Mastering the Art of AI Collaboration: A Journey from Hilarious Hallucinations to Harmonious Dialogues with ChatGPT
Think in terms of integrating man and machine, not man instructing machine
ChatGPT is an obsequious little servant. It responds dutifully and almost immediately to any prompt you throw at it, sometimes to hilarious or hallucinatory effect. And when either hilarity or hallucination strike, many people get frustrated with it, conclude that it is overhyped, and give up.
But there’s a better way! I wrote about it here. This strategy entails telling ChatGPT to ask you clarifying questions about your prompt before it executes the instructions contained in your prompt. To make this more concrete, this strategy uses the following sequence of events:
Initialize your session by telling ChatGPT that you have a project for it, but that before it executes your instructions, it is to ask you clarifying questions;
ChatGPT acknowledges your initial instructions;
Provide the prompt you want it to use to generate output;
ChatGPT asks clarifying questions;
Provide answers to its clarifying questions;
ChatGPT EITHER generates its output OR asks further clarifying questions.
In other words, you have to conceive of using ChatGPT as the integration of man and machine, or silicon and carbon, rather than man merely instructing machine.
You can read Ethan Mollick’s review of his students’ struggles to get good output from ChatGPT here. He provide a lot of valuable tips for getting it to generate good output. I think that combining his suggestions with clarifying questions is even more powerful.
I relate all of this because this morning I decided to try an experiment. I wanted to give ChatGPT the prompt to write a blog post, and I wanted to influence its output with two paragraphs of text which I thought interesting and relevant. The text which I thought to be interesting comes from this blog post about AI pessimism.
What follows below the section break is my interaction with ChatGPT and the blog post it wrote. Note that in the prompt I provide below, two paragraphs, contained within quotation marks, are sourced from the blog post I link to in the preceding paragraph. My prompts are preceded by Prompt and ChatGPT’s output is preceded by Output.
Prompt: I have a complicated project for you. Before you execute my instructions, I want you to ask me clarifying questions. Do you understand?
Output: Yes, I understand that you have a complicated project for me, and you would like me to ask clarifying questions before executing your instructions. Please provide details about the project and any instructions you have in mind, and I will ask questions to ensure I fully understand your requirements.
Prompt: Write a blog post about how the rise of AI affects human endeavors like chess and art. The blog post should be aimed at an educated, generalist audience. Use an active voice and provide relevant examples. Make sure the sentence structure is varied, word choice is varied, avoid cliches. Provide some information about what it means to be human and what the ineffability of humanity implies relative to that of AI. The post should be between 750 and 1000 words. Use this text as context for the post:
"The loss of the 1500-year-old tradition of chess kibitzing sessions may appear trivial. But the greater danger is that similar dynamics will start to manifest in art. The currency of chess was always understood to be intelligence—chess players have taken the advent of a more “intelligent” chess-playing entity with relatively good grace. But in art, the currency isn’t really excellence, it’s the ability to express one’s truth. The musician Nick Cave put this well in a recent diatribe. When a well-meaning fan sent Cave an AI-generated song “in the style of Nick Cave,” Cave reacted badly. “Songs arise out of suffering and as far as I know algorithms don’t feel,” Cave wrote. “Thanks for the song, but with all the love and respect in the world, this song is bullshit, a grotesque mockery of what it is to be human.”
For me, this is the line in the sand to draw. When I sit down to play a game of chess, I do so not because I’m seriously hoping to play at the heightened, god-like level of the human-machine hybrids; I’m playing because I enjoy the process of playing chess and because I’m looking to surprise myself by playing a little better than I played last time. In other words, the enjoyment of the game begins only once the gadget is put away. Similarly, if I sit down to write something, I really couldn’t care less if it’s at some heightened, world-beating level. What I do care about is if it reflects the truth of my very human experience."
Output: I understand that you would like me to write a blog post about the impact of AI on human endeavors, specifically in the areas of chess and art. Before I begin, I have a few clarifying questions to make sure I understand your requirements:
Are there any specific AI-related advancements or milestones in chess and art that you would like me to include in the blog post?
You mentioned providing information about what it means to be human and the ineffability of humanity. Should I discuss this concept throughout the post or focus on it in a specific section?
Do you have any preferences for the structure of the blog post, such as starting with a general introduction to AI, followed by its effects on chess and art, and then discussing the implications for humanity?
In addition to Nick Cave's quote, are there any other examples or quotes from notable figures that you would like me to incorporate?
Once I have your answers to these questions, I will be better prepared to write the blog post according to your specifications.
Prompt:
AI-related advancements in chess: IBM Watson's chess engine beating Garry Kasparov. AI-related advancements in art: large language models being used to generate images, for example, DALL-E.
Discuss the concept throughout the post. The goal is to understand how the ineffable qualities of humanity are informed by and affected by the rise of AI in fields historically dominate by humans.
No--I'm leaving that up to you.
No.
Output: Thank you for your responses. With the provided information and your clarifications, I will now write a blog post about the impact of AI on human endeavors, focusing on chess and art, discussing the ineffable qualities of humanity throughout the post, and incorporating relevant examples such as IBM Watson's victory over Garry Kasparov and DALL-E generating images. I will also use the provided text as context for the post. Here is the blog post:
Title: The Ineffable Human Touch in an AI-Dominated World
In a world increasingly infiltrated by artificial intelligence (AI), we must examine the role it plays in human endeavors, such as chess and art. While AI has made remarkable strides in these fields, it has also raised questions about the ineffable qualities of humanity and our ever-evolving relationship with technology.
In 1997, IBM's Deep Blue chess engine defeated the reigning world champion, Garry Kasparov, marking a turning point in the world of chess. The loss of the 1500-year-old tradition of chess kibitzing sessions may appear trivial, but the greater danger is that similar dynamics will start to manifest in art. The currency of chess was always understood to be intelligence—chess players have taken the advent of a more “intelligent” chess-playing entity with relatively good grace.
But in art, the currency isn’t really excellence; it’s the ability to express one’s truth. The musician Nick Cave put this well in a recent diatribe. When a well-meaning fan sent Cave an AI-generated song “in the style of Nick Cave,” Cave reacted badly. “Songs arise out of suffering and as far as I know algorithms don’t feel,” Cave wrote. “Thanks for the song, but with all the love and respect in the world, this song is bullshit, a grotesque mockery of what it is to be human.”
As AI progresses, it continues to infiltrate the art world, with large language models like DALL-E generating realistic images. But can a machine-generated image truly capture the human spirit, the emotions and experiences that make us who we are? The ineffable quality of human art lies in the depths of our souls, the unquantifiable essence that makes us unique. It is this same ineffable quality that leads us to cherish the creative process, to find meaning in the act of creating something from nothing.
When we sit down to play a game of chess or create a piece of art, we aren't necessarily striving to achieve the god-like level of human-machine hybrids. Rather, we engage in these activities because they allow us to explore our own humanity, to push our limits and, ultimately, to surprise ourselves. The enjoyment of the game or the creative process begins only once the gadget is put away.
As AI continues to permeate our lives, it is essential to recognize that it doesn't diminish our humanity but challenges us to redefine what it means to be human. It is a reminder of our own ineffable qualities, those aspects that cannot be replicated by machines. In the face of rapidly evolving AI, we must hold steadfast to the truth that our individual human experiences, our complex emotional landscapes, and the intrinsic value of our creative pursuits are ultimately what set us apart.
The rise of AI should not be seen as a threat to our humanity but as a catalyst to explore the depths of our own ineffable qualities. By engaging with AI, we are given the opportunity to understand ourselves better, to appreciate the complexity and richness of our emotions, and to further refine the human touch in our pursuits.
In conclusion, as AI continues to impact human endeavors like chess and art, we must not lose sight of the ineffable qualities that define us as human beings. Instead, let us embrace the challenge to dig deeper into our souls and forge a stronger connection with our own humanity. By doing so, we can celebrate the intrinsic value of our creative processes and find new ways to express our truths in a world increasingly influenced by AI. Ultimately, it is through this exploration of our ineffable human nature that we can best navigate the ever-changing landscape of technology and maintain a unique, authentic voice amidst the rise of artificial intelligence.
In an AI-dominated world, it is up to us to remember that our humanity, our ineffable qualities, and our ability to express our truths are what truly matter. Whether we're playing chess or creating art, we can continue to embrace our human nature, push our boundaries, and surprise ourselves with our own potential. As we adapt to the evolving relationship between humans and AI, let us cherish the human touch in our endeavors and seek to enhance, rather than diminish, our understanding of what it means to be human.