Apple and OpenAI pursue opposite hardware strategies for the AI era

By reading this article, you will understand the following:

• The specific reasons why smartphone screens are getting larger in the AI era
• The reasons why Apple and OpenAI are trying to create polar opposite devices
• Which is the real contender: foldables, AR glasses, or screenless AI devices?

I recently ran into a problem while using generative AI on my smartphone. The AI's answers themselves are quite smart. It can read documents, create comparison tables, and edit text. However, when I try to check the original source material, the screen switches, and when I return to the AI's answer, it has forgotten the numbers I just looked at.

"The AI has become smarter, but only my thumb is scrolling infinitely."

Even though I'm supposed to be using a tool of the future, what I'm doing is like someone from the Edo period opening and closing sliding doors.

Amidst this, Apple announced its first foldable iPhone, the "iPhone Duo." It is 5.4 inches when closed and 7.6 inches when opened. Apple describes it as a large canvas for browsing, gaming, and multitasking.

Meanwhile, OpenAI has integrated the 'io' team, including former Apple Chief Design Officer Jony Ive, and is attempting to create products and interfaces for the AI era that are different from the past. The first candidate being reported is, surprisingly, a screenless AI device.

One side is expanding the screen. The other side is eliminating it.

To conclude, neither one is the only correct answer. Devices in the AI era will diverge into larger screens for humans to view information and a 'screenless' approach where processing is left to the AI, reducing the amount of time spent looking at a screen.

And what will ultimately be important is not the size of the screen, but where humans confirm the results entrusted to the AI.

So, this time, I will connect the foldable iPhone, spatial displays, AR glasses, and OpenAI's AI device to examine the 'future of screens' in the AI era. I hope you find this helpful.

[Conclusion] AI-era devices are heading in two directions

Apple chose 'see more'

First, as a premise, what Apple is trying to solve with the iPhone Duo is a simple contradiction that smartphones have held for many years.

The larger the screen, the more convenient it is. However, the smaller the device, the easier it is to carry.

Conventional smartphones have forced their way through this contradiction by making the device itself huge. They shove a board in the late 6-inch range into a pocket and insist, 'This is a mobile phone.' It's basically a small cutting board now.

The iPhone Duo is 5.4 inches when carried and opens to 7.6 inches only when needed. In other words, it changed the screen size from a fixed value to a variable one.

According to Apple's official announcement, the iPhone Duo's inner screen is 50% larger than the iPhone 18 Pro Max, and iOS 27 has been redesigned to change according to how it is folded. The important thing is not the novelty of the hinge; it is that the iPhone has begun to swallow the role of a small iPad when necessary.

OpenAI aims for "not having to look"

However, if you try to solve the problems of the AI era solely by expanding the screen, things get a bit strange.

If AI processes a massive amount of information but humans still have to re-read all of it, the workload doesn't really decrease. If a human gets worried and checks 100 pages after the AI has summarized them, you've just added one more coworker to the task.

Therefore, an alternative solution from the OpenAI side is not to make the screen larger, but to reduce the number of operations and checks required.

In July 2025, OpenAI announced that the io Products team had officially integrated into OpenAI. The official statement notes the awareness that while computers have reached the stage of seeing, thinking, and understanding, the experience remains bound by old-fashioned products and interfaces.

Even though AI has changed, should the vessels we use to access AI remain smartphones and PCs?

This is where the path diverges from Apple. Apple expands the existing screen. OpenAI attempts to reduce the existing operations themselves.

The conclusion of the previous chapter is that there are two directions for AI devices: "seeing more" and "not having to look." So, why does using AI in earnest require a large screen? Next, let's consider this through specific tasks.

[Larger Screens] 6 inches is not enough when AI becomes your workspace

Generative AI has shifted from a chat interface to an editing desk

If you are only asking AI one-off questions, a large screen is unnecessary. If you ask "What's the weather today?" and it replies "Rain," that's the end of it. In fact, a wristwatch or voice interface would suffice.

However, current generative AI is changing from a mere know-it-all chat bot into a workspace.

For example, consider situations where you use AI for work.

• Reading a PDF proposal while looking for key points and contradictions
• Comparing customer emails with internal rules to draft a reply
• Comparing prices, performance, and warranty conditions of multiple products
• Checking spreadsheet figures while correcting graphs or reports
• Giving instructions on corrections while showing an image

What you need is not just a conversation with AI. You want to see the source material, the AI's response, the finished product, and the approval button all at the same time.

However, on a regular smartphone, every time you look at one thing, something else disappears. You copy the original email, paste it into the AI, copy the response, return to the email, and then switch screens again to check the recipient. The AI, which should be convenient, is just adding new legs to a copy-paste relay race.

If you use AI as a consultant, a small screen is sufficient. However, if you use AI as a collaborator, you need the space of a desk.

Comparison, approval, and correction require display area.

A large screen is especially necessary when you are verifying AI answers rather than taking them at face value.

AI is good at providing options. However, final judgment requires comparing option A and option B, the original information, costs, and risks. The more a task involves consequences where a mistake cannot be laughed off—such as booking flights, changing insurance, proposing to clients, or transferring money—the more screen space is required.

With a foldable device, for example, you can place the original document on the left and the AI's analysis on the right. If you place correction instructions or approval controls at the bottom, your smartphone becomes a small workstation.

In other words, the 7.6-inch screen of an iPhone Duo is not just a luxury for watching videos in a larger format. It is the area needed for humans to compare, correct, and responsibly approve the vast amount of information generated by AI.

That said, after reading this far, you might think, 'If AI is smart, don't I not need to see everything in the first place?' That question is correct. Next, we move on to the logic of turning off the screen.

[Screenless] The smarter the AI, the less it needs a screen.

An alternative solution: reducing operations.

To put it simply, as AI becomes truly intelligent, screens will become unnecessary for simple daily tasks.

Currently, to check tomorrow's schedule, you take out your smartphone, unlock it, open the calendar, navigate to the date, and read the schedule. We don't question it because we are used to it, but if you look at it calmly, there are many steps.

With an AI device, you just need to ask, 'Am I busy tomorrow?'

If the AI understands your schedule, travel time, weather, and unread emails, it can reply, 'You have a client visit at 10:00 AM, and since it's raining, you should leave by 9:00 AM. The afternoon document is still unfinished.'

What you need is not a calendar screen, but the answer required for decision-making.

The same applies to shopping, cooking, traveling, and operating home appliances. If an AI device equipped with a camera and microphone understands its surroundings, you can ask, 'What is this document?', 'What can I make with the ingredients in the fridge?', or 'Why is this lamp red?' You can skip the ritual of taking photos, attaching them, and selecting apps.

While smart speakers were 'boxes that listen to commands and execute them,' AI-native devices aim to be tools that see, hear, understand context, think, and act if necessary.

What is confirmed and what is at the reporting stage regarding OpenAI devices?

I want to separate facts from speculation here.

What OpenAI has officially announced is the integration of the io team and that Jony Ive and LoveFrom will be deeply involved in the design and creative aspects of OpenAI as a whole. Specific product names, shapes, prices, and release dates have not been officially announced.

On the other hand, according to reports from Bloomberg and others introduced by Reuters, the first consumer product is expected to be a portable, screenless smart speaker type equipped with cameras and sensors, with a launch envisioned for 2027 or later. Other reports mention smart glasses and smart lamps as future candidates.

Therefore, it is not accurate to say that "OpenAI has decided to release a screenless device." However, their official concerns and the reported content are pointing in the same direction. They are trying to build a new vessel suited for AI from scratch, rather than forcing AI into a smartphone.

Of course, screenless devices have major weaknesses. These include AI mishearing or misinterpreting, privacy concerns, resistance to always-on cameras, inability to use them in quiet places, and the difficulty of reviewing results in a list.

The fact that pioneering products aiming to replace smartphones, such as Humane's AI Pin, have struggled cannot be ignored. Getting rid of the screen does not automatically lead to the future. If you remove the bowl from a beef bowl and call it a "next-generation beef bowl," you just end up with sticky hands.

In the previous chapter, we saw that a large screen is necessary for verification and editing. In this chapter, we learned that if you leave simple operations to AI, you can eliminate the screen. So, is there a way to remove the screen from the physical device while still showing it only when necessary? That is where the third path, spatial display, comes in.

[Spatial Display] The path of moving the screen into space rather than erasing it

The current state of aerial displays

Like in sci-fi movies, projecting a 30-inch screen into empty air where anyone can see and touch the same image—unfortunately, such technology is not yet at a stage where it can be placed in an average home.

Ordinary projectors reflect light off walls or screens to make images visible. Since empty air barely scatters light, the image cannot be seen as a screen without a projection surface.

However, similar technologies do exist.

These include aerial imaging that uses special optics to make light appear as if it is coming from a point in mid-air, methods that use fog or fine particles as a screen, and aerial plasma that uses ultra-short pulse lasers to make the air emit light.

However, viewing angles, device size, resolution, brightness, power consumption, and safety are all hurdles. The future where you pull a smartphone out of your pocket and expand an Excel sheet into the air at a cafe is still a bit far off. If your quarterly sales figures overlap with the person next to you's pasta, that would be a nightmare in its own right.

AR glasses are the most realistic third path

At present, the most realistic approach is not to create images in the air itself, but to use AR glasses to make it appear as if there is a screen in the space.

With AR glasses, you can place multiple virtual screens only within your own field of view. Email on the left, AI in the center, and documents on the right. You can create a large workspace only when you need it, without having to carry around a physical display.

This method lies between "large screen" and "screenless."

Usually, you leave things to voice AI. Only when you need to compare or approve something do you bring information into your field of view. You don't need to carry a giant device at all times, nor do you need to trust everything based solely on voice.

However, AR glasses also face issues with weight, battery, heat, field of view, price, social acceptance, and wariness of cameras. It requires not just technology, but a design that ensures the person sitting next to you on the train doesn't have to worry, "That person's glasses aren't recording me, right?"

So far, we have seen three paths: Apple expanding the physical screen, OpenAI reducing operations, and AR displaying only the necessary screens in space. Finally, I would like to consider which one will win.

[Future Prediction] What remains in the end is not the screen, but the right to verify

See, delegate, and verify

I don't believe the future of devices will converge into a single form. I think we will move between three modes depending on the task.

1. When you want an answer immediately, leave it to AI via voice.
2. When you want to compare or edit, use foldables or spatial displays to see more.
3. Before sending money, contracts, or work, humans must explicitly confirm.

A 7.6-inch screen is unnecessary for weather, timers, or checking simple schedules. On the other hand, it is too scary to leave a customer quote or a 200,000 yen travel booking to a screenless AI that just says, 'I've taken care of it nicely.'

Therefore, the real point of contention is not whether the screen is big or small.

How far should AI process things, and at what moment should it return judgment to the human?

A good AI device is not one that does everything on its own. It is a device that allows you to switch stages according to the level of importance: handling it with just voice, displaying only the key points, or opening the source material for verification.

In that sense, the screen will not disappear; it will change from something you look at constantly into a verification device that appears only when necessary.

If it were me, I would wait for a foldable right now.

The iPhone Duo is interesting. However, for the first-generation foldable iPhone, I want to gauge the price, weight, durability, and app optimization. A first-generation device with a new form factor is a ticket to the future, but it is also easily a paid beta test participation pass.

Rather than jumping in right now, I would wait for post-launch real-world evaluations and a few generations of improvements. And for daily AI use, I would first separate 'tasks that require a large screen' from 'tasks that can be left to AI.'

• Document comparison and text editing on a PC or large screen.
• Schedule checking and simple searches via voice.
• Contracts, money transfers, and external communications must always be confirmed on a screen.

Even just by using it this way, you will be less likely to be swayed by the convenience of AI.

To repeat, devices in the AI era will not simply become massive, nor will screens disappear entirely.

Apple opens the screen so that humans can see the results of AI more broadly. OpenAI looks for a new vessel to reduce the very work humans do to look at things. AR glasses offload only the necessary screens into space.

And what remains in the end is not a slab called a display, but a place where humans can confirm, stop, and correct AI's judgments.

The foldable iPhone might not be the final form of the smartphone. Rather, I think it is a quite interesting transitional product that exists between the 'era where a screen is necessary' and the 'era where a screen appears only when necessary.'

I hope this serves as a reference for thinking about the future of AI devices. The end.