The news that Microsoft has invested $16mn in Mistral, a French artificial intelligence start-up (Report, February 27), is a welcome development and testament to the fact that open source — where technical details are released publicly as opposed to being closed to third parties — is going to play a pivotal role in shaping the future of AI.
Until now, people thought of Microsoft as an “OpenAI shop”, so it’s good to see it spreading its bets. Not content to compete with their own models alone, hyperscaler cloud service providers are also now vying to be the place to host and run a variety of AI models.
These “AI model marketplaces”, like Google’s Model Garden, will make AI more accessible, helping to drive new areas of innovation.
I believe most of the innovation around developing AI models will be driven by open source in the future. We’re already seeing a trend towards more tailored, more specific, lighter-weight, and possibly more price-efficient models that can be targeted at very specific challenges and issues, which plays well to open source’s strengths.
AI presents as many questions as answers — particularly around safety, with rogue models already in use — but it’s here to stay. Instead of fighting the tide, it’s better to embrace AI and help to steer it on a safer course.
Matt Barker
Global Head of Workload Identity Architecture, Venafi, London WC2, UK
This post was originally published on Financial Times
