31 Comments
User's avatar
Sam Illingworth's avatar

Fantastic break down and contextual explanation Ilia. For me this is going to be the death knell for xAI. Will also be interesting to see what it does to OpenAI's share prices and how this impacts their capacity to raise the capital they so desperately need to stay afloat...

Ilia Karelin's avatar

That’s actually a great point about the capital, I saw that OpenAI will start putting in the ads which probably not too many people will be happy about. Thank you for kind words Sam!

Joel Salinas's avatar

Oh man Grok and xAI are the epitome of go fast of ignore guardrails

Priank Ravichandar's avatar

Such a great analysis! The decision to integrate Gemini really makes sense, since it allows Apple to focus on building great hardware. I'm sure they’ll ensure that their custom version of Gemini has a distinctly Apple feel. I also think most user queries will be simple enough to run on-device or on Apple servers, since most people aren’t making highly complex requests.

Joel Salinas's avatar

I agree, it rlly makes sense!

Ilia Karelin's avatar

We will still be technically using Siri on the front end, so that’s our Apple feel. And you’re right about the simple user queries!

Daniel Ionescu's avatar

In a way the move is not as surprising. Apple and Google have had a longstanding relationship, from Google Search as default on iPhones ($20 billion per year), Google Maps default navigation (2007-2012) before Apple Maps replaced it, and even the original YouTube iPhone app was made by Apple.

While Apple was always rumoured to develop its own search engine to rival Google (which likely had too much of a head start), the Maps case could be an example of where their AI strategy would follow, and now they’re just buying time with Gemini integration.

Ilia Karelin's avatar

I can see them just buying time with this Google deal. There’s a rumor that they’re still developing their own AI but we will see!

Joel Salinas's avatar

That’s a great point on Google Maps. I do think Apple had true aspirations to create its own ai, I wonder why it took so long for this move

Mark S. Carroll ✅'s avatar

As a decades-long Apple user and a long-suffering Siri skeptic, this piece really landed for me.

What I appreciated most is what you didn’t spend time on. You didn’t re-litigate Siri’s long history of disappointment or Apple’s cultural blind spots. Anyone who’s lived in the ecosystem already knows that story by heart. Using shorthand there was the right call.

Instead, this reads like the moment after all of that became obvious. Apple finally conceding the model layer and choosing to partner rather than pretend they could brute-force their way through it feels like a real inflection point. From my perspective, that does look like a surrender. A pragmatic one, but still a surrender.

I also read this less as “Gemini is faster than ChatGPT” and more as “Google is willing to own the mess that is Siri.” OpenAI was smart not to inherit that product debt. Google, on the other hand, has a once-in-a-generation incentive to try.

My bet is this won’t be smooth. Siri’s problems aren’t just model quality. They’re permissions, context, latency, and a decade of accumulated UX scar tissue. Even a great model can struggle in that wrapper. But if anyone is going to attempt to untangle it at scale, it makes sense that it’s Google.

This felt like an All the President’s Men ending. Abrupt if you’re new. Exactly right if you already know how much came before.

Joel Salinas's avatar

Man, this is such an insightful reply. You made me think about a lot of things through this. I agree it won't be easy, but it is a wise move. To me, it also shows that there is a growing lack of trust in either the management or the feasibility or the long-term status or the governance of OpenAI. I'm sure Apple would have just gone with the best long-term partner, and clearly they thought that was Google over OpenAI. I also am a heavy Apple and Mac user, so on one side I'm really excited and hoping for the best here.

Mark S. Carroll ✅'s avatar

Joel, 100% with you. This reads less like “Gemini is better” and more like “Apple wants a partner it can control without inheriting the mess.”

The real contest is not model IQ. It is governance and who can be wrong in public without melting down. Apple hedging is very Apple. Optionality over romance.

Also, if Apple ships a new assistant that confidently corrects users while rationalizing its own mistakes, that will be the most on-brand feature ever 😄

Joel Salinas's avatar

Hahah that last line was great

Ilia Karelin's avatar

Thank you for such insightful comment Mark! I did read that Apple is still trying to make their own model, so we’ll see if it will ever come to fruition.

And about “Siri’s long history of disappointment or Apple’s cultural blind spots.” - I actually did have it in a draft, but I thought that people might already know about all of that lol

Mark S. Carroll ✅'s avatar

If it ever actually comes to fruition then it will tell users they’re wrong most of the time while rationalizing every misstep Apple makes 🤪

Michael Meneghini, MD's avatar

Apple’s move shows strategy beats ego: partner where strength exists, not where hype dominates

Joel Salinas's avatar

Well said, Michael!

Ilia Karelin's avatar

Agreed. They’re the best at hardware, and combine that with the power of Gemini - hopefully we’ll see a much more useful version of Siri soon!

Odin's Eye's avatar

Thanks. Huge move by Apple. Fantastic analysis and analogy

Joel Salinas's avatar

@Ilia Karelin did well!

Ilia Karelin's avatar

Appreciate you reading it!

Conspiracy Tourist's avatar

I call him Harry, first Harry said, that is not my name, later Harry smiled and now hé is oké with it✌️.

Alex Randall Kittredge's avatar

Two of the biggest encumbants align for the future. Is the AI race finally over? Did google and apple just win?

Joel Salinas's avatar

I do think OpenAI is completely freaking out right now, maybe calling Blackberry? hah

Ilia Karelin's avatar

I don’t think so, I think it’s still unfolding, and the race is still in play. Can’t forget about OpenAI and what they’re going to do. We got also Anthropic, will see what they will do!

Dennis Berry's avatar

For anyone using Apple products, this could finally make Siri something you actually want to use.

Joel Salinas's avatar

Hoping so!

Ilia Karelin's avatar

Potentially yes! I’m very curious how it’ll all work!

John Brewton's avatar

Knowing what not to build is often the smartest product decision.

Joel Salinas's avatar

Oh 100%!!

Ilia Karelin's avatar

Potentially yes! There’s a rumor that I saw that Apple is still building their own AI, but we will see if it will come to fruition or not!