We’ve now got official confirmation that Apple and Google are working together on what I’ve been calling the “Siri brain transplant.” Instead of an Apple LLM technology, it appears they’ll be using some version of Gemini. There are many unanswered questions here.

For instance, is the Gemini-based model going to be both local and in Apple’s Private Cloud Compute space? I like the idea of a Gemini-caliber PCC engine I can access privately from my Apple devices.

The original sin here isn’t Apple’s failure to build its own LLM over the last few years, but rather its lack of attention toward Siri throughout its lifespan. Anybody who cares about this stuff has experienced frustration as Siri never evolved the way it should have. Indeed, Siri has been the butt of popular culture jokes for at least a decade.

I’ll always believe that if Apple made Siri a priority ten years ago, they wouldn’t be in this position today.

Nevertheless, all of this feels more like a black eye than a knockout. There is still time for Apple to evolve and we’re all still figuring out what LLM-based AI means for all of us.

If I were going to look for a silver lining on this, at least Apple had the maturity to swallow their pride and go get help.

But let’s not forget this is all still just news. I still haven’t seen Siri work the way I’d have expected in 2026.