Apple Breaks Barriers: Craig Federighi’s New Approach to Siri with Third-Party AI Models

Apple's New Approach to Siri According to a report from The Information, Craig Federighi, Apple’s senior vice president of Software Engineering, has made significant changes in the development of Siri. For the first time, Apple engineers are now allowed to use third-party large language models (LLMs) to build Siri features, marking a departure from the previous policy that restricted them to only benchmarking against external models.

Background on Previous Constraints

Until recently, Apple's engineers were limited to using in-house models for developing Siri features. This restriction frustrated many team members who found Apple’s internal LLMs underperforming compared to those developed by companies like OpenAI. In 2023, managers explicitly prohibited the inclusion of third-party models in final Apple products.

New Policy and Implications

Under Federighi's new directive, engineers have the flexibility to use any open-source LLMs to enhance Siri’s capabilities. This change is expected to improve Siri’s performance and user experience, as it allows Apple to leverage the best available technology regardless of its origin.

User Perspective

For users, this pivot means a more capable and responsive Siri, which ultimately enhances their overall experience with Apple devices.