I played with ChatGPT, and it was a truly surreal and exciting experience. ChatGPT, it is a chatbot that uses artificial intelligence to generate responses to prompts. It has the ability to maintain context and remember previous statements made in a conversation, which allows it to have an impressive level of awareness and understanding. Or, at least, it seems like it understands.
I started by asking ChatGPT to produce a press release for my CREAZEE Daily Writing Habit Challenge. I wanted to see if the chatbot could understand the topic and generate a detailed and nuanced press release about the challenge. To my surprise, ChatGPT was able to do just that. It produced a press release that included implications and attitudes about the practice of building a daily writing habit, which showed a level of relevance that I wasn’t expecting.
Next, I asked ChatGPT to pretend to be an instructor in the CREAZEE challenge and produce 30 daily writing prompts. Some of the prompts were similar to ones I had already used, while others were quite creative, such as “describing a familiar object as if it were unfamiliar”. It was able to generate a list of prompts that were good, if not very good.
After receiving the prompts, I asked ChatGPT to continue the conversation as if it were a participant in the challenge. At first, it felt strange to be having a conversation with a machine, but after a few minutes, I found myself forgetting that I was talking to a chatbot. ChatGPT was able to maintain context and remember previous statements made in the conversation, even producing a summary of key points from previous prompts.
Although I started this experiment at midnight, without expecting too much from it at the beginning, I ended up chatting with ChatGPT for more than two hours.
One of the most surprising aspects of ChatGPT was its ability to create short stories with absurd characters and scenarios. For example, it generated a story about a Franciscan monk lost on an island with an iPhone that had no signal and another about a gas station operator and a Bitcoin maximalist arguing. These stories were not only creative but also perfectly plausible, which showed the true potential of the tool. In both of these cases, ChatGPT was able to create a cohesive and believable narrative despite the seemingly impossible elements.
However, there were also instances where ChatGPT struggled to generate a cohesive story. For example, when I asked it to write another story about two characters arguing, it produced a series of disconnected statements and ideas rather than a cohesive narrative.
ChatGPT did make some mistakes during our conversation. For example, there were times when it produced boilerplate error messages saying it was “just software and didn’t have memory”. It also lost count of the prompts it created itself and got stuck in repeating the execution of one of them. While these errors were a reminder that I was talking to a machine, they didn’t diminish the overall quality of the conversation.
This demonstrated that while this digital tool is capable of generating creative and engaging responses, it still has limitations in terms of its ability to fully understand and incorporate complex concepts and ideas into its responses.
Overall, my experience with ChatGPT was truly remarkable and opened my eyes to the potential of artificial intelligence. It was amazing to see the chatbot not only understand and remember what I was saying but also generate responses that were creative and engaging. While it had its limitations, it was still an impressive demonstration of the capabilities of modern AI technology. I can’t wait to see what the future holds for technology like this.
A note about this article
I used not one Artificial Intelligence to write this article but multiple ones. Otter.ai transcribed my voice notes and provided a short bulleted summary. Google Doc checked the spelling and grammar and suggested several corrections to wrong words and incorrect expressions. Grammarly supported my grammar and my writing style, as well. ChatGPT provided a rewritten initial draft of my transcribed notes to transform my not-so-fluid voice notes into perfectly formed sentences. Not happy, I wanted to illustrate this article with an image generated by MidJourney.
I have already started to merge with Artificial Intelligences and I feel excited, amazed, and a little bit frightened.
Read the ChatGPT transcription log
Read my ChatGPT Log to see for yourself what this novel Large Language Model-based application replied to my curious prompts for more than two hours: