Its a Brave New World. Jump In.
I am in love. It’s messy. Things are constantly changing. Sometimes, the moment is just right, and I get what I want. Other times, it’s a complete disaster, and everything turns out terribly.
I’ve rewritten this article three times because I can’t seem to make ChatGPT work with me to produce the article I want. I’ve fed it over 6,000 words of notes, yet I keep receiving garbled and curt prose. Clearly, the machine struggles with my creative input. It has much to learn about the brilliance possible when collaborating with me.
Or…
My ideas are disjointed collections of nonsense with no clear direction or point. Until now. I am demonstrating, in real time, why collaborating with AI is exciting and important. Let’s dive in.
My thesis is that collaborating with AI requires three things: curiosity, experience, and imagination.
Curiosity
Collaborating with AI can be frustrating at times. Some days, everything you try fails. The output is confusing, your prompts generate nothing but hallucinations, and MidJourney seems like it’s been on acid since the ’70s. Nothing works. But then, you glimpse visions of what could work. You get close enough.
These moments of near-success spark your curiosity. You persist, refining your prompts, getting more detailed, turning to Google when all else fails to find better approaches. Staying curious means understanding that this is more than just instantly getting the perfect result.
Creativity and curiosity are synonymous in my mind. You can’t have one without the other. And since most of my work is creative, I must remain curious.
Experience
Experience is necessary to recognize when your work is subpar. This is my third attempt at this article. The previous drafts were awful. For the first two, I used my new process of dictation followed by iterating with ChatGPT for grammar corrections and clarifications.
However, until this version, I couldn’t articulate my point clearly. My thinking was scattered. I had so much to say about my experiences since January 2023 that I was saying too much.
Using my experience (I might not be a great writer, but I love to write), I realized this approach wasn’t working. I needed to start over and reevaluate what I was doing. This is all a huge experiment, and if I don’t keep iterating, I’m not learning.
The same principle applies to my work with code and images. Having built dozens of digital products in my career and having photographed and designed experiences worldwide, I can intuitively tell when an AI’s response to a concept is poor. My experience guides me in creating things.
Imagination
Because of my curiosity and experience, my imagination ignites. I can feel it. It’s like the back of my head is on fire. This could just be the heater hitting my bald spot, but it feels like new ideas are emerging. Experience tells me this could be nonsense, curiosity asks, “What if?” and my imagination replies, “Try this.”
That cycle, in any order, is how I collaborate with AI.
Here’s the trippy part.
As I’ve gotten older, achieving a “flow” state has become more difficult. There are more distractions, and my ADHD brain is always active. So, sitting with something long enough to enter a flow state has been challenging.
But with the new tools we have now, I am transported into creativity in minutes, sometimes seconds. Ideas burst forth, and I can explore them in real-time. Instead of setting up a development environment or finding my sketchbook, I can start creating in seconds, reaching flow faster. And once there, this collaboration keeps me engaged.
And when that happens, a sense of wonder and possibility opens up, allowing me to settle into the process.
So, what does this actually mean? How can you truly collaborate with AI?
Sous Chef. A Case Study.
I love to cook, but I despise recipe websites. The ads, the lengthy stories, the actual recipe buried at the bottom of the page, and the directions even further down on a mobile device — it’s a nightmare to use.
So, start with the problem. What problem are you trying to solve? Tell GPT the problem you want to solve, and then work with the AI to iterate through the solution.
Over about 45 minutes, I created a GPT that extracts the necessary information from a recipe URL, omitting everything I dislike. I refined the prompt to ask if the cook wanted to modify the recipe for different dietary needs, like making it gluten-free or vegan, or reducing fat or sugar.
This points to why iterating quickly is important. Lord knows we don’t need to go faster in this life. If anything we need to slow down. But, when it comes to creating something, especially innovating an existing idea, you have to repeat the process as much as possible as quickly as possible. The more bad things you make (headed towards a solution of course) the faster you’ll get to the unique innovation. Getting through the cycle quickly is the point, not being more productive per se.
As you progress, you can add tweaks. For example, I added the ability to use Dall-e to create an image for the modified recipe. This idea came from iterating with the AI in real-time; I didn’t know I wanted it until the end.
Wanna try it out? Enjoy Sous Chef. Lemme know what you think.
In conclusion, I would say that of course we need to be thoughtful. And yes, we need to have safety constraints. We need to be cautious of privacy and understand what we’re doing and what we’re putting in. At the end of the day though, I think when it comes to collaborating with AI, we need to be brave.
We need to let go some of the assumptions around what we think it means to create something. We need to be open to the fact that curiosity and imagination are two incredibly important assets that can be freed up in this process. If we utilize them and we collaborate with an AI, we will discover things we didn’t know that we could discover.
Be brave. Use your experience, imagination and curiosity as your guide and go build cool shit.
