Apple is reportedly just weeks away from revealing the outcome of its expanded collaboration with Google’s Gemini team, which will power key aspects of Apple Intelligence and the revamped Siri. According to Bloomberg News’ Mark Gurman, the company has been preparing to unveil the upgraded Siri in the second half of February, complete with live demonstrations of its new capabilities.
Mark Gurman for Bloomberg News:
Whether that takes the form of a major event or a smaller, tightly controlled briefing — perhaps at Apple’s New York media loft — remains unclear. Either way, Apple is just weeks away from finally delivering on the Siri promises made at its Worldwide Developers Conference back in June 2024. At long last, the assistant should be able to tap into personal data and on-screen content to fulfill tasks.
To introduce those features in iOS 26.4 — scheduled to enter beta testing next month and roll out publicly in March or early April — Apple needed Gemini. Internally, the company has labeled the technology Apple Foundation Models version 10, making it seem entirely homegrown. It runs at roughly 1.2 trillion parameters — a measure of AI complexity — and is hosted on Apple’s Private Cloud Compute servers.
But that’s only the first phase. As I reported last week, Apple plans to unveil a fully reimagined Siri at this year’s WWDC. Code-named Campos, the new system is a fresh architecture and interface designed from the ground up for the chatbot era. It will debut in iOS 27, iPadOS 27 and macOS 27, which arrive in beta form this summer.
This Siri will be conversational, aware of relevant context and capable of sustained back-and-forth dialogue. Essentially, it matches what users already expect from ChatGPT, Google Gemini and Microsoft Corp.’s Copilot. It, too, will rely on Gemini, but a far more advanced version internally known as Apple Foundation Models version 11. The model is expected to be competitive with Gemini 3 and significantly more capable than one supporting the iOS 26.4 Siri.
To improve accuracy and responsiveness, the two companies are discussing running this version of Siri directly on Google’s cloud infrastructure and its high-powered tensor processing units, or TPUs, rather than Apple’s own servers.
The shift to Gemini represents a fundamental reset of Apple’s AI strategy under Federighi, who effectively assumed control of the company’s AI direction early last year — around the time I reported that Chief Executive Officer Tim Cook had lost confidence in Giannandrea. Federighi concluded that, at least for now, Apple would be better served partnering with Google than relying solely on its internal models. It wasn’t an ideal outcome, but it was the only viable one.
MacDailyNews Take: Apple Intelligence. Powered by Google
.
We don’t want to use or be dependent on fscking Google, Apple. We want to be able to choose our LLMs. We don’t use Google Gemini for myriad reasons: privacy; security; we’d rather use the best LLMs for each job. If you are incapable of creating your own, which Apple clearly is, then give users choice.
Apple would not be in this position if it wasn’t for far too long saddled with an operations manager masquerading as a CEO, but instead was led by a visionary CEO who wasn’t chasing fads (VR goggles) and failing at moonshots (EVs), but was instead skating to where the puck was going.
Tim Cook criminally neglected Siri for his entire iterative, beige, boring, seemingly unending tenure.
Now, Tim Cook’s “solution” is for Apple to stoop to paying Google, of all companies, to fix it.
As we wrote last August:
Google Gemini? Why not just get a Samsung Galaxy phone which already integrates Google’s Gemini AI as a core component of their AI-powered features?
Google Gemini on an iPhone offers precious little differentiation from Samsung, the chief iPhone knockoff peddler.
If you’re going to with an external AI partner, why not choose the smartest one? We find xAI’s Grok to be more accurate and useful than Google’s Gemini, ChatGPT, and the rest. – MacDailyNews, July 21, 2025
Apple would likely need a CEO who is more open to thinking outside the box than the one with which it’s currently saddled.
Google Gemini. Puleeze. OpenAI’s ChapGPT or Anthropic’s Claude would be better choices. Even better would be for Apple to allow users to choose – gasp! – which AI model they’d like to underpin Siri.
Regardless of what Apple chooses, they will at least be safely behind Apple’s privacy wall.
The issues are: Google’s Gemini is not the best and everyone knows it, Google has a poor reputation for privacy that will tarnish Apple’s, and Google, hello, ripped off the iPhone with Android. Enough with the Google, Apple!
Please help support MacDailyNews — and enjoy subscriber-only articles, comments, chat, and more — by subscribing to our Substack: macdailynews.substack.com. Thank you!
Support MacDailyNews at no extra cost to you by using this link to shop at Amazon.
The post Apple said to unveil new Google Gemini-powered Siri in mid-late February appeared first on MacDailyNews.
No comments:
Post a Comment