待翻译：Apple and AI: Siri the Early Years

AI 服务暂时不可用，以下为来源正文，待恢复后补全翻译。

Babbage Aug 13, 2026 One of the reasons for studying technology history is to look for recurring patterns: characteristics of either the technology itself or of firms that are built around it that are repeated time and time again. These can be particularly useful as predictors of future behaviour at a firm when that firm has strong DNA that is passed down from one generation of management to the next. It’s generally agreed that Apple is one firm with a track record of acting to preserve and strengthen its DNA. The appointment of 25-year Apple veteran John Ternus to succeed Tim Cook as Apple CEO is just the latest and most prominent example. Which brings us to the subject of this post and the next: Apple and AI. Spoiler alert! These posts suggest that there is a credible scenario where Apple is a huge winner - possibly even the biggest winner - in the ‘AI revolution’. I want to say upfront that there is too much uncertainty about the final shape of the technology underpinning AI to come to any firm conclusions. I’d also like to acknowledge the post Apple is the King of AI and Nobody Knows It by Limited Edition Jonathan which prompted me to think about Apple and AI. I want to emphasise though that I’ll be considering Apple’s position from a very different perspective and that, in the end, I reach conclusions that, although aligned in some ways, are distinctive in others. The core of the argument is that Apple has a track record of scaling the production of leading edge technology to make it available to its millions of customers ahead of its competition. This has been accompanied by long term investment across the whole technology stack, from custom silicon, through compilers and application software. These capabilities, and a willingness to use them, are central features of Apple’s DNA. There is a lot more to the say about this though! We’ll go into (much) more of the detail that supports these arguments in the next post. In the rest of this post we’ll take stock of where Apple is now and its AI journey so far. Siri Arrives: The Dream and The Reality … for decades technologists have teased us with this dream that you're going to be able to talk to technology and it'll do things for us. Haven't we seen this before over and over, but it never comes true. We have very limited capability you just learn a syntax: call a name, dial a number, play a song. It is such a let-down. What we really want to do is just talk to our device ask a simple question: what's the weather going to be like today and get a response. In fact we don't want to be told how to talk to it, we want to talk to it any way we'd like. Someone else might ask ‘what rain Cupertino’ or ‘is the weather going to get worse today’ or ‘do I need an umbrella today’ and your device, in this case your phone, will figure out what you mean and help you get what you want done. Phil Schiller, Apple SVP Marketing at Apple Keynote that launched Siri What year and with which phone did Apple launch Siri? Don’t cheat! The answer is below after the short break. The answer is 2011 with the iPhone 4S at an October special event. Yes, it’s nearly fifteen years since Siri made its first appearance. It was a momentous few days in Apple’s history. The event was Tim Cook’s first as CEO. Steve Jobs would pass away the very next day. Here is a link to that Siri introduction: Phil Schiller’s opening words neatly set out the promise of Siri: an assistant that you could talk to in natural language. If we didn’t get the message from the introduction then Scott Forstall’s live Siri demonstration reinforced the point, in Forstall’s own words (my emphasis): … not only did it understand the words I said it understands the meaning and then goes and gives me this weather forecast … the exact words I say aren’t important it’s the meaning behind the words so I could ask this in a completely different way … After asking for the weather today Forstall then goes on to ask if he ‘needs a raincoat today’, to which Siri responds ‘It sure looks like rain today’ followed by the weather forecast. Perhaps the cloudy outlook was a metaphor for Siri’s prospects? Readers who used Siri in the decade after it was launched will recognise the acute irony in Schiller’s opening words: Apple was continuing the long tradition of teasing us with a dream. The reality was that Siri would consistently fall short of Schiller and Forstall’s promise. We’ll come to this again in a moment. Put this all aside for now though and consider again how early this is and the prominence that Apple’s management gave to the launch of Siri. The iPhone 4S, also launched at this keynote, used a dual core 32-bit Arm Cortex A9 processor with 512MB of DRAM. This was nowhere near powerful enough to even attempt the voice recognition part of Siri, let alone ‘understand the meaning’ of what the user was saying. So the voice data was sent to Apple’s (probably rented) servers in the cloud. Siri’s Origins Siri started its life at the non-profit research institute SRI International - hence the ‘SiRI’ name - before being spun off into into start-up Siri Inc. The pre-Apple story of Siri is neatly summarised on the SRI website: Siri, the first virtual personal assistant, arose from decades of artificial intelligence (AI) research at SRl. The technology was developed through the SRI-led Cognitive Assistant that Learns and Organizes (CALO) project within DARPA’s Personalized Assistant that Learns (PAL) program – the largest-known AI project in U.S. history – and joint work with EPFL, the Swiss Institute of Technology. The Computer History Museum has a fascinating interview with Siri Inc’s founders which positions Siri as the first Agentic AI tool: There’s no topic as white hot in today’s tech press as “agentic” AI. That’s a newish word for what used to be called intelligent agents or intelligent assistants, and from Salesforce to Microsoft companies are investing billions. “Agentic” means AI designed to do things for us, rather than simply paint pictures or answer questions like MidJourney or ChatGPT. Think “order me the best Pad Thai you can get delivered here in 40 minutes” or “build us a factory nearby.” When it comes to shooting the breeze, Siri is not nearly as fluent as modern chatbots. But unlike them, she was designed to get things done, and only a portion of her abilities were fully deployed. What can we learn from a long-secret interview with the team that created Siri, the first wildly successful intelligent assistant, and recent follow-up interviews with both the technical architect and visionary main investors? The full interview with the founders of Siri can we watched here: The, interview, filmed in 2013 was embargoed for a decade until it was finally released in 2023 (and at the time I started writing this post still had only 66 views!) SRI would move to commercialise that research in 2007: SRI spun off Siri, Inc. in 2007 to bring the technology to consumers, raising $24 million in two rounds of financing. Siri in Apple Siri Inc. first came to Apple’s attention when they launched a Siri App on the iPhone App Store. Within weeks Steve Jobs was inviting Siri Inc CEO Dag Kittlaus to his home and the deal to buy Siri was completed within two months. The launch of the ‘integrated Siri’ in 2011 wasn’t the end of Apple’s work on Siri and notable enhancements followed at regular intervals over the next decade. WWDC 2014 : Added ‘Hey Siri’ hands free activation and Shazam music identification WWDC 2016 : Siri added to the Mac in macOS Sierra. WWDC 2018 : Siri shortcuts, allowing custom multi-step macros, introduced. WWDC 2021 : On device speech processing. Worthwhile improvements, certainly, but nothing that would bring Siri close to fulfilling the ultimate promise that Schiller and Forstall had held up in that first Siri announcement. The ChatGPT Moment If Siri’s weaknesses had been apparent over the decade after it launched, then they were thrown into sharp relief in November 2022 when OpenAI launched ChatGPT based on GPT 3.5. This was world-changing technology and squarely in the territory that Siri should have made its own. Apple had been caught flat-footed and needed to respond. In July 2023 Bloomberg’s Mark Gurman shared that: Apple Inc. is quietly working on artificial intelligence tools that could challenge those of OpenAI Inc., Alphabet Inc.’s Google and others, but the company has yet to devise a clear strategy for releasing the technology to consumers. The iPhone maker has built its own framework to create large language models — the AI-based systems at the heart of new offerings like ChatGPT and Google’s Bard — according to people with knowledge of the efforts. With that foundation, known as “Ajax,” Apple also has created a chatbot service that some engineers call “Apple GPT.” and in an update later that year: “There’s a lot of anxiety about this and it’s considered a pretty big miss internally,” a person with knowledge of the matter told Power On. There can be no doubt that Apple took ChatGPT very seriously indeed. These efforts culminated in the launch of Apple Intelligence at WWDC in June 2024. The press release made the promise clear: Apple today introduced Apple Intelligence, the personal intelligence system for iPhone, iPad, and Mac that combines the power of generative models with personal context to deliver intelligence that’s incredibly useful and relevant. Apple Intelligence is deeply integrated into iOS 18, iPadOS 18, and macOS Sequoia. It harnesses the power of Apple silicon to understand and create language and images, take action across apps, and draw from personal context to simplify and accelerate everyday tasks. Here is the WWDC presentation: And Apple was confident enough to include Apple Intelligence in TV ads for the iPhone 16: Except is was a false dawn again. In March 2025 the firm announced that these features would not be ready on the planned schedule: We’ve also been working on a more personalized Siri, giving it more awareness of your personal context, as well as the ability to take action for you within and across your apps. It’s going to take us longer than we thought to deliver on these features and we anticipate rolling them out in the coming year. In fact the announcement understated the poor state of development of Apple Intelligence. The features were nowhere near ready to ship. John Gruber in Something is Rotten in the State of Cupertino was annoyed furious not about Apple lagging on generative but on kidding everyone - including apparently itself - that it was on track to ship Apple Intelligence. The fiasco here is not that Apple is late on AI. It’s also not that they had to announce an embarrassing delay on promised features last week. Those are problems, not fiascos, and problems happen. They’re inevitable. Leaders prove their mettle and create their legacies not by how they deal with successes but by how they deal with — how they acknowledge, understand, adapt, and solve — problems. The fiasco is that Apple pitched a story that wasn’t true, one that some people within the company surely understood wasn’t true, and they set a course based on that. Siri AI Apple needed help. It found it in the shape of Google. When Siri AI was announced in June 2026 the firm acknowledged Google’s role: This year, we embarked on a deep collaboration with Google, leveraging the technologies behind their Gemini family of models. Together, we created the next generation of Apple Foundation models for our integrated Apple Intelligence experiences, and adapted these new models to run on-device and on servers using Private Cloud Compute. This time Apple was making sure that everyone could see that the features demonstrated were running on an actual iPhone. As an aside I think it’s interesting to note that Federighi stated that the collaboration with Google started in 2026 when th [truncated for AI cost control]