Artificial General Intelligence — but why?

Ramesh Kumar Katreddi
10 min readMay 26, 2024

--

Over the past few months, I have had many conversations with people regarding Artificial General Intelligence (AGI) and have concluded that different people have different interpretations of what it means. For some, it is the robots in Hollywood science fiction movies. For others, it's understanding how our brain works, creating a system that can function similarly, among many other interpretations.

There can be multiple definitions, and all of them could be right because it is a theoretical concept that suddenly seemed possible in the decades of this century. For now, you can assume any definition of AGI— as an example, a robot that is almost an identical replica of you in terms of thinking.

The Internet is the foundation for all the innovations we see right now. While we could not have imagined these possibilities, and they might have looked like far-fetched ideas a couple of decades ago, the Internet’s core was solving the problem of everyone having access to information so that everyone could make informed decisions.

From a problem solver’s point of view, AGI is a solution or perceived destination, but what is the problem? Why do we even want a system to be Sentient? While we can say it will bring intelligence to the systems and open up opportunities for countless issues, defining what intelligence means and the real problems this technology is supposed to solve in the short and long term will set the right expectations for the users on what to expect and carve out the precise direction in which it will evolve.

I am not saying issues are not identified; I am sure the technology frontiers would know it to the granularity I can't even imagine. However, it is not generally available knowledge, and most of the discussions I see are on solutions rather than problems, thus leading to speculations, fear, or anxiety.

I started thinking about this topic more deeply and began with the first-principles-thinking to explore the problem. In this series of articles, I aim to understand it better and provide an opinion on what are the limits of human intelligence, what issues we should ideally solve, what would be an ideal definition of AGI, how far we are from it, and whether it is the destination we are even chasing, and better to understand the prospects beyond the LLM euphoria we are in.

I plan to structure this exploration in the following six articles, starting with understanding human intelligence and gradually moving toward Artificial General Intelligence.

  • The Premise: The building blocks of “Intelligence” and especially “Human intelligence.” (Discussed in this same post)
  • The Problems: Identify challenges/limitations in each building block of human intelligence and its impact.
  • The Solve: What are the solutions for the above, and is AGI the ultimate solution for those problems?
  • The Path: If AGI is solving them, where does it stand now? Which building blocks are solved, and which must be solved to reach AGI?
  • The Destination: How do we know if we have reached AGI? What is the comparison? Is it a Turing test? Is human intelligence the minimum bar for AGI or the maximum limit we want to impose for safety?
  • The Future: Where is the future headed, and what are the opportunities?

This article represents my views, and I plan to mature the content as we dig deeper. The article is forward-looking and intends to help us better understand the problems, the path, and the future. If you are looking for knowledge about quick solutions to implement in your business at the current time, this might not be helpful.

Please feel free to share your constructive criticism.

Let's dive in!!!

The premise: Building Blocks of Human Intelligence

As a first step on our learning journey towards AGI, we must familiarise ourselves with how human intelligence works.

Let us focus on what “intelligence” means and precisely what “human intelligence” means.

  • Following are Wikipedia's definitions of Intelligence, Human Intelligence, and Cognition.

Intelligence has been defined in many ways: the capacity for abstraction, logic, understanding, self-awareness, learning, emotional knowledge, reasoning, planning, creativity, critical thinking, and problem-solving. It can be described as the ability to perceive or infer information; and to retain it as knowledge to be applied to adaptive behaviors within an environment or context.[1]

Human intelligence is the intellectual power of humans, which is marked by complex cognitive feats and high levels of motivation and self-awareness.[22] Intelligence enables humans to remember descriptions of things and use those descriptions in future behaviors. It gives humans the cognitive abilities to learn, form concepts, understand, and reason, including the capacities to recognize patterns, innovate, plan, solve problems, and employ language to communicate.

Cognition is the “mental action or process of acquiring knowledge and understanding through thought, experience, and the senses”.[2] It encompasses all aspects of intellectual functions and processes such as: perception, attention, thought, imagination, intelligence, the formation of knowledge, memory and working memory, judgment and evaluation, reasoning and computation, problem-solving and decision-making, comprehension and production of language. Cognitive processes use existing knowledge to discover new knowledge.

Wow, there are a lot of nouns, adjectives, and keywords. You would have glanced over it or tried to understand it, but it looked like many discrete topics that were coherent but unrelated were kept together.

Let's structure them to make sense of them and form the building blocks of human intelligence. To improve the definitions, I chose to segregate all the critical elements described in the definition into logical building blocks. These are my interpretations of understanding intelligence.

  • Knowledge and comprehension.
  • Problem-solving and Application
  • Cognition

Let us understand each of these building blocks using the following Image. It's not a fancy picture, but I believe when we are talking about such universally accepted paradigms, we should be able to use any image and be able to explain; I just opened my camera and clicked what was in front of me :) Someday, I hope this picture will become “The historic balcony picture” ;)

Observe the details of the picture for 10–15 seconds before jumping to read the further sections.

To understand intelligence, let's individually go through each building block explained in the above diagram.

Knowledge and Comprehension.

  • Perceive or Infer Information — Through our sensory organs, which can process information in front of it selectively, we infer and perceive the information.

I am sitting on the sofa. It feels soft. My eyes are seeing tall structures in front of me with varying heights and colors. I feel the air coming through the door, and I see the cloth hanging from movable rods in front of me moving. I also notice the leaves of the plant moving due to the air.

(Read the above sentence carefully twice to ensure you follow the further explanations)

  • Learning and forming concepts — This is a process of acquiring knowledge.

The tall structures are called “Buildings”. The clothe covering balcony is called “Curtains”

  • Remember things

The above is only one instance of the information our sensory organs capture. Every millisecond, we continue to capture thousands of such cases through our eyes, noses, and ears. Say the ear transmitted the below information to the brain one minute before the above observation.

Hey I am finding it cold.

  • Convert into knowledge

It's essential to understand the difference between Information and Knowledge.

Information: The raw sensory data I was experiencing and describing so far.

Knowledge: The interpretation, understanding, and application of that information by combining it with the existing information to make decisions or gain insights.

Example:

Since curtains are open and air is coming, I am feeling cold. I think they are related. I will close the door.

Problem-solving and Application

In the above example, we saw a form of Deductive Reasoning.

Air is coming towards me.

Air can cause body to feel cold.

I am feeling cold.

I deduce that I am feeling cold because of air

Reason is the capacity to apply logic consciously by concluding new or existing information to seek the truth. The reason need not be correct.

I saw the large structure, it is called called building

I see similar structure, i think that is also called building

(This is called Anological Reasoning)

Let us one more example.

I observe that the leaves of the plant are moving because of the air.

The air is touching the curtains before reaching me.

Thus, the air has the capability to move objects.

Thus I deduce that The curtains are moving because of the air.

(one more example of deductive reasoning)

We do this throughout the day in the process of acquiring knowledge (as seen in the above section) and applying it to solve a problem. This is the most critical aspect of our functioning.

  • Critical thinking — focuses on analyzing and evaluating information, evidence, or arguments to form a judgment or decision.

This is a crucial part of getting our knowledge and application right. So far, we have seen an intuitive process of how it is easy to consume information, reason with it, and convert it into knowledge, but our ability to reason can lead to biases. (Check out the link to learn more about biases)

Example of conformity bias —

I already decided in my mind that air coming through curtains is causing the cold.

I find evidence to support it and I find the reason like the above.

That is where our questioning abilities and information evaluation capabilities by deep questioning come into the foray.

(continuing the above example of curtains moving due to air).

Is air the only reason the curtains are moving?

If Rods are movable, Could someone be moving the rods?

If air can move objects, why am I not moving?

After these questions, more reasoning might emerge, all of which might seem plausible. That is where the next topic comes in.

  • Logic — One of the most significant lessons learned through this process is that Reasoning comes before logic. Logic is nothing but the correct reasoning.

So far, we have captured a lot of information and tried to reason with it. Using our critical thinking, we generated more reasons, and now the time has come to decide which one is true.

Logic works with certain premises and conclusions. Premises and conclusions express propositions or claims that can be true or false.

Premises:

The leaves of the plant are moving.
The air is touching the curtains before reaching you.
The movement of the leaves indicates that air has the capability to move objects.

Conclusion:

The curtains are moving because of the air.

If each of the Premises is true, the conclusion would be proper. The only thing to evaluate is whether we need to include any premises.

The logic would look as follows.

I do not see anyone in the picture touching the rod, thus it might not be the reason.

I am heavy, thus I am not moving.

  • Recognize the patterns

With the knowledge acquired through the information and reasoning, we continue to evaluate any new information we receive compared to the existing information we possess to form patterns.

Leaves are moving due to air, so is curtain — both of them are lightweight and are in the path of air flow.

  • Abstraction — Abstraction generalizes concepts by removing specific details from instances. This is an essential aspect of human intelligence

When a light weighted object is in the path of airflow it moves.

  • Planning

I want to reduce the airflow into the room and I plan to replace the curtains. I need to find the curtain shop. I need to find the material. I need to find measurements. I should install it.

Cognition

I am conscious of the softness of the sofa, the sight of the tall buildings, and the feeling of the air coming through the door.

I am aware that someone observing me might think my actions are unusual.

  • Emotional knowledge

I feel calm seeing this balcony, but also sad when reminded of the trees cut down to build these structures.

  • Self-awarenessself-awareness is the experience of one’s personality or individuality.

I am aware that I am the one perceiving these experiences and that I am sitting on the sofa, reflecting on my surroundings. I do not like the windows obstructuring my view. I have certain likes, dislikes, choices, preferences.

  • Imagination and Creativity (So far, everything above is about applying the knowledge, but this focuses on the generation of expertise)

I want to add ten more pots and cover the balcony with greenery and improve the asthetics

  • Adaptive behaviors with the environment

My flatmate does like to have tea in this balcony every morning and I cannot fill this with pots completely. He might not find this idea welcoming

I have pets which might be alergetic to some pots and i should not keep them here

Beyond this, Humans have physical abilities that, combined with this intelligence, make us much more powerful.

  • Our sensory organs
  • Manual dexterity
  • Our mobility and navigation
  • Coordination between our body and mind.
  • Our ability to communicate using language

I hope this provides an overview of intelligence. You can refer to the original Wikipedia definitions described in the first section again. This is an essential first step in exploring further topics.

Each of us has a different level of expertise in each of the building blocks of intelligence, making us unique. If a machine could be at the maximum capability level across all the building blocks, would it be the Singularity? We will find it in further articles.

In the following articles, we will use the above building blocks to understand the challenges of human intelligence and make roads to understanding AGI.

--

--

Ramesh Kumar Katreddi

Passionate problem solver with expertise in building large scale systems and mentoring ownership driven technology teams. Forbes India Top 30 Technology leader