Imagine a world where an infinite supply of digital white-collar workers stands at humanity’s disposal. This is not the plot of a sci-fi movie but the reality that dawned on us on March 14, 2023. On this day, GPT-4 was unveiled, artificial intelligence so advanced that it can surpass most humans in a vast array of cognitive tasks, even acing a medical examination. Welcome to the dawn of the AI age.
The world has been gifted the marvel of a cognition engine that:
- Grasps and follows intricate instructions
- Excels in multiple languages, with an impressive command of English
- Parades a deep and wide-ranging knowledge base, spanning from programming to philosophy
- Transforms and manipulates language, almost indistinguishably from humans
- Holds an internal representation of worldly operations, including a grasp of the Theory of Mind—the ability to attribute unobservable mental states to others
Even before this breakthrough, its predecessor GPT-3.5 had set a record as the most quickly adopted technology in human history—taking just two months to reach 100 million users.
Now, where is it all going? Allow me to share my humble forecast of this evolving landscape over the next few years.
Game over for AI labs—game on for digital product studios
In a world increasingly dominated by generalized models such as GPT-4, the balance is shifting. The staggering costs required to train these behemoths, reaching into the tens of millions of dollars, is a luxury accessible only to the biggest tech giants. This, coupled with the fact that these models often outshine custom-developed alternatives, puts into question the economic feasibility of in-house training. The marginal gains of doing it will soon not justify the associated cost.
The world is vast, however, with nearly infinite nooks and crannies in every value chain just waiting to be optimized by deploying generalized models. The tech giants will not bother with these smaller niches. The revenue potential is too small for them. Instead, their attention will be focused on:
- Training foundational models like GPT
- Building robust MLOps infrastructure
- Rolling out products for the broadest use cases
(e.g. GPT for business, integrating Large Language Models (LLMs) into office suites and so on)
The baton of successfully deploying artificial intelligence will therefore pass from AI labs to agile digital product studios. These studios will roll out solutions based on generalized models tailored to the end user’s needs in months, not years.
The cost of cognition equals the cost of computing
At present, utilizing GPT-4’s cognition capabilities comes at a significant premium, primarily because it’s the first and only LLM of its kind available. But with Google’s recent launch of PaLM 2, a model comparable to GPT-4, we can expect a downward trend in these costs.
Every subsequent model with comparable performance will further diminish them. In a few years, cognition will be as accessible and commoditized as electricity, billed in tokens. The provider will matter less, as long as the cognition keeps flowing.
The future is multimodal and hyper-personalized
GPT-4 is a testament to the power of multimodality. The future won’t differentiate between a voice command, a picture, or a WhatsApp message. The AI will digest it all and respond in the most fitting format.
On top of that, it offers a world where everything is personalized for you. Every stream of information can be fine-tuned to your liking, from news to university lectures. It’s like having a personal Socratic tutor, a news anchor who speaks your language, and a friend and advisor who’s always by your side—understanding, empathetic, and helpful.
The future favors the pioneers
The first movers will hold a significant advantage in this rapidly evolving landscape. Understanding the massive paradigm shift and learning to harness the power of multimodal AIs will be pivotal. The sooner you embark on this journey, the better your chances are of outpacing the competition.
The good news is, if you’re reading this, you’re not late. The starting line of this race was drawn just half a year ago with the launch of ChatGPT 3.5. Most of the world is yet to catch up with the new reality.
Start by thinking about how AI can enrich your business or personal life—perhaps providing new insights and solutions or enhancing your productivity and learning. Dedicate a few hours a week to experimenting with AI’s potential in easing your life’s burdens, and you’ll be amazed by what you can achieve in a few months.
The bad news, however, is that the world of AI moves at a lightning-fast pace. Seminal papers that redefine the game are published every week. It takes time for state-of-the-art advancements to permeate the public consciousness, and the harsh reality is that solutions developed today might become obsolete in a year or two. The key is to comprehend the core concepts and periodically integrate the latest advancements into your solutions.
The future is calling, and we are answering
In this new era where cognition is as accessible as electricity and personalization becomes the norm, we see opportunities to redefine our lives. It is an era of liberation where AI frees us from mundane, repetitive tasks and allows us to focus on what truly matters—our friends, families, hobbies, passions, and the world around us.
As a teal organization, one of our guiding principles is listening to the evolving needs of mankind. Humanity does not need yet another software product studio. It needs an applied AI product studio—a group of professionals who can wield the power of artificial intelligence and use it to build products fitting into the nooks and crannies that this revolution has uncovered.
We’re answering the call, focusing all our efforts on developing capabilities to help our clients navigate the new landscape. Our mission is to assist our clients in building AI-powered products that save humanity millions of hours of meaningless work.
As we step into this bold new era, we invite you to join us. Together, we can navigate the exciting challenges and unprecedented opportunities that the AI age presents. The future is calling, and we are ready to answer. Are you?