Apple’s new Siri arrives as ChatGPT threat

new AI-powered – Apple said its AI-powered Siri will finally reach iPhone users this fall, shifting the smartphone assistant from long-promised feature to everyday alternative for many users—especially the free tier of ChatGPT. The change also lands with sharper commercial sta

Apple spent 2024 telling the world it had created its own take on AI. The plan was simple: fold it into Siri, the iPhone virtual assistant that, until now, “has never really delivered on its promise.”

But Apple didn’t just miss expectations—it also ran into a wall. It became unable to deliver the new Siri.

Then came the reset. At its annual developers conference this week, Apple said the new, AI-powered Siri is actually ready. iPhone users will finally see it this fall.

The positioning is careful. Apple says the new Siri isn’t meant to compete with the most sophisticated versions of AI engines like ChatGPT or Claude. It also won’t replace the idea of “vibe-code a new business.”

Yet for many people, the real impact is more straightforward. The sum total of Apple’s announcements, as Nilay Patel puts it, is “a replacement for the free tier of ChatGPT.”

Patel’s comparison is built from Apple’s own history of shipping features by absorbing what users already liked elsewhere. He points to the “Sherlock” era—an app Apple built called Sherlock that “just destroyed” an earlier Watson tool. He says Apple’s pattern is that each year it looks through the ecosystem of third-party apps and turns the best parts into built-in capabilities.

In this latest move. Patel argues Apple has “nakedly done” the Siri equivalent of that same approach: “Sherlock the free version of ChatGPT.” His core question is blunt: why use free ChatGPT when Siri is “right here behind a button on your phone” and. for many everyday tasks. can do “all the same stuff?”.

Apple’s pivot also raises a bigger question about what changed since 2024. Two years ago. Patel says. Apple’s approach to AI seemed to be. “We don’t make our own AI models. but we’ll let you use everyone else’s.” The idea mirrored how Apple often works—watch what happens in the broader market and then build a product that wins once the pieces are proven.

What didn’t happen was what Apple may have hoped would stay in the future. Patel says the models kept getting more capable. and the product turned out to be “writing code for you.” He adds that. as far as he can tell. there isn’t a great consumer AI pendant yet that threatens the iPhone—phones still dominate because consumer AI still doesn’t fully replace what users already do on their devices.

That’s where the economics and the strategy collide. Patel argues Apple didn’t want to make its own chatbot because it would rather let others build the best models and bots first. letting users use their iPhones to access them. In his view. the longer-term risk is that cloud-based agents do more and disintermediate the app model—potentially even disintermediating Apple services revenue. which he describes as a “big line of revenue” that drives much of the business.

If chatbots can operate in the cloud—placing an Uber. arranging a DoorDash order. or shopping for you—then the phone is less essential and Apple’s services exposure could shrink. Patel is also direct about the human reality of the device ecosystem: people may talk about “glasses and pendants. ” but “right now it’s the phone.” He says he uses Gmail on his iPhone. Google Maps instead of Apple Maps. and Spotify instead of Apple Music. In his view. Apple can prefer users to switch to Apple’s versions of those tools. but it doesn’t have to force it immediately—because buying power anchored to the hardware still matters.

What Apple can’t assume, Patel says, is that the phone remains dominant for another 10 or 15 years. If users shift from tapping apps to “talk[ing] to an agent” that can “go get things done. ” that could lead to new hardware form factors. He references Jony Ive’s effort at OpenAI as one example of where that could go.

The device question ties directly to why Apple is now answering the competitive pressure. Patel frames Apple’s version of AI as protection for the idea that “you can talk to the phone.” His scenario is simple: if users aren’t using Siri. they might use ChatGPT. and if they use ChatGPT—or Gemini—enough. then later they could be targeted with hardware changes that reduce Apple’s relevance.

In that picture, he says the pitch could become, “Buy a point-and-shoot camera and our little pendant, and you’ll be set,” paired with the promise that “you won’t have social media to distract your kids.”

There’s also the commercial fracture line between Apple and OpenAI. Patel points out that two years ago, OpenAI was Apple’s chosen partner, and now Apple appears to be competing with it. He also notes that OpenAI is considering suing Apple.

For Patel, that tells a story about organizational style and ambitions. He calls OpenAI “vastly more chaotic. ” points to a “messy breakup with Microsoft. ” and describes the company as trying to become “the big consumer company of the next generation.” He says OpenAI is building its own products and trying to do a lot at once without what he sees as enough focus.

That uncertainty may matter to the enterprise. If a customer depends on OpenAI’s services, Patel says they may start wondering whether the company will survive to an IPO, whether Sam Altman will stay CEO “forever,” and whether executives could be “fired tomorrow.”

So what is really happening in the relationship?. Patel questions whether Apple is seeking “more control over AI” or whether it simply wants to “work less with OpenAI.” His answer leans toward Apple’s DNA—summarized as the “Cook Doctrine.” He recalls that Tim Cook told shareholders many years ago: “We want to own and control the primary technologies that we base our products on.”.

In this specific case. Patel says Apple may not own Gemini. and it may not invest in making that kind of acquisition. But he argues Apple does own elements that support the AI stack: private cloud compute. “the code that’s running on those servers in Google Cloud. ” and “the Nvidia GPUs and how they’re being used for this stuff.”.

Patel sees that as a strategic goal with teeth: OpenAI ultimately wants to be a competitor to Apple. not a “frenemy relationship.” He also says OpenAI isn’t yet positioned to sustain the kind of structured rivalry that other big tech firms maintain. noting that there’s no Instagram without the iPhone and “there’s no iPhone without Instagram. ” so the two sides keep distance.

Patel’s final point is about maturity. Other companies, he says, may want to become enormous rivals, but they haven’t built the skill set—or the “thick enough skin”—to pull it off.

Apple’s promise to deliver a truly AI-powered Siri this fall turns a long-running product gap into a direct competitive moment. For millions of iPhone owners, the question won’t be about model benchmarks or Claude-level capabilities. It will be simpler: whether Siri. newly retooled. can replace the free ChatGPT experience from the button they already use every day.

Apple Siri artificial intelligence ChatGPT Claude OpenAI iPhone developers conference Tim Cook doctrine Nvidia GPUs cloud compute