Cogito Ergo Creo

I think, therefore I create.

The advent of Generative Pretrained Transformer (GPT) models has brought a fundamental shift in the way we think and create. And in direct contrast to what many “thinkers” have proclaimed, GPT has is becoming an instrumental in elevating human cognition to new heights, allowing us to think better, deeper, and more creatively. With its advanced language processing capabilities, GPT is driving a revolution in the way we approach tasks and solve problems, making it possible for us to create at a level never before seen in human history. It’s the inflection point that many have been looking for in many areas of innovation. But it’s surprisingly emerged in a way that is technologically advanced, yet striking accessible and useful to many people

GPT has been trained on vast amounts of text data, which has enabled it to develop a deep understanding of language and its nuances. This understanding has enabled GPT to generate highly effective prompts that can spark new ideas and thoughts in its users. These prompts can take the form of questions, statements, or even creative prompts that challenge us to think in new and innovative ways. The output is less a function of craft (think paintbrush) but much more a function of thought or prompt that drives the computation.

With GPT’s help, we can now explore new ideas and concepts that were previously beyond the reach of many. Its ability to generate new and unique perspectives has opened up new avenues of thought and created new avenues for creativity. Whether we’re writing a story, composing a song, or developing a new product, GPT can help us tap into our inner creativity and unleash our full potential. Or perhaps even foster a new and expanded capacity that didn’t exist prior to GPT augmentation.

While the idea of thought in the context of AI an GPT remains both controversial and even inflammatory, new data pushes the boundaries of understanding the very nature of GPT and its potential. An interesting paper looks GPT’s “ability to understate” the mental state of someone. This study uses the “theory of mind” task—a type of psychological test that assesses an individual’s ability to understand the mental states of others. The term “theory of mind” refers to a person’s understanding of the thoughts, beliefs, and emotions of others, and the way that these mental states can influence behavior.

This study administered tasks, commonly used to assess theory of mind (ToM), to various language models without any pre-training or examples. The results showed that language models developed prior to 2022 had virtually no ability to solve ToM tasks, while the January 2022 version of GPT-3 (davinci-002) performed at a level comparable to that of seven-year-old children (70% accuracy). The November 2022 version of GPT-3 (davinci-003) performed even better, with accuracy comparable to that of nine-year-old children (93%). These findings suggest that ToM-like abilities, previously thought to be exclusively human, may have arisen as a result of language models’ improving language skills.

GPT is driving a new era of human cognition, one in which we can think better, deeper, and more creatively than ever before. With its advanced language processing capabilities, GPT is helping us create at levels never before seen in human history, elevating us to new heights and establishing a new perspective. From “Cogito Ergo Sum” to “Cogito Ergo Creo” — I think, therefore I create!

Thinking is at the core of this new dynamic. The question that will emerge is just who (or what) is doing the thinking…