Chapter 1: Understanding Creativity

The world of AI artists and their creations is expanding rapidly. (Source: AIArtists.org) Our enduring affinity for classic music poses significant questions about the essence of artificial creativity.

Every so often, a notion returns with the reliability of a comet: the idea that we are now in the era of creative AI. Proponents argue that machines are not just expediting discoveries; they are independently making them. This is particularly timely given the increasing complexity of the world, which may be beyond human comprehension.

However, this perspective has been flawed for centuries. Back in the 17th century, Francis Bacon suggested that our salvation lay in an autonomous scientific method that derives knowledge from observations. He was mistaken, but his view still captivates many today. This misconception echoed in 1900 when Lord Kelvin claimed that physics had nothing new to unveil, and again when John Horgan popularized this "wrong idea" in his book, The End of Science in 1996. Even in 2008, Chris Anderson celebrated what he called the end of theory due to big data. This notion resurfaced once more with discussions around the limits of physics and remarkable advancements in automated discovery.

Yet again, this idea is fundamentally incorrect.

Instead of focusing on this error, I want to highlight why creativity is an inherently human trait. It’s essential to differentiate between applied creativity and fundamental creativity, as this distinction sheds light on the specific area where AI may soon excel.

Chapter 2: The Challenge of Fundamental Creativity

Joe Rogan Experience #1309 - Naval Ravikant – This episode delves into the nature of innovation and creativity, discussing the limits and potential of AI.

As we consider the difficulty of achieving fundamental creativity, we think of figures like Einstein, Austen, and Picasso—those whose groundbreaking work typically aligns with the promise of Artificial General Intelligence (AGI). These creations not only introduce novelty but also compel us to reassess the boundaries of creativity. Such fundamental innovations challenge existing evaluative criteria, which often remain undefined prior to their emergence.

To better understand this dilemma, let’s turn to music. Imagine all the melodies that could be crafted using a musical scale. As notes are combined in various ways, the possibilities rapidly grow into millions or even trillions. The issue of creativity is often mistakenly simplified to a problem of scale, leading to the assumption that greater data and computational power will eventually resolve the challenge. Yet, the scale of a problem is not the core issue of creativity.

If one aims to create an entirely new genre of music, how can they determine, beforehand, the new criteria by which an original melody might be evaluated? Moreover, when striving for something fundamentally innovative, how can one establish the new evaluative standards that distinguish a masterpiece from a mere random collection of sounds?

A classic example of this conundrum is the diabolus in musica (Latin for "the Devil in music"), which refers to a dissonant interval known as a tritone. While you may recognize it from songs like "Enter Sandman," "Maria," or "Purple Haze," these melodies would have been rejected during Medieval times—not due to their absence in the soundscape but because the necessary evaluative criteria hadn’t been established.

This illustration of classical music and the tritone serves as a small glimpse into the vast challenge of creativity, which is a dynamic and emergent process.

Reflect on your fondness for classic music. Have you ever pondered why we cherish the songs from our youth? Our evaluative standards tend to solidify during adolescence and early adulthood. As a result, we often perceive contemporary music as inferior to that of previous generations. The music we love encapsulates the criteria we hold dear, while new music embodies fresh evaluative standards, often disregarding the old.

You could even create a system that "proves" the superiority of older music by selecting attributes that favor past criteria. It seems today’s songs feature less melodic complexity, are more repetitive, have fewer chord changes, and exhibit narrower dynamic ranges. Furthermore, today’s artists may appear overly self-absorbed. However, the quality of music and the criteria for evaluation cannot solely be defined by these characteristics.

Such analyses often overlook the new and largely implicit criteria that may be active within our cultural communities. Misunderstandings surrounding the nature of scientific arguments are as prevalent as misconceptions about creativity.

I find this cultural barrier to creativity to be quite intriguing. I wonder how many brilliant ideas have been conceived but ultimately rejected because they did not conform to the prevailing cultural norms. Artists working in isolation often internalize the creative limitations of their communities, becoming their own harshest critics. The crumpled papers in their trash cans serve as evidence of this process.

The real challenge lies not so much in the creative variations (such as having your dog play the piano) but in the means of selection. This same environmental challenge also constrains the effectiveness of evolutionary approaches to creativity.

In reality, the intricate interplay of psychological and social dynamics that fuels creativity remains largely enigmatic. It is both mysterious and beautiful. At the right moment, when a creative mind intersects with the unpredictable nature of their social context, something truly novel and significant can arise.

This will occur again. A brilliant individual, operating within the right community at the right time, will unravel the complexities of creativity. We’ll know they’ve succeeded because they’ll have the ability to articulate their insights.

Soon after, AI may also achieve true creativity.