in the generative AI race.

According to the reports Microsoft has been setting aside a large cluster of servers equipped with Nvidia’s graphic processing units. In addition to this they are also  working with large amounts of data to improve the new model.

Follow FE Tech Bytes on Twitter, Instagram, LinkedIn, Facebook