Apple is reportedly testing a revamped Siri powered by Google Gemini
A new report by Mark Gurman at Bloomberg tells us that Apple works with Google to test a personalized version of Gemini who could feed the Vocal Siri vocal assistant from the next generation of Apple in the cloud.
The report indicates that Apple is in “early discussions” and that Google is currently forming a version of its Gemini AI assistant which can work on Apple’s cloud infrastructure. This is similar to the partnerships that Apple explored with Claude d’Anthropic or the Openai Chatpt. Apple creates a new generation, “LLM Siri” using its own models while developing an infrastructure that could see the new SIRI propelled by a third -party model. We wrote on this effort here.
According to Gurman’s report, Apple is still “several weeks” to decide to follow the opportunity to follow Apple’s own internal models for the new Siri or use external partners technology.
Apple uses an “cooking” approach, opposing the new generation Siri internally (Linwood, named code) to a new generation SIRI propelled by external models (Glenwood named by Code). The company has not yet decided which external technology was the best, and has not been financial with them. Supposedly, Claude d’Anthropic was the first favorite, but Anthropic demanded such a high price that Apple began to reach out to other companies for alternatives.
This work to use Google Gemini is distinct from the effort previously declared to add Gemini like another option for general requests in Apple Intelligence. Today, Chatgpt is the only third party support for this type of general knowledge requests, but Apple plans to add Gemini like another option. This new report, on the other hand, concerns Google in “brain” development based on Gemini to serve as a fundamental model for the new Siri, operating on Apple servers.
