The article analyzes Microsoft's AI Copilot strategy, highlighting its initial reliance on OpenAI's models and subsequent efforts to reduce this dependence by developing its own and customizing open-source models. This shift is driven by cost concerns associated with the high token usage of OpenAI's models and aims to improve Copilot's efficiency and affordability.

The author speculates on the potential commoditization of generative AI algorithms, drawing parallels to the evolution of internet access. Ultimately, the piece explores the evolving landscape of AI development and its implications for businesses and consumers.

This article is also available in English on my Medium page, «What does Microsoft’s AI Copilot strategy tell us about the future of AI?»