A new report from Mark Gurman at Bloomberg tells us that Apple is working with Google to check out a customized model of Gemini that would energy Apple’s next-gen Siri voice assistant within the cloud.
The report says Apple is in “early discussions” and that Google is presently coaching a model of its Gemini AI assistant that may run on Apple’s cloud infrastructure. That is just like the partnerships Apple has explored with Anthropic’s Claude or OpenAI’s ChatGPT. Apple is making a next-gen, “LLM Siri” utilizing its personal fashions whereas on the identical time creating an infrastructure that would see the brand new Siri powered by a third-party AI mannequin. We wrote about that effort right here.
In accordance with Gurman’s report, Apple remains to be “a number of weeks away” from deciding whether or not to go together with Apple’s personal inside fashions for the brand new Siri or to make use of tech from exterior companions.
Apple is using a “bake-off” method, pitting the internally-developed next-gen Siri (code-named Linwood) towards a next-gen Siri powered by exterior fashions (code-named Glenwood). The corporate has not but determined which exterior expertise is greatest, nor has come to monetary phrases with them. Supposedly, Anthropic’s Claude was the early favourite, however Anthropic demanded such a steep worth that Apple began reaching out to different firms for options.
This work to make use of Google’s Gemini is separate from the previously-reported effort so as to add Gemini as an alternative choice for basic queries in Apple Intelligence. Right now, ChatGPT is the one supported third-party AI for these kinds of basic information queries, however Apple has plans so as to add Gemini as an alternative choice. This new report, in contrast, is about Google creating Gemini-based “brains” to function the foundational mannequin for the brand new Siri, working on Apple’s servers.