On Monday, Microsoft revealed a collaboration with French startup Mistral AI in the field of artificial intelligence. This partnership could potentially decrease the company’s dependence on OpenAI, the creator of ChatGPT, for its future chatbot and generative AI offerings.

Mistral AI was founded less than a year ago, but has already been recognized by Microsoft as an “innovator and trailblazer” in the forefront of developing AI systems that are both efficient and affordable.

The specific amount of money involved in the deal between Microsoft and Mistral was not revealed, but Microsoft did mention that it includes a minor investment in the startup company based in Paris. This indicates that the investment is much smaller compared to the billions of dollars Microsoft has invested in OpenAI over the course of several years. This has caught the attention of antitrust regulators in both the United States and Europe.

On Monday, Mistral made a public test edition of their chatbot, Le Chat, available. Due to overwhelming interest, a company representative stated that it was temporarily inaccessible for some time during the day.

The company also announced its newest large language model, Mistral Large, which it claims is in the same league as competitors such as OpenAI’s GPT-4, Anthropic’s Claude 2 and Google’s Gemini Pro and will be available on Microsoft’s Azure cloud computing platform. Mistral has also previously said it is teaming up with other big cloud providers including Amazon and Google.

Mistral has caused a great stir by securing substantial financial backing from investors and earning a valuation in the billions shortly after its launch last spring. The company was founded by three former Google and Meta employees from France: CEO Arthur Mensch, Chief Scientist Guillaume Lample, and Chief Technology Officer Timothee Lacroix.

The company has promoted a transparent method for creating AI through the release of key elements in certain systems, in contrast to companies like OpenAI who keep them confidential. However, their newest and most advanced model, Mistral Large, will not follow this approach, based on the company’s website. Mistral did not provide a comment in response to requests on Monday.

As the European Union was finalizing its Artificial Intelligence Act, which includes a range of regulations for AI, Mistral objected to the limitations being placed on the foundation models used in generative AI systems. Mensch expressed concerns on social media that the EU’s suggested approach of a two-tier system would hinder new and innovative players in the field.

Source: wral.com