Experimenting with an AI writing partner

Winkletter  •  9 Dec 2022   •    
Screenshot

I’ve published an experimental article that was partly based on a conversation I had with ChatGPT. Setbacks Are Essential Data started as an outline I created on paper, but I then asked the AI some questions based on my outline and used its responses to help fill out my article.

  • Is failure an essential part of success?
  • What are some strategies I can use to not feel discouraged by failure?
  • Do you know any authors who have written on this topic?
  • This also reminds me of Nassim Taleb’s book Antifragile. How can I create an antifragile system that benefits from failure?
  • I’m trying to think of a rival viewpoint about failure that sees it as less valuable. For example, experimenters often don’t publish negative results. Can you provide some reasons why they might prefer to not publish negative results of an experiment?
  • Can you extrapolate from the reasons above why an individual might devalue failure when trying to accomplish a personal goal?
  • What are some of the emotional needs that someone is attempting to satisfy when trying to accomplish a personal goal?
  • Can you list some examples of goals an individual might attempt to accomplish where they will encounter a high rate of failure before eventually succeeding?
  • How about some more examples that someone might attempt as a New Year’s resolution?

Overall, I found it increased the breadth of the article. I didn’t save any time writing the article since I didn’t just copy and paste the results. But after this experience I almost feel like I would be doing the reader a disservice not to use ChatGPT for brainstorming.

Where I might have one or two points to make, the AI had four or five and could easily find ten more. So the article may have too much information. Usually in an article I try to focus in on one viewpoint and an emotional need they are attempting to satisfy, but the AI gave me four different emotional needs and they all seemed important to explore.

My final section is about four times as long as I usually try to make them.

Also, it took a lot of work to pull everything together into a stream of thought. I’m worried it might still be too disjointed and inconsistent in tone. If it is, I will likely see a low read through rate.

I’ll have to see how this article performs compared to others I’ve published recently.

Comments


Discover more

Sourced from other writers across Lifelog

Ooops we couldn't find any related post...