How the Study of Birds in Aviation History Shapes AGI Development

Artificial general intelligence (AGI) will possess human-like intelligence capable of understanding, learning, and applying knowledge across a wide range of tasks and domains. An AGI system would be able to transfer learning from one area to another, reason abstractly, and solve unfamiliar problems without specific prior training–much like a human can.

It’s important to differentiate AGI from artificial intelligence, which refers to systems designed to perform specific tasks or a narrow range of functions. These systems are typically specialized and excel in their designated areas but lack broader understanding or adaptability outside their training domains. Examples include: 

• large language models
• image recognition software
• language translation tools
• chess-playing programs

In essence, current AI is narrow and specialized, while AGI would be broad and adaptable, mirroring or surpassing human cognitive abilities across the board.

Artificial general intelligence will be the most important scientific discovery in the history of humanity.

Much attention today is being given to large language models (LLMs) as a beachhead towards AGI. The consensus is that LLMs will naturally give birth to AGI. But this is unfortunately not the case. In an essay for the New York Times, Dr. Jeffrey Watumull, along with Dr. Noam Chomsky and Dr. Roberts posit that, unlike LLMS that are essentially prediction engines, “The human mind is a surprisingly efficient and even elegant system that operates with small amounts of information; it seeks not to infer brute correlations among data points but to create explanations.”

The Wright way to invention

If LLMs will not birth AGI, what will? 

The study of nature provides a significant source of inspiration, leading to technological innovations, medical breakthroughs, and a deeper understanding of the world around us. Indeed, scientific advancements often arise from the study of nature as researchers observe natural phenomena, analyze patterns, and develop theories to explain them. 

The development of flight by the Wright Brothers is a prime example of how scientific advancements can be derived from the study of nature. Their successful creation of the first powered, controlled airplane was deeply influenced by their observations and understanding of natural phenomena.

And it’s not as though the Wright Brothers simply saw birds flying and said, “Hey, let’s do that.” No, they meticulously observed how birds angled their wings for lift and maneuverability, noting the subtle adjustments birds made to control their direction and altitude. This led to the crucial realization that a successful aircraft would need a similar system to achieve balance and control. Their method of wing-warping, which allowed for the controlled twisting of the wings, was directly inspired by the way birds manipulated their wing shape to navigate the air.

Further, a significant influence on the Wright brothers was Otto Lilienthal, a German aviation pioneer who had extensively studied the flight of birds. Lilienthal’s experiments with gliders and his published works on aerodynamic mechanisms provided valuable insights into the principles of lift and control. 

The Wright brothers were particularly obsessed by his concept that curved surfaces could generate lift more effectively than flat ones, which he demonstrated through his numerous glider flights. Lilienthal’s tragic death in a gliding accident underscored the risks involved, but his work remained a crucial reference for the Wright brothers as they pursued their own experiments.

Building on Lilienthal’s theories, the Wright brothers conducted their own experiments with gliders, refining their designs through rigorous testing and observation. Indeed, they developed a wind tunnel to test different wing shapes and configurations, which allowed them to gather precise data on aerodynamic forces. 

By combining their empirical research with insights gained from observing birds and studying aerodynamic mechanisms from Lilienthal’s findings, they overcame many of the challenges that had stymied earlier aviation efforts. Their breakthrough in controlled flight, achieved through their innovative approach to wing-warping and a practical propulsion system, was a testament to their ability to synthesize natural observations with scientific experimentation. 

This blend of nature-inspired design and rigorous engineering ultimately led to the development of a practical and controllable aircraft, marking a significant milestone in the history of aviation.

Mimicking nature

Developing a synthetic and fully capable AGI must involve leveraging the process of learning from nature, applying scientific principles, and implementing technological innovations, much like the Wright Brothers did for flight. 

They mimicked the bird’s aerodynamic mechanisms and we need to do the same things by mimicking cognitive mechanisms of the human mind to develop AGI. 

By observing natural intelligence, applying cognitive principles, and leveraging technological innovations, the process of developing AGI can be systematically approached. 

For centuries flying machines were impossible, unreal, and unsafe. Yet the Wright Brothers brought us to the aviation era by the study of avian aviation. Similarly for almost seven decades, AGI has been science fiction. Unreal and unsafe. In the last few years that zeitgeist has shifted and we now believe it to be within the realm of near term possibility.

LLMs have inherent limitations that will prevent them from birthing AGI. And thus we should seize the moment to expand upon other scientific avenues to achieve AGI.

Powered by our ability to mimic the cognitive mechanisms of the human mind, we have a real shot.