OpenAI’s iPhone Moment is Still Coming
If 2023 was the world’s introduction to the true potential of AI, 2024 already seems to be the year when we see the technology’s full commercialization. Companies like Samsung and Google have announced a myriad of new AI features that they plan to integrate across their product lines. Apple is also highly rumored to be launching its own Generative AI capabilities at WWDC this year.
Yet, what is most intriguing is the number of new AI hardware announcements made within a span of three months. We have seen new product releases from newcomers such as Humane, Rabbit, Tab, and Rewind, each offering their unique vision of what AI’s future in hardware form might entail. This prompts the question: why focus on hardware at all? There are two main reasons that make the hardware aspect of AI as crucial for businesses as the advancement of the technology itself. It all boils down to businesses striving to establish themselves as platform leaders and the inherent limitations current consumer electronics face in the context of AI.
The Race to Become a Platform Leader
Why did Meta go all in on VR/AR while it seemed like the world was still grappling with its value proposition? Why did Tesla commit fully to building its own Supercharger network for its vehicles? It was all a strategic bid to become the platform owner, where newcomers entering the market would be essentially dependent on your platform. And no company has felt this pain more than Meta. Meta has never owned the platform on which they publish.
This, a common reality for many software giants, has particularly plagued Meta, especially considering their heavy reliance on user data for revenue generation. Apple’s changes to iOS in 2021 alone cost Meta $10 billion dollars. In business, leverage is everything, and no one has had more leverage than Apple in the last decade. So much so, that companies are resorting to the legal system to compel Apple to make some “concessions”, like taking their 30% cut on external payment methods as well as their internal ones.
Establishing New Realms vs. Dethroning Kings
Toppling a king from his throne presents a far greater challenge than establishing a new kingdom and eventually subjugating the old king. This very strategy was at the heart of Meta’s strategic pivot towards AR/VR, underscoring what AI represents for these companies today. The tech world is in a relentless pursuit of what might be termed their ‘iPhone moment’ — a golden opportunity to forge a new, dominant platform poised to redefine the technological landscape for the next decade and beyond. In this quest for innovation, while various contenders may posit differing visions, it is OpenAI that appears most favorably positioned to emerge as the definitive AI platform. This perspective is underscored when considering ChatGPT — if this can be seen as OpenAI’s iPod moment, then the strong indicators suggest OpenAI is ambitiously crafting their equivalent of the iPhone.
By Why Do We Even Need New Hardware?
But some of you might be wondering, “Why do we even need new hardware? Why not use old designs like those we have with smartphones or smartwatches and load AI onto them?” I think about it the same way you could make the argument with Blackberry and an iPhone form factor in the past. Could you run third-party, non-productivity apps on a Blackberry and interact with them? Yes, you can. Is the hardware designed around that, like an iPhone is? No, not at all.
In the same way, just as Blackberry wasn’t designed for third-party, non-productivity mobile applications, modern-day smartphones and electronics have not been designed from the ground up to interact with AI. You can throw in a neural chip for faster AI processing, add as many tensor cores as possible, and include a dedicated button to communicate with AI. However, even then, the hardware comes with limitations that don’t allow you to fully take advantage of the technology. These are some of the fundamental limitations I believe our hardware has today in regards to AI.
Context Gathering
Modern devices like smartphones, while sophisticated, lack specific design elements crucial for AI context gathering. For instance, AI requires an array of sensors to understand and interact with the environment effectively. This goes beyond the typical camera or microphone — it includes advanced depth sensors, thermal imaging capabilities, and more nuanced audio pick-up technology. These would allow AI to perceive the environment in a way that aligns more closely with human perception, thereby enhancing its responsiveness and utility.
Convenience of Interaction
The current mode of AI interaction, primarily through smartphones, limits the user’s ability to engage with AI seamlessly throughout their day. Imagine wearable devices that can detect gestures or even changes in your physiological state as a means of interaction.A device which is always there with you when you need it. Its part of the reason why Tab, Rewind and the Pin all rest out of your pocket so getting to this form of interaction is convenient. Meta’s smart glasses are another form factor that addresses this.
Mode of Interaction
Right now, our interaction with AI is largely voice-based, which, while useful, is just one aspect of human communication. Envision a future where AI can interpret not just what you say, but also how you say it — the tone, the emotion, the subtleties of human communication. This requires hardware capable of more sophisticated audio analysis and possibly even visual cues to fully understand the user’s intent. Additionally, AI should be able to communicate back in a way that is most effective for the situation, whether that be through voice, visual displays, or even haptic feedback.
Operating Systems
There is a big distinction between developing an operating system from the ground up for AI, as opposed to merely adding AI functionalities to existing systems. Current operating systems, while sophisticated, are primarily designed for traditional computing tasks. When AI features are added to these systems, they are often constrained by the foundational architecture not originally intended for AI. This leads to limitations in how deeply AI can be integrated and how efficiently it can operate.
Building an AI-specific operating system from scratch allows for a holistic approach where every aspect of the system is optimized for AI processes. This includes core functionalities like memory management, data processing, and user interaction, all tailored to complement and enhance AI capabilities. For instance, such an OS would have an advanced neural processing unit (NPU) integrated at its core, designed to handle complex AI computations more efficiently than traditional CPUs or GPUs could.
Furthermore, an AI-dedicated OS would inherently understand and anticipate user needs by continuously analyzing behavioral data, leading to a more intuitive and anticipatory user experience. This could manifest in features like proactive assistance, where the system anticipates the user’s needs before they’re explicitly stated, or adaptive interfaces that evolve based on user preferences and habits.
Another aspect is the integration of privacy and security within the AI OS. With AI processing vast amounts of personal data, an AI-specific OS can be built with advanced encryption and privacy-preserving technologies from the onset, rather than retrofitting these as afterthoughts.
Additionally, such an OS could better manage the energy consumption and computational demands of AI, optimizing battery life and performance in ways that current systems, which have to balance a wide range of tasks, may not be able to achieve.
In essence, building an operating system for hardware made specifically for AI allows for a more seamless, efficient, and user-centric experience. It’s akin to the difference between a multi-purpose tool and a specialized instrument; both can perform a task, but the specialized instrument does so with greater precision and effectiveness. This difference is crucial in leveraging the full transformative potential of AI in our daily technology use.
At the end of the day, the current hardware landscape is not designed with AI at its core. The journey ahead lies in discovering the optimal form factor that truly harnesses AI’s capabilities, much like the leap from traditional phones to smartphones revolutionized personal technology. The future holds immense promise for AI-integrated devices, potentially marking a new era in how we interact with technology.
So why OpenAI?
Why do I believe OpenAI is poised to capture this moment and have their “iPhone moment?” There are two main reasons underpinning this belief.
Firstly, OpenAI, as it stands, is indisputably a global leader in AI. The essence of the company is deeply rooted in fundamentally understanding and evolving AI from a research-centric technology to one that resonates with consumers. This transition marks a significant shift from the approach taken by Google DeepMind and other AI research entities, who, despite years of development, hadn’t brought AI into the consumer sphere until the advent of ChatGPT last year. Much of this transformation can be attributed to Sam Altman’s profound grasp of product development and his vision of harnessing technology to craft consumer experiences, rather than solely focusing on technological advancement. ChatGPT has rapidly integrated itself into the daily lives of many, positioning OpenAI as a synonym for AI innovation in the collective public consciousness.
The second, and perhaps most significant X factor, arises from a newly announced collaboration with a singular, influential figure: Jony Ive. As the mastermind behind the design of Apple’s modern product lineup, Ive crafted the quintessential form factor that has become the blueprint for consumer electronics worldwide. His visionary approach to design not only set a new standard within the industry but also inspired a myriad of companies to adopt a similar aesthetic for their products.
In recent years, Ive’s impact on technology seems to have been somewhat overlooked. Post-Steve Jobs, especially towards the end of his tenure at Apple, the lack of Jobs’ balancing influence on form versus functionality became apparent. This was evident in products like the 2016–2019 MacBook Pros, which ultimately contributed to Ive’s departure from the company. However, the announcement of Ive’s collaboration with OpenAI on new hardware design is particularly intriguing. If there’s anyone who can fundamentally reinvent hardware design for optimal AI interaction, it’s Jony Ive. I believe Ive is at his best when he’s tasked with reimagining a completely new product category from scratch — a challenge that OpenAI has now put before him.
The path to success is not without its enormous challenges and a constellation of critical elements must fall into perfect alignment. Sam Altman’s involvement in product development is crucial; he must emulate Steve Jobs’ legendary ability to harmonize form with function. OpenAI is venturing into uncharted territory, tasked with creating an innovative hardware product without relying on an established line of existing products — a challenge of a magnitude not seen since the debut of the groundbreaking iPhone. The success of this venture hinges on a visionary product concept that not only captures the essence of AI’s future but also resonates deeply with end-users. Achieving this monumental task necessitates the collaborative brilliance of the globe’s most gifted engineers, channeling their collective genius to transform an audacious idea into a groundbreaking reality.
However, if OpenAI can somehow do it, we are in for a paradigm shift that rewrites the narrative of technology as we know it. On January 9, 2007, years of research, ingenuity, and technological innovation converged in perfect harmony to transform the world as we knew it. The stage is set, the players are ready, and the world watches — waiting for OpenAI to potentially stick the landing and redefine the way we interact with technology.
I recently started a bi-weekly newsletter about IRL events, startups and building products! These articles will be posted there first so feel free to subscribe
Thanks for reading,
Daivik Goel
