This is the first of three posts where we will in turn take a deeper dive into intelligence, information, and complexity. 

One of the key pillars upholding Frankenstein’s World is the rapid increase of intelligence. Much like the eponymous creature, intelligence arises out of what appears to be thin air. The situation is similar to the Cambrian explosion of 540 million years ago, in which life emerged seemingly out of nowhere and most major layouts of animals appeared. It marked the beginning of the rise of complexity on Earth. Today we are undergoing a Cambrian explosion of intelligence. In this post, we will cover the types of intelligence with us today as well as the long term consequences of the explosion itself. 

We can divide intelligence into two main buckets; biological and artificial. Biological intelligence is the natural cognitive ability of humans and animals. For the remainder of this post, I will be talking about human intelligence. Animal intelligence, while fascinating in its own right, has not had the transformative effects on the Earth as human. 

Human intelligence emanates from the eight billion humans inhabiting the planet. Everyone perceives the environment around them, predicts what will happen next, and acts according to their own interests and needs. Human intelligence, while coming out of individuals, is primarily collective, which means groups of humans sharing their knowledge with each other and acting collectively as a group. Collective intelligence is needed whenever understanding a problem or situation lies far beyond the capabilities of a single individual. It is with collective intelligence that the explosion really begins. 

Individuals band together in groups to from a sort of collective hive mind. These groups range in size and sophistication from a small group of friends to states and alliances of countries cover billions of individuals. There is a virtually infinite number of combinations of collective groups that can be formed in the world today (there are ten million nonprofits/NGOs alone), with each group perceiving the world around it and making a collective decision on what to do next. These perceptions and decisions are altered by other groups and individuals also perceiving the world and making decisions. The interactions between groups and individuals is exploding exponentially with no way to effectively estimate how many there are in the world. 

In addition to being collective, human intelligence creates another type of intelligence that is just as pervasive: embedded intelligence. Every human artifact you see around you is the result of human intelligence, whether collective or individual. Just think about all of the intelligence in the car you drive and the house you live in. Tentacles of intelligence span the globe and are embedded in the cities we live in and the products we use. 

Embedded intelligence is not a new line of inquiry. MIT scientist Cesar Hidalgo laid out a similar line of reasoning in his book Why Information Grows which was instrumental in Yuval Harari’s book Nexus. What Hidalgo refers to as information I refer to as intelligence. Hidalgo explains how networks of individuals link together to create solutions far beyond the capabilities of any one person. Harari takes the idea one step further and shows how these information networks he calls a nexus are used to create wisdom, truth, and power in the form of stories. According to Harari, the most powerful story ever told is religion. 

Embedded intelligence has proliferated throughout history and is a function of the number of people living in the world. From dams and airports to skyscrapers and satellites, embedded intelligence is woven into our day to day existence. 

On its own, human intelligence is powerful. Yet, it has been enhanced by technology. We are not talking AI (yet), we are talking regular old computers and networks. Together they have amplified the power of human intelligence in three ways. First, networks around the world have led to an enormous increase in the ability of humans to form into groups. Crisscrossing the globe, computer networks mesh intelligence together in tight or loosely coupled arrangements, ranging from vast global supply chains to connecting people together from all walks of life. Second, technology enhances the creativity process of human intelligence through the accumulation of knowledge and wisdom and putting it at our fingertips. From online databases to teaching courses to vast videos, information is simply a click away. And thirdly, technology provides access to the stories and information collective human intelligence disseminates. In this capacity, technology also acts as an access device. 

The other bucket of intelligence is artificial intelligence. For the first time in history, humans are attempting to create intelligence that rivals our own from scratch using silicon chips. This is a goal that has fascinated humans for centuries with early versions being purely mechanical. Efforts here have diverged into two broad categories; narrow or weak AI and general AI.

Narrow AI has captured all of the buzz lately. From ChatGPT and Grok to Claude and Gemini, narrow AI is the rage. And yet, there is nothing intelligent about them. All of these systems solve very specific problems called path problems that are well structured and where finding the solution to the problem is simply searching through the space of all possible solutions. 

OpenAI’s ChatGPT is designed as a conversational agent that responds to prompts, analyzes information, and can generate code. Under the hood though is nothing more than a model based on predicting text it learned via massive datasets called Large Language Models (LLMs). It takes advantage of the brute force calculation power of chips and combines it with a predictive text model loosely based on Bayesian reasoning. An algorithm in short. The key to past success for ChatGPT has been the massive amounts of data fed into it for training purposes. That diminished with ChatGPT 5 where even though its hallucination rates (incorrect answers) dropped by 26%, the improvement over ChatGPT 4 was incremental vs an exponential cost for the training. 

The other Achilles Heel of narrow AI is its inability to distinguish between cause and effect. It understands correlation instead of causation. While it will see a correlation between two related variables, it falls short in deciding which variable is dependent on the other or if both variables are dependent on some other variable. This shortcoming seems to suggest narrow AI models are an evolutionary dead end for developing human like reasoning, or general AI. 

So why the hype around AI if these models fall short of it? It all comes down to selling investors on the idea in exchange for funding. OpenAI has raised north of $100 billion for essentially developing a very sophisticated text prediction model. Without flashing the specter of general AI, it is doubtful that OpenAI would have generated anywhere near that much money. As it stands now, OpenAI is losing key partners such as Microsoft and Disney as the limitations of ChatGPT become more widely visible. 

General AI refers to the capacity of computers to construct a problem from scratch and solve it using any means necessary, including interdisciplinary thinking (thinking across multiple domains such as math, physics, economics, etc). It is here that progress has been slow. Rather than being driven by algorithms, general intelligence in humans is driven largely by experience, learnings, and heuristics. Heuristics are mental shortcuts and rules of thumb humans developed out of learning and experience. Heuristics were vital when ancient humans lived on the Savannah. Heuristics as shortcuts provided effective reasoning for a relatively low amount of calories. This conservation of calories was critical when calories were hard to come by. Since we are the descendants of those that survived over millions of years, those heuristics remain with us today (its one reason we instinctively fear spiders and snakes many of which were lethal back in the day. We are the survivors of those who learned to avoid them). 

To date, there is no discernible path on how to build general AI systems. One major stumbling block is that of requiring some form of worldview to reason with. As stated before, a worldview consists of models and heuristics that explain how the world works. It is essential for humans to navigate the world. Without one, we would not understand how to drive a car, cook dinner, interact with others. A worldview tells us how the world acts and how we should behave within it. It is constructed through our teachings and our experiences. All of our heuristics are based on some worldview. A worldview provides us with a strong understanding of cause and effect. Without that understanding, our actions become meaningless.

It seems very likely that general AI will require some sort of worldview if it is to achieve the sort of success at solving unstructured problems that are not susceptible to being solved by algorithms. Unstructured wicked problems such as income inequality, global terrorism, global warming, and nuclear proliferation require keen insight generated from multidimensional thinking, that is looking at the problem from multiple angles. There is no single simple solution waiting to be found by a clever search algorithm. Solving these problems will take a patient approach where experimentation is the key to finding solutions over time. 

We are now at the point where we can summarize where we are at. We live at a particular moment in time where human intelligence, both natural and embedded, is coming together with narrow artificial intelligence in a exponential exploding combination of forms.

Think about going to the bank. The interaction either online, at an ATM, with a teller masks a veritable web of connected intelligence of algorithms, data, analysis, and decision making to provide financial services and products to the general public. Over time, banks have exploded dramatically in their complexity, beginning with loans in the form of grain in the Middle and Far East in roughly 3000 BCE, to modern banking structures housing entities that span a wide variety of financial sectors both domestic and global, with hooks into derivatives, EFTs, traditional and crypto currency. 

The Cambrian explosion of intelligence woven into the fabric of everyday life that is at the heart of Frankenstein’s World will increase exponentially in the next few years. As more embedded, human, and artificial intelligence becomes available, it will combine with existing intelligence to create new collective groups of intelligence which will network with other groups. Much like the way Frankenstein was made from spare body parts, we too will be creating monstrous complexities from the intelligence we fuse together.