What if the strongest product-market fit for a foldable iPhone has almost nothing to do with folding a phone?
The more interesting opportunity may be solving what I call the Open-Laptop Problem: the moment when an AI agent is already doing the work in the cloud, but the human still needs to open a laptop simply to inspect, approve, redirect or validate what the agent is doing.
That friction looks small today. In an agentic computing environment, it could become a major product opportunity.
What is the Open-Laptop Problem?
Traditional computing assumes the human is performing most of the work. The laptop therefore needs a keyboard, large screen, multitasking interface and enough local computing power to support sustained production.
Agentic AI changes that relationship. Increasingly, the human describes the outcome, delegates work to an AI system and then returns at decision points. The agent may research, write, code, analyse, operate software or coordinate a multi-step workflow remotely.
The human's role starts shifting from operator to supervisor.
But today's device hierarchy was not designed around that behaviour. A smartphone is excellent for initiating a conversation, but sustained inspection of a complex agent workflow can feel cramped. A laptop solves the visibility problem, but opening one can be disproportionate when the user only needs to review a result, compare two states, approve an action or give the agent its next instruction.
That gap is the Open-Laptop Problem.
Why agentic AI changes mobile product-market fit
The question for future mobile hardware is no longer only, “How much work can this device perform locally?” A better question may be, “How effectively can this device let a human direct and supervise work happening elsewhere?”
Cloud AI makes this distinction important. If the expensive reasoning, coding, rendering or automation is happening remotely, the personal device increasingly becomes a control surface.
That control surface needs to be instantly available, comfortable for reading, large enough for inspection, excellent for voice input and capable of showing enough context to make decisions confidently.
This is where a foldable phone becomes strategically more interesting.
The possible product-market fit for an iPhone Duo
Imagine an iPhone-sized device when closed and a substantially larger workspace when opened. The value proposition would not need to be “a laptop inside your pocket.” Trying to reproduce the laptop may actually miss the opportunity.
The stronger proposition could be: the fastest interface for supervising AI work wherever you are.
Closed, it remains the communication device. Opened, it becomes the agent review surface: inspect generated work, compare alternatives, read research, review a storefront change, approve an advertising campaign, watch a coding agent progress or speak the next instruction.
The product is then positioned between the smartphone and laptop not by computing power, but by decision latency. It reduces the time between an agent asking for human judgment and the human being able to provide it comfortably.
Running an AI agent is different from supervising one
This distinction matters.
A device does not necessarily need to execute the agent locally to become central to agentic work. In many workflows, the agent can operate in cloud infrastructure while the user moves through the physical world.
The valuable hardware experience is therefore the interface around the agent: notifications, persistent task state, voice interaction, visual review, approvals, interruption handling and rapid switching between multiple running jobs.
In that model, the foldable display is not merely additional screen real estate. It becomes an expandable management console.
Why voice could be as important as the fold
Agentic computing also reduces the importance of typing for many tasks. A user can describe intent conversationally, let the agent translate that intent into operations and intervene when judgment is required.
This makes a combination of voice + expandable display + cloud agents particularly powerful.
Voice handles instruction. The larger screen handles verification. The cloud handles execution.
That is a fundamentally different interaction model from shrinking desktop software onto a phone.
Could foldable phones become AI agent control surfaces?
Yes, but only if their software evolves with the hardware. A larger display by itself does not create product-market fit.
An agent-first mobile interface would need persistent workflows rather than disposable chat sessions. Users should be able to see which agents are working, which are waiting for approval, what changed since the last review, what evidence supports a recommendation and what action requires human authority.
The winning experience may look less like opening ten applications and more like managing a portfolio of delegated work.
The broader signal: computing is moving from doing to directing
This is the larger product shift worth watching.
PCs were optimized for humans doing digital work. Smartphones were optimized for communication, consumption and lightweight interaction. Agentic computing introduces another mode: humans directing digital labour.
If that mode becomes common, hardware categories can be reevaluated around supervision rather than execution.
The best device may not be the machine capable of doing everything itself. It may be the device that lets the human remain connected to everything being done on their behalf.
A Next Product Move hypothesis, not a hardware prediction
This is best treated as a product hypothesis rather than a prediction about any specific Apple roadmap.
The signal is the changing job-to-be-done. As AI agents become more capable, users may spend less time continuously operating software and more time initiating work, reviewing intermediate states and making high-value decisions.
That creates a possible unmet need between phone convenience and laptop visibility.
In Codified Commerce terms, this is the kind of signal a Next Product Move analysis should surface: not “What feature should be added next?” but “What new customer job is emerging because the surrounding technology has changed?”
And it is exactly the kind of hypothesis that should remain a Moving Target. Agent capabilities, mobile interfaces, voice interaction and user behaviour will continue changing. The opportunity needs to be observed continuously rather than frozen into a one-time conclusion.
The question worth tracking
If AI agents increasingly perform the work, how often will people still need a laptop simply because their phone is not a good enough place to supervise that work?
If that frequency becomes significant, the next breakthrough in mobile computing may not come from putting more desktop capability into a pocket.
It may come from eliminating one surprisingly persistent action:
opening the laptop.
Ramesh Babu J · Founder, Codified Commerce · Independent Shopify Partner since 2020.
