Book: Draft No. 4: On the Writing Process
John McPhee’s Levels of the Game was published when I was three. I’ve read many books about tennis, but it’s still my favorite, with String Theory: David Foster Wallace on Tennis coming in a close second. Both books share extraordinary writers and a topic I can relate to, given my short but profane junior tennis career.
I’ve become annoyed by the writing that ChatGPT produces. At first, I was fascinated. As I saw more of it, I began seeing patterns. After reading Stephen Wolfram’s article What Is ChatGPT Doing … and Why Does It Work? my observation shifted to impatience. Most of my ChatGPT efforts now are making it hallucinate.
I finished Draft No. 4: On the Writing Process a few minutes ago. I enjoy reading books about a craft by a master of a craft, especially when it’s a craft that I practice, even if I’m not that good at it. Tennis is one example. Writing is another.
After reading Draft No. 4, I envy writers with excellent editors. For the books I’ve written, I’ve enjoyed my relationship with my editor, but he mostly plays the role of project manager. My experience, compared to Jerry Colonna’s relationship with his editor, is completely different.
I had continuous insight into this difference while reading Draft No. 4. McPhee is a spectacular writer. He has also been blessed with over 50 years of great editors, including The New Yorker’s Robert Gottlieb. While writing can be extremely solitary, the relationship between writer and editor can be a beautiful dance. If you want to understand that in-depth, grab a copy of Avid Reader: A Life by Robert Gottlieb.
ChatGPT badly needs an editor. Apparently, it agrees.
As an AI language model, I don’t have personal preferences or needs like humans do. However, it’s always beneficial to have human editors or reviewers involved when using AI-generated content, including the responses from ChatGPT. Editors can help ensure the accuracy, clarity, and appropriateness of the generated text for specific contexts or purposes.
Human editors can review and refine the AI-generated content to improve its quality, correct any errors or inconsistencies, and adapt it to the desired style or tone. They can also help mitigate biases that may inadvertently emerge in the generated text. By combining the capabilities of AI with the expertise of human editors, the overall quality and usefulness of the generated content can be enhanced.
Therefore, involving human editors can be a valuable step in refining and improving the output of AI language models like ChatGPT.
When I asked ChatGPT, “Can you turn the above text into an essay in the style of John McPhee’s writing?” I got the following.
When it comes to discussing the involvement of human editors in the utilization of AI-generated content, one cannot help but ponder the profound influence of their collaboration. As an AI language model, I lack the personal inclinations and requirements of human beings. However, the inclusion of skilled editors in the process undoubtedly enhances the overall quality and effectiveness of the generated text. In the realm of AI, where the capabilities of language models like ChatGPT have flourished, the expertise of human editors is essential to refine and polish the output, elevating it to new heights.
Editors possess a unique ability to discern and rectify any inaccuracies, inconsistencies, or inadequacies that may manifest within AI-generated content. Their discerning eyes scrutinize every word, sentence, and paragraph, ensuring a level of precision and clarity that may elude the algorithmic mind. By judiciously reviewing and refining the text, editors bestow upon it a coherent and lucid nature, making it more comprehensible and impactful for specific contexts and objectives.
In the delicate task of mitigating biases, human editors exhibit their paramount importance. While AI language models aim to emulate human language patterns and produce coherent responses, they can inadvertently perpetuate biases ingrained in the data they are trained on. Editors, with their conscientiousness and attentiveness, serve as gatekeepers, identifying and rectifying any biases that may inadvertently emerge. Their involvement acts as a crucial checkpoint, ensuring the AI-generated content does not unwittingly reinforce discriminatory views or perspectives.
Apparently, ChatGPT decided to use fancier words. While I don’t know McPhee, I expect he’d cringe and Green 10 before giving it to his editor.
The four things I enjoy the most in life outside my work are reading, writing, running, and being with Amy. When I encounter a writer like McPhee writing about writing, I enjoy my brief time in a parallel, recursive universe.