Apple's New CEO Pushes for Faster AI Development as Siri Prepares for a More Autonomous Future
From
mummycullen@mummycullen@gmail-dot-com.no-spam.invalid (MummyChunk) to
comp.sys.mac.advocacy on Mon Nov 9 14:23:40 2026
From Newsgroup: comp.sys.mac.advocacy
Apple's New CEO Pushes for Faster AI Development as Siri Prepares for a More Autonomous Future
Apple's new chief executive, John Ternus, is reportedly pressing the company's software leadership to accelerate artificial intelligence development, signaling a more ambitious approach to a technology that is rapidly changing how people interact with computers and smartphones. According to Bloomberg's Mark Gurman, Ternus recently held a private meeting with Craig Federighi, Apple's senior vice president of Software Engineering, expressing concern that the company is not moving quickly enough to match the capabilities being developed by its competitors.
The reported discussion comes just two months after Ternus succeeded Tim Cook as Apple CEO on September 1, 2026, and only weeks after the company introduced Siri AI, its substantially redesigned digital assistant. Although the new Siri represents an important improvement over previous versions, Ternus reportedly believes Apple must look beyond conversational assistance and develop AI systems capable of performing increasingly complex tasks with greater independence.
The immediate objective is to expand Siri's ability to interact with applications, manage information and complete multistep requests. Apple's longer-term ambition, according to Gurman, is to create a more autonomous assistant that can pursue ongoing objectives, remember relevant context and determine what actions are needed without requiring users to issue separate instructions at every stage.
The reported management pressure coincides with changes already taking place inside Apple's software organization. Engineers working on iOS 28 and other future operating systems are being given more time to experiment with new features, while the company has opened an expanded version of Siri AI's application integration system to third-party developers.
Apple has not publicly confirmed the private meeting or announced a release date for a fully autonomous Siri. Nevertheless, the combination of official product announcements, developer initiatives and detailed reporting about internal changes provides a clearer picture of how the company is approaching its next phase of AI development.
John Ternus Wants Apple to Move Beyond Improving Siri
The reported meeting between Ternus and Federighi represents one of the first indications of how Apple's new chief executive intends to influence software strategy.
Ternus spent most of his career overseeing hardware development, joining Apple's product design organization in 2001 and eventually becoming senior vice president of Hardware Engineering. During that period, he contributed to the development of numerous iPhone, iPad and Mac generations, along with products such as AirPods and Apple Watch.
Apple announced his appointment as chief executive in April 2026, with Tim Cook moving into the role of executive chairman. The transition was completed on September 1, placing Ternus in charge of a company whose future increasingly depends on the relationship between advanced hardware, intelligent software and cloud-based services.
According to Gurman's November 8 reporting, Ternus has become concerned that Apple's software organization is approaching artificial intelligence with less urgency than competitors developing increasingly capable assistants and autonomous agents.
In a closed-door meeting held in recent weeks, he reportedly instructed Federighi to accelerate Apple's software roadmap and broaden its ambitions beyond the improvements already planned for Siri.
The distinction is important because Apple has recently delivered one of the largest upgrades in Siri's history. The assistant now supports natural conversations, understands personal information across supported applications and can perform substantially more complicated actions than its predecessor.
However, the broader AI industry is moving toward systems that can plan and execute tasks rather than simply respond to individual requests. Ternus reportedly wants Apple to participate more aggressively in that transition.
His public comments have been more positive than the concerns described in Bloomberg's report.
During Apple's November 2 earnings conference call, Ternus said the company had finally delivered the version of Siri it had long wanted. He also emphasized that the assistant's architecture leaves considerable room for future improvements.
When asked about the growing importance of autonomous AI agents, Ternus acknowledged their potential but stressed that users must understand what information and capabilities they are granting to such systems.
There is no necessary contradiction between those statements and the reported private discussion. Apple can regard the current Siri as a significant accomplishment while also recognizing that the technology must advance quickly to remain competitive.
The question is whether the company's development organization can deliver those advances at the pace Ternus reportedly expects.
Siri AI Has Improved Considerably, but Its Capabilities Remain Limited
Apple introduced Siri AI in September 2026 as part of iOS 27, iPadOS 27, macOS 27 and its other major operating system updates.
The release represented a substantial departure from the traditional Siri experience, which had largely centered on voice commands, straightforward questions and relatively simple interactions with system features.
The redesigned assistant uses the latest generation of Apple Intelligence to understand natural-language requests, maintain conversational context and retrieve relevant information from supported applications.
A user can ask Siri to locate details from previous messages, retrieve information contained in email, examine photographs or respond to questions about material currently displayed on the screen.
The system also includes broader knowledge capabilities, allowing it to answer questions using current information rather than relying exclusively on a comparatively limited set of predefined commands.
Apple has integrated the assistant with a system orchestrator that coordinates different capabilities and determines which tools are appropriate for a particular request. Its underlying architecture combines on-device processing with additional AI infrastructure where necessary.
These improvements have made Siri more useful for activities that previously required manually searching through applications or transferring information between them.
However, the assistant remains primarily request-driven.
A user asks Siri to perform an action, the assistant attempts to complete that request, and the interaction generally ends. The system may understand context and combine several supported operations, but it is not yet designed to independently manage long-running projects or continually pursue objectives in the background.
That is the distinction Apple reportedly wants to address.
The current Siri also faces practical restrictions involving hardware, language support and regional availability.
Apple Intelligence requires compatible devices, including the iPhone 15 Pro models and newer supported iPhones, recent Apple silicon Macs and qualifying iPads. Older devices can receive operating system updates without gaining the full range of AI features.
The release of iOS 27.2 on October 26 expanded Siri AI beyond English to include French, Japanese, Korean, Portuguese and Spanish. Apple plans additional language support, but not every Apple Intelligence language currently provides access to the same conversational capabilities.
Siri AI also remains unavailable on iPhones and iPads in the European Union, where Apple has cited regulatory concerns involving the Digital Markets Act. This creates a substantial difference between the features available to customers in different markets.
Furthermore, some AI capabilities relying on server-side processing are subject to daily usage limits, with Apple indicating that expanded access may eventually require payment.
These limitations do not diminish the significance of the technology already released, but they demonstrate that the current Siri platform is still evolving.
Apple Wants Siri to Become a More Autonomous AI Agent
The next major objective involves what the technology industry increasingly describes as agentic artificial intelligence.
Traditional AI assistants generally operate through individual conversations or commands. A user provides an instruction, the system produces a response or performs a supported action, and further progress requires another interaction.
An AI agent can potentially operate with greater independence by maintaining an objective, identifying the steps required to achieve it and using available tools to complete those steps.
For example, a conventional assistant might help someone find information about upcoming travel. A more autonomous system could potentially monitor changing flight arrangements, identify scheduling conflicts, compare alternative options and prepare relevant adjustments without requiring the user to restart the process each time something changes.
Similarly, an agent managing a professional project might maintain a list of outstanding tasks, examine new information and propose the next appropriate action based on previously established objectives.
This kind of functionality requires more than a language model capable of producing convincing text.
An agent needs access to tools, mechanisms for remembering relevant information, methods for evaluating progress and safeguards governing which actions it can perform independently.
The distinction becomes particularly important when an assistant can interact with personal information, financial services, communication systems or applications that make changes on the user's behalf.
Companies including OpenAI, Anthropic, Google and Meta are developing different approaches to these capabilities. Some systems can interact with computer interfaces, while others use dedicated application connections or operate as persistent assistants that maintain context between conversations.
OpenAI's Dots and Meta's Muse are among the examples cited in Bloomberg's reporting, illustrating the industry's movement toward systems that can pursue objectives over longer periods.
Apple's current Siri architecture provides elements of this approach, particularly through its personal context understanding and ability to interact with supported applications.
However, Apple has not yet introduced an equivalent general-purpose system for controlling an entire computer interface.
Some competing AI products can examine screenshots, interact with graphical controls and navigate applications using simulated mouse and keyboard input. This approach is commonly called computer use and can potentially operate software that lacks a dedicated AI integration.
Apple's strategy has so far emphasized structured application access through its own software frameworks rather than unrestricted interaction with graphical interfaces.
There are advantages to that decision. Direct application interfaces can provide more reliable actions, clearer permissions and less dependence on the precise arrangement of buttons and menus.
However, they also require developers to expose supported functions. An agent cannot automatically perform every operation available in an application simply because that application is installed.
Gurman believes Apple could move toward a more autonomous Siri as early as 2027. That represents a prediction based on his reporting, not an officially announced launch window.
The final implementation could differ considerably from the persistent agents being developed by competing companies, particularly if Apple continues prioritizing tightly controlled application access.
iOS 27.3 Is Already Expanding Siri's Access to Applications
While the more ambitious agent capabilities remain under development, Apple has already taken a concrete step toward making Siri more useful within applications.
On November 3, the company announced a new Siri AI Developer Preview included in the beta versions of iOS 27.3, iPadOS 27.3, macOS 27.3 and visionOS 27.3.
The preview opens Siri's expanded capabilities to third-party developers through Apple's App Intents framework.
App Intents allows developers to describe information and actions their applications can provide to the operating system. Siri can then use those capabilities to respond to natural-language requests without requiring the user to navigate every application manually.
The updated framework introduces additional support for complex actions, including a new concept called skills. Developers can combine several related operations into a more sophisticated task that Siri can carry out through a single request.
Apple has demonstrated possibilities involving checking deliveries, managing reservations, placing orders and interacting with application content based on what the user is currently doing.
A person using a compatible ride-sharing application, for example, could ask Siri to change a pickup location while an active trip is already being tracked. The assistant may be able to infer the relevant service from the application's current state rather than requiring the user to specify every detail.
The framework also expands integration with Spotlight's semantic index, making additional application information available for contextual searches.
Apple has already begun introducing comparable capabilities across many of its own applications, including Calendar, Files, Finder, Mail, Maps, Messages, Music, Notes, Photos, Reminders, Safari and Shortcuts.
These integrations can support more detailed interactions with existing content.
For example, users may be able to ask Siri to locate a business they visited previously, organize information in Safari or identify files that have not been opened for an extended period.
The Verge reports that forthcoming improvements could allow Siri to perform more precise operations in Apple applications, including adjusting photographs and organizing browser tabs.
Third-party application integration is expected to become available more broadly during spring 2027.
Apple's official developer announcement confirms that developers can begin experimenting with the functionality now, while applications using the new capabilities are expected to become eligible for App Store submission in the spring.
This distinction is important. The existence of an API in a developer beta does not mean that every iPhone owner can immediately use Siri to complete complex tasks in arbitrary third-party applications.
Developers must implement the relevant functionality, Apple must complete its testing process, and supported features must become available in public software releases.
The expanding App Intents framework nevertheless provides an important foundation for Apple's longer-term AI ambitions.
By allowing applications to expose structured actions, Apple is creating a system through which Siri could eventually coordinate more complicated activities across multiple services.
Federighi's Management Style Faces Renewed Scrutiny
The reported pressure on Craig Federighi also reflects a broader discussion about how Apple develops software and makes technical decisions.
Federighi has been responsible for Apple's major operating systems for many years, overseeing platforms including iOS, iPadOS, macOS, watchOS and visionOS.
His responsibilities expanded significantly during 2025 following changes to Apple's artificial intelligence leadership.
In December of that year, Apple announced that John Giannandrea, its former senior vice president of Machine Learning and AI Strategy, would step down before retiring in spring 2026.
The company subsequently placed important AI responsibilities under Federighi, including oversight of Apple Foundation Models, machine learning research and AI safety.
Apple also recruited Amar Subramanya, a former Microsoft AI executive who previously worked on Google's Gemini Assistant, to lead several critical research and development areas while reporting to Federighi.
These changes were intended to bring Apple's foundational AI research and practical software development into closer alignment.
However, Bloomberg's latest reporting suggests that some employees have concerns about Federighi's management approach.
He is described as highly involved in important technical decisions, frequently challenging proposals and debating details with considerable intensity.
That level of engagement can help maintain consistency and technical discipline across complex software platforms, but some employees reportedly believe it also leaves insufficient room for alternative ideas to influence product direction.
The concerns are attributed to people familiar with Apple's internal organization and have not been independently confirmed by the company.
Another reported issue involves planning.
According to Gurman, Federighi tends to organize software development around a relatively short horizon of approximately 12 to 18 months.
Apple's hardware organization, by comparison, commonly works with plans extending several years into the future, reflecting the lengthy development cycles required for processor architectures, manufacturing technologies and new physical products.
Software development naturally operates differently because applications and operating systems can be modified more frequently. However, advanced AI platforms increasingly require substantial long-term investment in models, infrastructure, developer tools and specialized hardware.
A company planning a future generation of autonomous assistants may need to coordinate those technologies several years before the complete product becomes available.
This creates a potential challenge when software development is focused primarily on delivering the next annual operating system release.
The apparent objective of Ternus's intervention is not simply to make Federighi's organization work faster, but to encourage a broader approach to planning and innovation.
There is currently no indication that Federighi is being removed from his position. Bloomberg reports that succession planning has not been a major priority for him, and Apple's official leadership information continues to identify him as senior vice president of Software Engineering.
Apple Is Also Changing How It Develops iOS 28 and macOS 28
Alongside the leadership discussions, Apple is reportedly experimenting with a revised development process for its next generation of operating systems.
According to Bloomberg, the company normally organizes the development of major annual software releases into three principal milestones.
Each milestone typically consists of approximately four weeks of coding and feature development, followed by four weeks of integration and stabilization.
The integration period, often called convergence, allows different engineering teams to combine their work, resolve conflicts and prepare the operating system for subsequent testing.
For the first development milestone of iOS 28, iPadOS 28 and macOS 28, Apple reportedly changed that arrangement.
Engineers received six weeks for coding, experimentation and feature development, followed by two weeks of convergence.
The total milestone remained eight weeks, but the amount of time allocated to creating and exploring features increased by 50 percent.
According to Gurman, employees have responded positively because the revised schedule provides additional opportunities to develop ideas before decisions become fixed.
This is a noteworthy change for a company whose software engineering processes must coordinate work across numerous hardware platforms, applications and operating system components.
Allowing engineers more time to experiment could encourage more ambitious features, particularly when the underlying technologies are still evolving.
However, reducing the initial integration period also creates a potential trade-off. Complex software requires substantial testing to identify compatibility problems, performance regressions and interactions between components developed by different teams.
The practical effect will depend on how Apple manages subsequent milestones and whether additional testing compensates for the shorter early convergence period.
The change has so far been reported for the first development stage, not as a permanent reduction in integration time throughout the entire software cycle.
Apple is also making greater use of AI-assisted coding tools, which can help engineers generate code, investigate problems and experiment with alternative implementations.
These tools may improve productivity, but the generated results still require review, testing and integration into Apple's larger software systems.
The first milestone has reportedly been completed, with the second stage beginning in November.
Apple has not announced the feature lineup for iOS 28 or macOS 28. Based on its established release pattern, the company would ordinarily introduce the next major operating systems during its Worldwide Developers Conference in June 2027, followed by public releases later in the year.
Those dates remain expectations rather than confirmed announcements.
The development changes are significant because they suggest that Apple's AI ambitions are beginning to influence not just individual products, but the way its engineers organize their work.
Uncertainty Surrounds the Future of Siri Engineering Leadership
The pressure to accelerate AI development comes at a time when Apple may also face another leadership transition.
Bloomberg reports that Mike Rockwell, the executive who helped lead development of Apple Vision Pro before taking responsibility for Siri engineering, is considering leaving his current position or departing Apple entirely.
Rockwell became closely associated with the company's efforts to deliver its redesigned assistant after earlier Siri improvements encountered significant delays.
His experience spans complex hardware and software development, including technologies involving computer vision, spatial computing and sophisticated user interfaces.
That background made him an important figure in the effort to transform Siri into a more capable AI system.
According to Gurman, Rockwell is considering a change as early as 2027. However, no departure has been announced, and there is no confirmation that he has made a final decision.
A leadership change could introduce additional uncertainty at a sensitive stage of Siri's development, particularly as Apple attempts to expand the assistant's capabilities and strengthen integration with third-party applications.
On the other hand, Apple's AI organization now includes several senior engineers and researchers with extensive experience developing machine learning systems.
Amar Subramanya's responsibilities include Apple Foundation Models, research and AI safety, while Federighi maintains overall oversight of the company's core AI efforts.
Bloomberg has suggested that responsibility for Siri engineering could eventually be distributed among other senior leaders if Rockwell departs, but no official succession arrangement has been announced.
The more immediate concern is maintaining continuity while expanding the assistant's capabilities.
Developing a reliable AI agent requires coordination among language model researchers, operating system engineers, security specialists, application developers and teams responsible for cloud infrastructure.
Changes in leadership can affect how those groups work together, making organizational stability especially important when ambitious product deadlines are approaching.
Privacy and Security Could Determine How Autonomous Siri Becomes
One of the largest challenges facing Apple is reconciling increasingly autonomous AI functionality with its established approach to privacy and security.
Apple has repeatedly emphasized that its AI systems should process information on the device whenever practical, while using additional infrastructure for tasks that require more substantial computing resources.
Siri AI incorporates personal context understanding, meaning that it can work with information contained in supported messages, emails, photographs and other applications.
That information can make an assistant considerably more useful because users no longer need to provide every relevant detail manually.
However, expanding an assistant's authority from retrieving information to performing actions introduces additional risks.
An AI system capable of changing calendar appointments, sending messages, modifying documents or placing orders must correctly understand what the user intends and recognize when further authorization is necessary.
Errors that are merely inconvenient in a conversational response could become consequential when the system is allowed to act independently.
For example, an assistant that misunderstands a restaurant recommendation may provide an unsuitable suggestion. An agent that misunderstands an instruction to reserve a table could potentially make an unwanted booking or provide incorrect information to a third-party service.
More sensitive tasks involving financial accounts, confidential documents or business communications require stronger safeguards.
There are also risks associated with prompt injection, in which malicious instructions are embedded in webpages, documents or other material an AI system encounters while performing an otherwise legitimate task.
An assistant must distinguish between instructions from its user and untrusted information contained in the content it is examining.
Apple's reliance on structured application interfaces could help establish clearer permission boundaries, although the security of the overall system will depend on how those interfaces are implemented.
The company is also tightening controls around sensitive permissions for third-party AI agents on macOS, including access to protected files and personal information.
During Apple's November earnings conference call, Ternus acknowledged the potential of autonomous agents while emphasizing that people need to understand what they are allowing those systems to access.
This approach may produce a different user experience from agents designed to operate with broad permissions and minimal intervention.
More frequent confirmation requests could limit convenience, but they may also reduce the risk of unintended actions.
The challenge for Apple will be providing meaningful autonomy without requiring users to surrender control over their information or accept unpredictable behavior.
Ternus Faces a Defining Test of Apple's AI Strategy
The reported discussion between John Ternus and Craig Federighi reveals a broader challenge facing Apple as artificial intelligence becomes an increasingly important part of the technology industry.
The company has demonstrated that it can integrate sophisticated AI capabilities into its operating systems, and the arrival of Siri AI represents substantial progress after a prolonged development effort.
However, the expectations surrounding digital assistants are changing rapidly.
Users are beginning to encounter systems that can do more than answer questions, summarize information or generate content. The next stage involves assistants that can coordinate applications, maintain objectives and complete complicated tasks with less direct supervision.
For Apple, the opportunity lies in combining those capabilities with an enormous installed base of personal devices, a tightly integrated operating system architecture and access to information that users already maintain within its ecosystem.
The company also controls much of the hardware on which its software operates, giving it opportunities to optimize AI processing, memory usage and security across devices.
Those advantages do not automatically translate into leadership in autonomous AI.
Apple must still develop reliable models, provide sufficiently capable developer tools and deliver software experiences that compare favorably with increasingly sophisticated alternatives.
The latest reporting suggests that Ternus believes achieving those objectives will require changes in both the company's ambitions and its internal processes.
Some of those changes are already visible. Siri's application integration framework has entered developer testing, Apple is expanding the assistant's access to system functionality, and software engineers are reportedly receiving more time to experiment with new ideas.
Other ambitions, particularly the development of a persistent, autonomous Siri capable of managing ongoing work, remain unconfirmed plans rather than available products.
The significance of the reported meeting is therefore not that Apple has suddenly discovered the importance of artificial intelligence. The company has been investing in machine learning for years and has already introduced extensive AI functionality throughout its products.
What appears to be changing is the level of urgency attached to translating those investments into a more capable personal assistant.
For Ternus, the central question is whether Apple can preserve the reliability, privacy protections and consistent user experience associated with its software while accelerating development sufficiently to meet rapidly evolving expectations.
The next major milestones will be the broader release of Siri's third-party application capabilities in 2027 and Apple's presentation of its next operating systems.
Those developments should provide a clearer indication of whether the reported internal changes are producing more ambitious software or simply reorganizing the process through which familiar improvements are delivered.
The longer-term measure will be whether Siri evolves from an assistant that responds intelligently to requests into one that can responsibly carry out meaningful work over time. That transition, rather than another improvement in conversational ability alone, is emerging as one of the central challenges of Apple's new leadership.
--
Via JLA Forums web gateway for comp.sys.mac.advocacy
--- Synchronet 3.22a-Linux NewsLink 1.2