What Would Be The Perfect AI device?
AI Series: Mid-Series Special 9
I am a software engineer.
Welcome to the first Mid-Series Special of the year, and our 40th publication on this series š¤ š„
Letās get started šØ
Among all the attempts at creating the perfect AI device thatāll replace the smartphone as we know it, all have failed. I wrote a piece on this on the 4th mid-series special, AI the Product vs AI the Feature: A Case Study of Humane AI Pin and Rabbit R1. Now, in that special, I mentioned the grand failures of Humane AI Pin and Rabbit R1 that raised $230 million and $30 million, respectively. I recommend you check that special out. There are other more recent products like the strange Friend AI Pin, which is an AI device you wear around your neck like a necklace that serves as your āfriendā. Isnāt that just sad? š
But really, why do these products fail? I think we should first try to understand a core reason why products succeed.
What makes a great product?
I think a core reason products succeed is Real-World Adaptation.
Letās start with maybe the most famous of them all, the iPhone. Steve Jobs famously introduced the iPhone as, āA widescreen iPod with touch controls, a revolutionary mobile phone, and a breakthrough internet communicator. These are not three separate devices. This is one device. And we are calling it iPhone.ā The iPhone was an adaptation of the 3 great technologies into one. Letās go a bit further back to when Steve Jobs and his team were working on the Macintosh OS, and he vehemently insisted on things like the naming of the āTrash Binā and the icon being an actual trash bin, same with folders, he insisted that it had to be intuitive and feel natural, something a child can figure out by themselves.
Fastforward to Apple Inc products like the AirPods, Apple Watch Series, and their famous M series chips. Itās all a story of Adaptation. Many, including myself, believe adaptation is Appleās playbook.
It is easy for something to be intuitive when it feels familiar. The learning curve becomes almost flat.
We can also see this with the latest strides Mark Zuckerberg is making with the smart Rayban Meta Glasses, adapting an already existing familiar product, weāve all seen and maybe used eye glasses right? š¶ļø
Weāve seen products from these companies as well (Meta and Apple Inc), like the Meta Quest VR headsets and the Apple Vision Pro. These are highly unnatural and strange devices. They are unfamiliar and so have proven difficult to succeed, even for enthusiasts and gamers. They are not adapted, more sci-fi (Ready Player One stuff), and not adapted from a familiar or natural object or product.
Mark Zuckerbergās Facebook name and Idea came from school. A "Facebook" is a student directory featuring photos and basic information. Adaptation, thatās why it succeeded, it was intuitive and familiar.
OpenAIās Sam Altman and the famous Jony Ive (Jony Ive is the person who designed the iMac, iPod, iPhone, Apple Watch, etc.) are reportedly working on a new AI device, which I really donāt have much confidence in. Just the Idea of a dedicated AI hardware device is just a wrong turn. I am also of the view that AI is a feature or software product, not a dedicated hardware device.
Now, if I keep writing examples of Adaptation in technology, I might as well write a book lol, maybe I should š. So, letās get back to our question: what will be the perfect AI device? Given what weāve learned about Adaptation in technology.
What should we expect from the perfect AI device?
The ideal AI device should be:
A real-world adaptation that feels intuitive and natural to use.
A hands-free device that can be conveniently carried by the user throughout the day.
A device with sufficient processing power and reliable battery life.
A device with a usable touchscreen that supports both voice and touch input. A camera would be a welcome addition.
A device that supports, or is open to, app stores and development tools for third-party developers.
A device designed to complement the smartphone, not replace it.
Compared to some of the AI devices that already failed, they all relied primarily on voice inputs only, which is very strange, considering that in crowded/public places, meetings, etc., they would be basically unusable. Also, only the Friend AI Pin is natural, hands-free, convenient enough, and connects to your smartphone.
What will be the perfect AI device?
I have been thinking a lot about the Apple Watch lately in my quest for fitness. I believe in the Roman saying, Orandum est ut sit mens sana in corpore sano, which means one should pray for a sound body in a sound mind. And in my case, work towards it as well.
I never had a smart watch, never saw the need. You see, I am not really a wristwatch person. However, after several days, maybe weeks of research on whether I need a smart watch for my fitness journey, I suddenly came to an epiphany.
The ephiphany is that general consumer AI is a feature, and the perfect device is the smart watch. Particularly starting with tightly integrated premium ecosystems like the Apple and Samsung ecosystems. Remember the key here is adaptation. I figured the Apple Series and Samsung Galaxy watches provide a lot of features and processing power. They have app stores for watch apps and even support cellular connectivity, which is just perfect.
The smart watch. A device you can conveniently wear 24 hours a day.
I think this ephiphany came to me from a culmination of several events, one of which was when I was at the dentist some days back. While waiting my turn, I observed the dentist carrying out a procedure through the glass wall. She had an Apple Series Watch on her wrist and a couple of times checked her wrist maybe for notifications, made some swipes (with equipments in the patientās mouth lol), and got back to work. How convenient, she couldnāt do that with her phone or any other device I could think of, the Humane AI pin and the Rabbit R1 device inclusive. The convenience was just perfect.
The smartwatch would be the perfect device for AI software products and features. Without carrying out your phone, you could do a lot of the things the Humane AI pin and Rabbit R1 promised with their Large Action Models (LAM) models, ideas that never really worked. But the technology is getting better and would only keep getting better. LAMs are basically agentic models that can actually get things done instead of just generating text and media like the regular LLMs. So you can tell the LAM to do things like order you food, book the cheapest flight, and so on. The agent would interact with several services and apps to get the action done. I wrote a piece about them here: AI Agents: The future of AI? You can check it out to understand AI agents better.
The convenience of the smart watch, with the key feature of great products to adapt, makes this just right. With your smart watch, you can carry out LAM actions hands-free. You can even use the processing power of your smartphone in the Samsung or Apple Ecosystem to carry out more intensive task and send it back to the smart watch. Before we proceed further, let me enlighten you a bit more.
Short History of the Smart Watch.
The history of personal watches started with pocket watches. The first wrist watches were made for women (Queens šø), which was the adaptation of bracelets with the pocket watch by Abraham-Louis Breguet himself in the 1800s. During the Boer War and World War I, soldiers found pocket watches impractical and began strapping them to their wrists for quick, hands-free time checks. Thatās how men started wearing wrist watches as well. You see adaptation, right? Natural and intuitive adaptation. As for smart watches, they started with digital watches in the 1970s and continued improving to what we have today, with the first Apple Watch launched at 2015.
The smart watch was adapted for health, fitness, and wellness because of its position on the wrist, where it can get a lot of health data. I believe we can go one more for AI.
Now, letās go back to our hypothesis.
This idea looks fantastic. It checks almost all the boxes. But I, for one, donāt have a smart watch, so what are the numbers of active smart phone users? There are over 100 million active Apple Series smart watch users and a total of over 450 million smart watch users in the world, with yearly growth of about 8 to 9 percent and an age range of 18-34, making up the majority, 58% of whom use it mainly for health and fitness reasons. You see, you already have a market šŗ
Now back to my dentist. This would open more possibilities for her to do more with her convenient hands-free smart watch, like order food A for me from Restaurant A and order food B for my husband. I want it delivered by 2 pm sharp. Or something like order a ride for me from Uber to go home by 5 pm, add a stop to supermarket C. But then, if this is all AI can do for the general consumers, then thatās underwhelming, right? There has to be more than Generative and Agentic implementations. The question is, is there more? š
I believe in this idea so much that I would get an Apple Watch (I donāt like the price though šŗ), but itāll be a worthwhile investment to at least try out this idea. Iāll then ensure I build an app for it and have it available in the Watch OS App store before the middle of the year. I am open to ideas on what to build, so share them with me. While working on this, Iāll keep working on fitness as well, as you know the motto by now, mens sana in corpore sano, a sound mind in a sound body.
There are some other devices that are worthy of mention, which Iāll call runner ups, like the Rayban Smart Glasses, AirPods, and Smart Rings. The Rayban is a cleverly adapted smart glasses thatāll for sure keep getting better, but it for sure does not have the convenience of the smart watch, and might seem more distracting than seamless. AirPods are not it, at least for people who donāt wear their AirPods 24 hours a day, and it lacks a display, so the convenience is just not there. The Smart Ring is very convenient, maybe even more convenient than the smart watch, but it can not do anything by itself. It needs to be connected to another device like a phone or smart watch. I see these other devices more as complementary devices with the smart watch as the primary AI device, the AirPods, the smart rings, and the smart glasses.
What do you think? Do you have a better device in mind? Let me know.
