In a bold prediction that challenges the foundation of modern mobile computing, Nothing CEO Carl Pei has declared that traditional smartphone applications are destined for obsolescence, to be replaced by proactive artificial intelligence agents. Pei outlined this transformative vision during an interview at the SXSW conference in Austin, Texas, on March 16, 2026, positioning AI not as an added feature but as the core interface of future devices.
The End of the App-Centric Smartphone Model
Carl Pei, the co-founder and CEO of the consumer electronics company Nothing, argues that the current smartphone paradigm has stagnated. He describes the familiar routine of unlocking a phone, navigating home screens, and opening full-screen applications as a relic of the past two decades. This model, he contends, creates unnecessary friction for users who must manually orchestrate tasks across multiple, disconnected apps. For instance, arranging a simple coffee meeting might require separate applications for messaging, mapping, ride-hailing, and calendaring. Pei’s critique centers on this inefficiency, suggesting that technology should evolve beyond requiring users to act as their own system integrators for every intention.
Consequently, Pei envisions a fundamental shift where the device’s operating system understands user context and executes tasks autonomously. The core value, therefore, migrates from individual app ecosystems to the intelligence of the agent itself. This perspective carries significant weight given Nothing’s successful $200 million Series C funding round in 2025, which was secured partly by pitching this AI-first device concept to investors seeking the next platform shift.
From Command-Based AI to Proactive Intelligence
Pei delineated a multi-stage evolution for AI integration into personal devices. The initial phase, already observable in the market by early 2026, involves AI assistants that execute specific user commands. These systems can book travel or make reservations when explicitly asked. However, Pei characterizes this step as “super boring,” viewing it as a simple automation of existing manual processes rather than a reimagining of the interaction model.
The subsequent and more profound phase involves AI that learns user preferences and intentions over the long term. In this stage, the system transitions from reactive tool to proactive partner. For example, if a user expresses a goal of improving their health, the AI agent could surface timely suggestions, coordinate schedules for exercise, or recommend dietary choices without explicit prompting. Pei compares this capability to advanced implementations of AI memory features, where the system’s deep understanding enables it to anticipate needs the user hasn’t consciously articulated.
A New Interface for a New User
A critical component of Pei’s vision is the complete redesign of the device interface. He asserts that the future does not involve AI agents clumsily mimicking human interactions with touchscreens, tapping through menus designed for fingers. Instead, it requires building entirely new digital interfaces specifically for AI agents to operate within. This means the primary user of certain system layers would be the AI itself, accessing services and data in a frictionless, backend manner to fulfill user intentions. The human-facing interface would then focus on presenting outcomes, suggestions, and confirmations, radically simplifying the user experience.
This transition does not imply the immediate disappearance of apps. Nothing’s own operating system, Nothing OS, currently includes features that allow for user customization and mini-app creation. The path forward, as Pei describes, involves a gradual evolution where the underlying functionality of services remains but is increasingly accessed and orchestrated by AI agents rather than directly by users through dedicated app icons.
Industry Context and Expert Perspectives
Pei’s comments arrive during a period of intense exploration within the tech industry regarding post-app interaction models. Major platform companies are investing heavily in AI agents capable of cross-application task execution. The development of large language models (LLMs) with reasoning capabilities and the expansion of AI-powered memory and personalization features provide the technical foundation for this shift.
Analysts note that while the concept is compelling, significant hurdles remain. These challenges include:
- Technical Complexity: Creating AI agents that reliably and securely perform multi-step tasks across independent services.
- Business Model Disruption: Rethinking app store economics, developer revenue, and platform governance in an agent-centric world.
- Privacy and Trust: Ensuring users are comfortable delegating significant intention and action to an AI that requires deep personal data access.
- Standardization: Developing protocols for AI agents to interact with diverse online services consistently.
Historical parallels exist in computing’s evolution, such as the transition from command-line interfaces to graphical user interfaces (GUIs). Each shift rendered previous interaction paradigms obsolete while unlocking new use cases and user bases. Pei’s argument suggests the mobile industry is on the cusp of a similar, agent-driven transformation.
Conclusion
Carl Pei’s prediction of disappearing smartphone apps underscores a broader industry trajectory toward ambient, proactive computing. His vision of AI agents replacing manual app navigation presents a future where technology acts on our behalf, driven by an understanding of context and intention rather than explicit command. While the timeline for this shift remains uncertain and the technical and commercial challenges are substantial, the direction signaled by industry leaders like Pei is clear. The next era of personal technology may not be defined by the apps we download, but by the intelligence of the agents we trust to use them for us.
FAQs
Q1: What did Carl Pei say about smartphone apps?
Carl Pei, CEO of Nothing, stated that smartphone applications will eventually disappear, replaced by AI agents that understand user intentions and execute tasks automatically without requiring manual navigation through separate apps.
Q2: What is an AI-first device?
An AI-first device is a conceptual smartphone or personal device where artificial intelligence is the primary interface. Instead of a user opening apps, an AI agent proactively handles tasks by accessing services directly in the background based on learned preferences and context.
Q3: Is Nothing building an AI smartphone?
Nothing has publicly discussed its vision for an AI-first device and used this concept during its 2025 funding round. While the company is actively exploring this direction, specific product details and launch timelines had not been announced as of March 2026.
Q4: How would AI agents work without apps?
The underlying services and functions that apps provide would still exist, but AI agents would access them through dedicated application programming interfaces (APIs) designed for machine-to-machine interaction. The user would interact with the agent’s decisions, not the individual service interfaces.
Q5: What are the biggest challenges to this vision?
Major challenges include developing sufficiently reliable and trustworthy AI, establishing new technical standards and business models, and overcoming significant user privacy concerns related to granting an AI deep access to personal data and the ability to act autonomously.
Updated insights and analysis added for better clarity.
This article was produced with AI assistance and reviewed by our editorial team for accuracy and quality.
