This week, we are excited to share our first featured community article: our very own Chappy Asel’s manifesto on AGI.

What do Joe Biden, Elon Musk, Barack Obama, Vladimir Putin, Xi Jinping, Bill Gates, Warren Buffet, and Mark Zuckerberg all have in common? They are unanimously calling attention to the profound social, cultural, and even cosmic implications foreshadowed by the recent breakthroughs in artificial intelligence (AI). The line between creator and creation is blurring — challenging the very notion of what it means to be human. More existentially, many claim we are living on the precipice of life’s most significant evolution in its 3.7-billion-year history¹.

As anyone living in Cerebral Valley will tell you, AI is the new land of opportunity. While Google Research kicked off the gold rush in 2017, it was OpenAI that roused mass awareness when it released ChatGPT to the public last November². Announced merely four months later, its successor, GPT-4, surpasses human performance on a majority of standardized tests and is too powerful for full public release³.

With this blistering pace of progress, are we on the verge of some grand, culminating crescendo? Will the rate of technological advancement continue its exponential increase to near infinity? Or are we instead in the midst of a seventh AI hype cycle, with progress doomed to tail off logarithmically?

In a quest to demystify the exciting world of artificial general intelligence (AGI), this four-part manifesto answers this question among many others. It is a concise yet comprehensive map, generously dotted with links and footnotes inviting deeper self-exploration (this first part alone has over a hundred). In this journey, I am as much an explorer as you are. Consider this not as the final word, but as the sowing of seeds for much-needed dialogue and debate on AGI and our shared future.

We will begin by defining AGI and its related terminology, in the process presenting the grand promise that AGI presents. “Part II: Prediction” will assuage realists’ concerns by investigating possible fallacies and laying out practical timelines for our path to AGI according to the latest insights from world experts. “Part III: Perils” will examine the myriad existential threats brought forth by alarmists standing between us and AGI. Finally, armed with this full array of perspectives, “Part IV: Progress” will present pragmatic advice for grasping the near-term implications and structuring our individual lives in light of this understanding. Let us embark on this thrilling adventure together.

Often overlooked is the fact that OpenAI — the biggest name in the AI game — has the stated mission to “ensure that artificial general intelligence — AI systems that are generally smarter than humans — benefits all of humanity.” Google DeepMind’s long-term goal is to “solve intelligence, developing more general and capable problem-solving systems, known as artificial general intelligence (AGI).” This begs the question: what is AGI and why is achieving it a central goal for titans of the AI industry?

Put simply, AGI⁴ is a computer program that surpasses human performance in almost all intellectual tasks⁵. Obviously, creating an intelligence greater than ourselves will have extraordinary implications, promising to revolutionize every conceivable facet of our lives⁶. However, one does not need to look that far ahead — the fruits of this future are already ripening. Goldman Sachs and McKinsey estimates agree that the recent boom in generative AI alone could nearly double rates of worldwide GDP growth and birth a trillion-dollar industry. Still, these projections could be wildly conservative if advancements maintain their current clip. So, how do we get from where we are to humanity-altering AGI?

Superficially, creating an AI system more capable than ourselves presents an imposing challenge. However, underlying the incredible surge in complexity and capability of AI models in recent years is a simple yet powerful concept: the accelerating rate of technological change.

Perhaps the most famous quantifiable example of this is Moore’s law. Gordon Moore posited in 1965 that the number of transistors on computer chips doubles every three years — an observation which has proven true for over 120 years now. Plenty more examples of this exponential rate of return also exist. Global data is doubling every three years. Supercomputers are doubling in computational power every 1.2 years. Lighting is 1,898 times cheaper now than it was in 1900⁷.

Exponential change is not just an abstract concept, either — it plays into our everyday lives. The move from hunter-gatherer societies to agriculture took thousands of years. The transition from closed autocracies to majority adoption of electoral-based governmental systems took 174 years. Electricity took twenty years to accomplish that same mark. The internet did so in just over ten years. Smartphones needed only five years. ChatGPT, released mere months ago, is now used by over 14% of Americans.

There is a broad consensus among experts that paradigm shifts are occurring faster than ever. OpenAI CEO Sam Altman’s “Moore’s Law for Everything” goes so far as to imagine a future of abundance where everything, from education to clothing, follows exponential progress. Are AI systems bound to track this same trajectory, with intelligence exploding exponentially?

Neural scaling laws suppose that factors such as computational power, parameter count, and dataset size all correlate with model performance and thus intelligence. Analyses from OpenAI and independent research show promising and persisting positive relationships⁸. What’s even more impressive are studies pointing to larger AI models exhibiting emergent abilities and the formation of complex internal representations. Thus, AI seems poised to ride on the coattails of compounding exponential curves.

Let’s assume that everything goes to plan, intelligence continues to scale with technological progress, and we successfully develop AGI. What happens when machine sapience eclipses human sentience?

This intelligence explosion is often referred to as the technological singularity: a hypothetical point in time at which technological growth becomes uncontrollable and irreversible, resulting in unforeseeable changes to human civilization. Many of mankind’s greatest minds have theorized about this event — rightfully so as it is difficult to overstate the cosmically monumental nature of this moment.

This is not merely another mile marker in the inexorable march of technological progress but a unique, singular moment where a “seed AI” capable of recursive self-improvement ushers in a new era of intelligent, non-biological life. Max Tegmark describes the nature of this transition well in Life 3.0:

As we continue to see this ecosystem evolve, we would love to hear your perspective on how you see the future of AGI changing the way we interact with generative technologies. If you have any ideas or insights you would like to share, hit up Eric on the community Slack or reach out via email at [email protected] for a feature in our next newsletter!

We hope everyone had a wonderful Thanksgiving! Such a joy to gather together for a warm and cheery virtual Friendsgiving! Who can guess what the number one topic of discussion was? 🃏☕️

About Eric Fett

Eric joined The GenAI Collective in early September to lead the development of the newsletter. He is currently an investor at NGP Capital where he focuses on Series A/B investments across enterprise AI, cybersecurity, and industrial technology. He’s passionate about working with early-stage visionaries on their quest to create a better future. When not working, you can find him on a soccer field or at a sushi bar! 🍣