The Collapse of Apple's AI: From Jobs' Vision to the Dilemma of Executive Missteps

Author: Moonshot

AI has been hot for almost three years now.

Major tech giants are racing to invest in this field, yet ironically, among this frenzy, Apple, which is closest to our lives, seems to be the farthest from AI.

The biggest giant seems to have become invisible in the face of the hottest trends.

At WWDC last June, Apple slowly released Apple Intelligence, but now nearly a year has passed, and for most users, Apple Intelligence remains more of a rumor than a reality.

The whole world sees that Apple’s AI is not doing well, but no one knows what exactly happened.

Famous Apple analyst Mark Gurman has just released a lengthy article in the foreign media titled “Why Apple Still Hasn’t Cracked AI,” revealing the internal ambivalence towards AI at Apple, the internal struggles, and the technological bottlenecks that are difficult to overcome.

It is worth noting that Gurman used the phrase “Still hasn’t”, which has already set the tone for Apple’s current situation.

This article will reorganize the original text to present Apple’s history, current status, root problems, and future challenges in the field of AI, analyzing why Apple has struggled in the AI arena, making AI its Achilles’ heel.

Siri from 14 years ago, the concept is already a large model.

On October 4, 2011, the day before Steve Jobs passed away, he left behind the legacy of Siri for Apple.

At that time, Siri was like a product from science fiction coming to life; it could understand users’ voice commands, make restaurant reservations, find movie theaters, or hail a taxi. Apple once again turned futuristic technology concepts into mainstream products.

Siri was released alongside the legendary device iPhone 4s | Image source: Apple

At that time, Jobs was not very interested in building a search engine.

One person who worked with him said, “Jobs didn’t trust users to actively search for things, and he thought Apple’s job was to carefully select and show users what they wanted.” This idea, like many of Jobs’s beliefs, continues to have a profound impact on Apple even after his death.

When Jobs first encountered Siri, which was then an application in the App Store, he was immediately captivated.

Dag Kittlaus, co-founder of Siri, recalled that the ultimate goal of Siri was “you can talk to the internet, and the assistant will handle everything for you. You don’t even need to know the source of the information; the problems with applications and websites will also be resolved.” This is exactly the most widespread application scenario of today’s large language models.

Steve Jobs immediately realized that Siri was much more than just an application. He quickly contacted Kittlaus and invited the Siri team to his home for a meeting. During the three-hour meeting, Jobs proposed to acquire their company. Kittlaus initially rejected the acquisition offer, but couldn’t withstand Jobs’ persistent calls persuading him for 24 consecutive days.

Eventually, Kittlaus agreed to sell Siri, and Jobs immediately listed Siri as one of Apple’s top development projects, fully investing in its research and development even during the end of his life.

At that time, Siri led the smart voice assistant market, but a few years later, competitors like Google, Amazon, and Xiaomi launched more advanced voice assistants and smart speakers, while Siri showed no significant improvement.

With the launch of Siri, Apple also began its research into machine learning, primarily for facial and fingerprint recognition, smart suggestions (such as reminding you when to leave based on traffic conditions), improving maps, and the then-focus projects: headsets and cars.

Use Siri to check the weather | Image source: Apple

In the initial years, the development of Siri focused solely on basic tasks, such as providing weather information, setting timers, playing music, and handling text messages.

Looking at the actions, Apple has long been laying out its plans in the AI industry. They have acquired several small AI companies, including machine learning companies Laserlike, Tuplejump, and Turi.

According to sources familiar with the matter, Apple is even considering acquiring Mobileye Global Inc. for about $4 billion, which could be Apple’s largest acquisition ever.

Mobileye focuses on the development of advanced autonomous driving systems and computer vision technology. However, Apple ultimately abandoned the deal, and in 2017, Intel acquired Mobileye for $15 billion.

So looking at the direction, Apple has not invested AI in the voice assistant field.

At that time, OpenAI, which had only been established for half a year, claimed it would create a “general” robot. But Siri was still, as it was at the time of its acquisition, setting alarms, checking the weather, and playing music on countless iPhones.

02 Expectations, Infighting, Exclusion

With money, status, and strength, why did Apple stumble in AI? Perhaps only veteran Apple analyst Mark Gurman has access to such insider information, as he extensively reported on the internal conflicts at Apple regarding AI.

In 2018, Apple poached John Giannandrea (referred to in the industry as JG) from Google to be the head of AI.

Apple AI head John Giannandrea | Image source: Apple

JG was previously in charge of Google’s Search and AI departments, leading the team in deploying AI technology in products such as Google Photos, Translate, and Gmail.

“JG is not only regarded as the most influential executive after the CEO at Google, but he also served as the Chief Technology Officer at the internet pioneer Netscape.” A person involved in the hiring process stated, “Can anyone be found who is more suitable than him?”

For Apple, mining JG not only deals a heavy blow to its competitor Google, but it is also expected to be the first step in transforming Apple into an AI leader.

In the press release published by Apple at that time, Cook stated, “Machine learning and AI are crucial to the future development of Apple. They will not only fundamentally change the way people interact with technology but have already brought significant improvements to users’ lives. We feel very fortunate to be able to work alongside John. As a leader in the field of AI, he will surely be able to drive our substantial progress in this key area.”

John Giannandrea joined Apple in 2018 to lead AI and machine learning | Image Source: Apple

However, looking back seven years later, this expectation and optimism have completely disappeared. Apple’s AI has not only failed to improve but has fallen even further behind.

The core issue is:

Executives have differing understandings of the AI roadmap.

Some senior executives responsible for software engineering believe that Apple should emphasize AI more in iOS. Around 2014, one executive stated, “We quickly realized that this was a revolutionary technology, far more powerful than we initially understood.” However, they could not convince Craig Federighi, who oversees iOS, to take AI seriously, saying that “many suggestions fell on deaf ears.”

However, Cook is unexpectedly optimistic about AI, a person who once worked with him said: “Cook is one of the most optimistic people about AI at Apple. He has always been frustrated that Siri is lagging behind Alexa. He is also unhappy that Apple has not secured a place in the smart speaker market.”

The person in charge, JG, has been wavering in his judgment about AI.

When JG joined Apple in 2018, other executives recalled, citing Apple’s closed software ecosystem as a unique advantage in instantly deploying the latest features to billions of devices.

But JG quickly discovered that Apple needed to invest hundreds of millions of dollars for large-scale testing and image and text annotation to train the large model. JG recruited top AI researchers from Google and assembled a team responsible for testing and data analysis. Afterwards, JG turned its attention to Siri, replaced the person in charge, and proposed cutting features that were rarely used by Siri.

However, JG’s efforts are often hindered. According to several colleagues, software head Craig Federighi is reluctant to invest heavily in AI, as he does not believe that AI is a core capability of mobile devices.

Craig Federighi is a familiar face at the press conference, having always been in charge of Apple’s software team | Image source: Apple

A former senior executive who worked at Apple for a long time said: “Craig is not the kind of guy who says, ‘We need to do something big, we need more budget and people.’”

Other leaders also have similar reservations. An old executive said, “In the AI field, you have to invest first to know what the product is. This is not how Apple operates. When Apple develops a product, they already know what the end goal is… Our usual strategy is to enter late, relying on over a billion users, and steadily defeat everyone in the end.”

But this strategy does not work for AI. The release of ChatGPT in November 2022 caught Apple off guard.

A senior executive stated that prior to this, Apple “didn’t even have the concept of Apple Intelligence.”

Another executive said, “OpenAI’s actions are not a secret; anyone paying attention to the market should see it and invest fully in it.”

Less than a month after the release of ChatGPT, Craig Federighi began using generative AI to write code for software projects. According to insiders, he suddenly saw the potential of AI and united with JG and other executives to start meeting with AI companies like OpenAI and Anthropic, rushing to learn about the latest models and market trends. From that time on, Craig Federighi requested that as many AI features as possible be incorporated into iOS 18 in 2024.

JG has started to assemble an AI team to develop large language models again, at which point they have already fallen behind their competitors for many years.

According to a number of employees, Apple’s various internal product development teams are responsible for part of the AI research and development, which makes it difficult to unify technology, progress, and compatibility.

At the WWDC in 2024, Apple Intelligence arrived as expected, but its capabilities were somewhat limited.

Apple’s generative AI capabilities are indeed lagging behind. According to a report by Bloomberg, there is an internal chatbot at Apple that can handle basic image generation, but it is at least 25% behind ChatGPT and significantly lacks accuracy in handling most queries.

In order to provide consumers with the AI products they truly want, Apple had no choice but to negotiate with competitors such as Google, Anthropic, and OpenAI.

This has caused internal divisions again, with JG advocating for the adoption of Google’s Gemini, arguing that OpenAI’s protection of personal data and sustainable development is inferior to Google’s. However, Apple’s corporate development team disagrees; they are optimistic about OpenAI. Therefore, it was announced at WWDC that requests that Siri cannot handle will be forwarded to ChatGPT.

The iPhone 16, which focuses on Apple Intelligence, has yet to "deliver"｜Image source: Apple

Apple’s lack of self-developed AI chatbots has made many executives uneasy, but JG does not believe that large models are the future of AI.

According to multiple employees, JG believes that it will be many years before AI Agents can truly replace humans, and that most consumers, like him, do not trust generative AI.

These employees said this explains why JG has not fully committed to developing a consumer-facing ChatGPT competitor. They revealed that JG had stated consumers do not want tools like ChatGPT.

Inside Apple, JG has borne most of the blame for delays and mistakes. According to several employees, JG has struggled to integrate into Apple’s core executive circle, “Those executives have worked together for decades, operating the company like a family business.”

JG’s position is quite awkward; he is an external hire and finds it difficult to push for core changes at Apple. According to an employee familiar with him, “JG should be more proactive in securing large amounts of funding, but he is not a salesperson; he is a technical expert.”

Some people also say that JG is not hands-on enough and has not set strict requirements for employees. One executive said, “Other engineering teams at Apple are fully dedicated and deliver on time, while JG’s team is not like this; they lack execution capability.” This “lenient” impression also extends to benefits.

Unlike other Silicon Valley giants, Apple headquarters employees have to pay for their cafeteria meals. However, during the rush period at Apple Intelligence, some engineers in JG often received free meal vouchers, causing dissatisfaction among other teams. One employee said, “Apple doesn’t provide free meals, but their team delivered a year later than others and still gets free lunch.” These subtle differences in treatment have led to divisions among different teams within Apple.

Left: Cook Center: JG Right: Craig Federighi | Image source: Bloomberg

The so-called lack of urgency from JG may not only be a personality issue but also a philosophical consideration regarding the understanding of AI.

He has always been conservative about the pace of AI development, doubting the value of chatbots, and he doesn’t see a threat from rivals like OpenAI, Meta, Google, etc.

Some Apple colleagues have stated that JG firmly believes that the AI assistant users want is the main interface of the device, rather than a specific application. Despite facing delays and setbacks, he remains committed to this vision.

In March of this year, JG was stripped of all product development control, including Siri and robotics projects. According to other executives, Cook lost confidence in JG’s ability to execute on new product development.

JG retains oversight of AI, large language model development, AI analytics, and parts of the development team. According to insiders, some executives have discussed the idea of reducing JG’s responsibilities or retiring him gradually, but Craig Federighi and others are concerned that if JG leaves, the top researchers and engineers he brought with him may also leave.

Currently, JG has chosen to stay; he told his colleagues that he does not want to leave before Apple AI is on the right track. He also admitted that not being responsible for Siri has relieved him.

03 Technological Backwardness, The Crossroads of Privacy

The mistakes of tech giants cannot be simply attributed to one person, JG.

The marketing and advertising teams promoted unfinished features in advance, Craig Federighi is the final decision-maker for the software projects, and Cook sets the product development culture for the entire company.

Even the former CFO was overly conservative when purchasing GPUs, and Apple did not leverage its market dominance and cash reserves, instead opting to gradually procure AI hardware as usual. As a result, most GPUs were snapped up by competitors like Amazon and Microsoft, leading to slower training speeds for Apple’s AI models.

According to Apple and other company executives, the number of AI employees at Apple is far less than that of its competitors, and the GPUs procured for training and running large language models are also fewer.

For Apple, missing out on potential disruptive technologies is not fatal. After all, Apple often lets its peers explore new technologies to validate the market first, and then refines its products before launching well-designed and more user-friendly versions to consumers.

This strategy has also shaped users’ impression of Apple as “not seeking the latest, but seeking the best.” Over the years, Apple has become the world’s most valuable technology company through its meticulously crafted products, curated content, and annual software updates.

Apple thinks so too. During the earnings call in May, when Cook was asked by shareholders about the delay in AI, he stated that it just needs more time to meet Apple’s quality standards, “there’s not much else to it, it’s just taken a bit longer than we expected.”

But the question is, how long is considered long? Apple has invested a large amount of resources in AI but has seen little return. Moreover, AI is a faster and more complex technology.

Looking back at history, Apple’s most successful products have been built on core technologies developed internally, such as the multi-touch of the iPhone and the self-developed M series chips of the Mac. However, when it comes to AI, it is difficult to see Apple’s technological strength.

Apple’s M series self-developed chips have fully penetrated into the hardware line | Image source: Figma

Moreover, Apple has a technological bottleneck in AI that other giants do not have: data utilization.

For many years, Apple has positioned user privacy protection as a selling point, but now it has become an obstacle to their AI development.

With 2.35 billion active devices, Apple holds web search, user habits, and communication data that surpasses many competitors. However, Apple’s restrictions on AI developers’ access to data are much stricter than those of Google, Meta, and OpenAI, which forces Apple’s researchers to rely on third-party licensed datasets and synthetic data (artificial data created specifically for training AI).

A person familiar with Apple AI and software development said: “Apple has countless objections to every move in AI; you have to fight against the privacy police to make any progress.” An executive with a similar viewpoint stated: “Look at X’s Grok, it keeps improving because they have all the data from X. What does Apple use to train its own AI?”

Apple stands at the crossroads of data and privacy, with its previously advertised advantages turning into technical disadvantages. In the chaotic AI market, Apple seems overly “elegant.”

04 A single move affects the whole body

Apple’s weakness in AI affects more than just Apple Intelligence.

Last year, Apple terminated its self-driving car project, which had cost billions of dollars and lasted for a decade, partly because AI could not deliver on the promise of fully autonomous driving.

According to Gurman’s speculation, failures in AI could jeopardize Apple’s future product plans, including AR glasses, robots, and Apple Watches and AirPods that can recognize surrounding objects.

Apple’s erroneous bet was on car manufacturing | Image source: Apple Explained

If Apple is unable to inject AI into its hardware in the future, it will not only subvert the company’s belief that “hardware is the carrier of software,” but also impact the hardware and software ecosystem that Apple takes pride in.

Apple’s Senior Vice President of Services, Eddy Cue, once told colleagues that Apple’s dominant position in the tech industry is precarious.

Last month, the search volume for Google on Apple devices declined. Eddy Cue stated, “This has never happened in 22 years, and AI is the reason.” He acknowledged that users are increasingly relying on large language models for information.

Eddy Cue pointed out that Apple does not provide essential goods like ExxonMobil (the famous American oil company), and he is worried that AI for Apple could be like the iPhone was for Nokia; Nokia fell before Apple, and Apple may fall before AI.

He even bluntly stated that the iPhone might become irrelevant in ten years, “It sounds crazy, but it really could happen.”

The greater challenge also comes from outside. According to informed sources, in order to meet the EU’s expectations for new regulations, Apple is working to modify its operating system to allow users to switch the default voice assistant from Siri to third-party options.

This means that if Apple does not make significant breakthroughs with its voice assistant, users may not use Siri in the future, opting instead for AI assistants from OpenAI, Anthropic, Meta, Alphabet, X, or even DeepSeek.

Apple is not sitting idly by. According to employees, Apple is building a completely language model-based Siri in its AI office in Zurich, aimed at making Siri more conversational and capable of information integration. This secret project is referred to as LLM Siri.

Apple also has thousands of analysts in its offices in Texas, Spain, and Ireland to review the accuracy of the Apple Intelligence summary, compare data discrepancies, and assess the frequency of AI hallucinations.

As for Apple’s self-developed chatbot, some executives now want to turn Siri into a true competitor to ChatGPT.

To this end, the company has begun allowing Siri to access open networks and integrate multi-source data. According to employees, the chatbots being tested internally by Apple have made significant progress over the past six months, and some executives believe their performance is now comparable to recent versions of ChatGPT.

After waiting a year for the new system, the result is a color palette? | Image Source: Apple

Insiders at Apple revealed that for next month’s WWDC, the company plans to focus on upgrading existing Apple Intelligence features and adding some new functionalities, such as AI-optimized battery management and a virtual health coach.

The significant upgrade of Siri, which was unlikely to be mentioned much at WWDC a year ago, is now being anticipated. According to sources, despite the high hopes for “LLM Siri” within Apple, the company is preparing to separate Apple Intelligence from Siri in its marketing.

On one hand, they are worried that the comprehensive lag behind competitors like Siri will hinder the company’s AI promotion. On the other hand, Apple also dares not announce new features months in advance anymore.

Do you remember Kittlaus, the co-founder of Siri? He remains optimistic about Siri’s AI capabilities, saying, “All model companies don’t know what an assistant is, and Apple has been researching this concept since 2010.” He believes that Apple just needs to make Siri smarter: “Apple has the devices and the brand; as long as they ‘change Siri’s brain,’ it has a complete chance to become the preferred assistant.”

Fourteen years ago, the emergence of Siri placed Apple at the pinnacle of intelligent interaction. Jobs’ vision ignited the voice assistant market and gave birth to the chatbots we can now converse with using natural language.

However, today’s Apple is struggling on the AI track, with the delayed and faded Apple Intelligence, the stagnation and obsolescence of Siri, and the failure of autonomous driving…

Mark Gurman, who has access to internal information, reveals not only Apple’s superficial delays in AI but also the internal strategic swings, the dilemmas of technological pathways, and the friction between core culture and the pulse of the times.

The product strategies and corporate culture that once made Apple proud have become stumbling blocks for the company in the face of AI, which is reshaping the tech landscape at an unprecedented speed.

Apple’s future is overshadowed by the shortcomings of AI, Cook said it’s just a matter of time. But how much time does Apple have left?

At least for next month’s WWDC, we won’t see the expected answers.