Over the last two years I’ve noticed something very disturbing. As AI continues to advance, the ability (and desire) to be creative, imaginative, and curious has diminished. The amount of people who see AI solely as a way to get out of “thinking” is staggering. They do not see it as a way to enhance themselves, but a means to replace themselves. One look at any Google search result tells the story better than I ever could. And sadly, this is the behavior all models and new tools cater to.

These three attributes define humanity—possibly even our greatest God given gift. Yet they are slowly fading away each day, and I’ve not once seen anyone question this. What might happen should this continue?

Now before some of you are tempted to leave… No, this is not an article written by some “anti-AI”. I’m rather a large proponent of it: I have seen beautiful things created with its aid; people who have had their creativity and imagination unleashed because of it; ideas that would never be now are. You’ve likely seen this yourself. However these people are the extreme minority. The 1% of the 1%. And if we were to only focus on them, which often we do, we would miss what is happening to the 99.9%.

Now let’s look at a perspective you may have glanced over. With a goal that we might make AI more powerful, beautiful, and—like the printing press that sparked an explosion of human creativity and learning—help usher in a new renaissance where AI amplifies our imagination rather than replacing it.

Use It or Lose It

Before we move on there is something we need to agree on: Your ability to be curious, imaginative, and creative, will fade (or disappear) if you stop actively doing so. For those of you who are teachers, especially of young children or those in the arts, you see this first-hand. But for those that are more scientifically inclined, humor me with a small thought experiment.

Between reading a book and watching a movie, which experience demands more imagination? Both make use of it, but surely we can agree a book demands far more of its reader (and recent studies support this). Your mind has to construct the “movie” for itself in real time after all.

Now I want you to consider the people you’ve met in your life you find particularly creative. The ones who approach problems differently, have wild ideas, or create fascinating things. Do they not tend to be more of a reader than those who largely only watch? I have personally found this to be true time and time again.

Should our quick thought experiment of reading versus watching be accurate, it suggests a strong correlation between actively using your imagination and its increased potency over time. And the opposite holding true for those who neglect it.

This observation happens to be backed by hard science—something known as neuroplasticity. Side Note: It was brought to my attention that there has been extensive research on this topic showing this to be true, but we’ll carry on with my original train of thought

I won’t belabor you with an entire paper on this phenomena. Simply put, neuroplasticity is the brain’s ability to quite literally reorganize and optimize itself by forming and changing neural connections throughout your life. Your brain adores being efficient and saving energy after all (and thankfully so—for its comparatively small size it consumes 20% of your body’s energy daily).

When certain neural pathways are not used regularly (like those associated with a skill), they become weakened and are eventually pruned away—leading to a decrease in proficiency of related activities. On the other hand, it also strengthens and optimizes pathways that are used more often… see why i titled this section “use it or lose it”? It is quite literally true.

So what does all of this have to do with “the death of creativity and imagination”? Well, let’s keep this science of “use your creativity or lose it” and neuroplasticity in mind as we examine what exactly we’ve been building with AI (we’ll stick to mainly looking at LLMs) and how people have been using it.

Let’s take all of the think out of it

Let’s start with the easy one. ChatGPT. Super powerful. Fantastic app. But here’s a question for you: When was the last time you began a conversation with any LLM by sharing your own thoughts first? Saying “let’s explore how we might accomplish this task. Here are my initial ideas”? Or do you, like most, simply ask it to do the work and be done?

If you’re being truthful, likely never or at best a few times.

You know this to be true, even if you’re among the rare few who do otherwise. And even if I showed you that thinking collaboratively before executing the task yields demonstrably better results, most won’t bother—it’s easier to say “do it” and be done. Why expend energy on “thinking” when you’re not rewarded for doing so and the outcome is “good enough”?

And then people realized “good enough” was not “good enough” and complained. And you know what, OpenAI listened and surprised many when they updated ChatGPT with o1. Now when you ask a question it does the “thinking” portion for you now as well (amazing). And people were shocked how much better the results for many tasks became. And thus we taught those people once again nothing of the value their own ingenuity could bring. A muscle that will now begin to atrophy.

Consider data analysis. Isn’t it incredible that you can now upload any dataset and simply ask “what seems interesting?” and receive an answer with extreme certainty? So incredible that experts across industries urge you to do the same. Yet every seasoned data scientist will tell you the most profound insights emerge when you challenge those initial results with your own perspective and domain knowledge—when you refuse to accept the first answer and dig deeper, beating it against an anvil like a blacksmith. But why bother when you don’t have to anymore? The tenacity required of curiosity withers.

Consider the art of writing itself. In mere years, I’ve watched countless individuals abandon creative processes and techniques that have been honed for thousands of years. The vital steps of brainstorming, outlining, drafting, reviewing, editing—all cast aside in favor of instant gratification. Why take the time to bring your own imagination into something when you can just ask, receive, and be done? (Though they’ll still complain when their output pales in comparison to those who embrace the ancient, tried and true process.) And another neural pathway withers and dies.

That should be enough, I hope. These tools, powerful as they are, do not encourage the beautiful imagination and knowledge residing within their users. Being creative is NOT EASY—it’s painful, exhausting… It’s hard. But the exhilaration when you break through is one of the greatest feelings in the world. While AI could help us become more creative than ever before, we’re instead teaching an entire generation that its power lies simply in “doing this.” That thinking deeply is unnecessary, a waste of precious time.

And what of the future? Well, the creators of Midjourney (a popular AI art tool) gave us a glimpse of it in 2022:

Kids will not fix this problem for us. They aren’t being taught how to do the difficult task of thought to language. One [of] the saddest realizations for me when we were scaling the @midjourney server at @discord in ‘22 was seeing millions of US gen z kids struggle to prompt. They literally don’t have the words. Broken english. Pidgin lingo. Translating thought to language is insanely hard for them. Full post on X

The end to that particular story, by the way, was a creation of a new feature: /describe. A feature allowing users to simply upload a reference image and have AI provide the words necessary to create similar results. Genius solution. But for most users it will only exacerbate the problem. Even less imagination is now required. Further stunting growth of those neural pathways until they disappear.

Consider, then, this sobering thought: could future generations rebuild society if the need arose? Will they possess the fundamental skills of imagination and creativity that have driven human progress for millennia? Because right now, we certainly aren’t cultivating them.

How ironic that we spend so much attention on “AI safety” (limiting what an AI is allowed to output) while giving no thought to safeguarding and promoting the creativity and imagination of its users.

Yes, your imagination does add COLOR

As a short aside, if you’re wondering if adding your own creativity and curiosity does matter at all, I’d like to give you an extremely simple example to show you it absolutely does.

Imagine you have a task of writing a fairy tale to be used in a D&D session with your friends. And, unfortunately, you don’t have time to write it entirely yourself so you solicit the aid of an LLM like Claude. Below are examples of a simple “do it for me” prompt, and another that adds the tiniest bit of my personal color to it (15 extra seconds of thinking and writing):

Prompt 1: “Please write for me a short fairy tale i could use in my D&D campaign with my friends”

1. Output 1: The Mirror of True Names
2. Output 2: The Mirror of True Sight

Prompt 2: “Please write for me a short fairy tale I could use in my D&D campaign with my friends. I’d like for it to have the flavor of a mixture of Nordic and Russian culture. And be written like many fairy tales were: as stores told by parents to discourage or encourage types of behavior in the children it is shared with”

1. Output 1: The Snow Sister’s Gift
2. Output 2: The Tale of LIttle Masha’s Winter Promise

Let’s look at the first thing you’re likely to notice in both prompts, especially if you’ve never done this type of exercise before. Notice how, especially in Prompt 1, both outputs are nearly identical. That is not a bug or a flaw, but an outcome of how LLMs and many other “AI” work. Your output is entirely dependent on the text and ideas you provide. The same input generally gives you the same output. So if you always go for the minimal approach of “do it for me”, you’ll find yourself with the same output as your neighbor down the street.

Now take a look at how much more color a few seconds of creativity adds. The output of prompt 1 appears great at first glance (great job Claude), but it’s a typical story you’ve heard many times… it’s devoid of color despite how well written it is. Contrast this with how much more personalized, and focused on the core message given in either output of Prompt 2. You’ve made something with real color and hints of your own personality in it with a mere 2 extra sentences and 15 additional seconds of typing.

My hope. Tools that believe in our creativity and imagination

So what would it look like if the AI tools we create actually believed in and relied on the user’s imagination and creativity? I could write an entire book on this, but let me instead answer your question with another question: What if a tool refused to simply answer your question… and instead asked you to think deeper, to explore outside the box, to provide more context before moving forward?

Remember the /describe tool from Midjourney? Extremely elegant. But imagine if they had a “brainstorm” mode where you conversed with the AI on what you’re looking to create? Where it shared examples of different styles, asked what you, the imagineer, liked and disliked, slowly building the vocabulary needed to express your ideas? I think we’d find far more imaginative pieces of artwork than the copy/paste prompts you see everywhere today provide. And if you’ve used these tools, you know exactly what I mean (8k ultra hd, hyper realistic, photorealistic, unreal engine 5, ray tracing, volumetric lighting).

What about a writing tool that guided you through the entire creative process: brainstorming, researching, and outlining. Not simply giving you “the result” but walking alongside you step by step. We’d see shallow articles transform from being a waste of bits on the internet (e.g., “how to paint your house”), to something far more fascinating (e.g., “the optimal time and process to paint your house in Virginia based on weather patterns and humidity”). Actually, don’t make that one, my team is already on it.

All of us have fascinating stories, skill sets, and places we draw inspiration from. And for millennia, most of us have lacked the ability to share our stories or execute our ideas because we lacked the ability to. AI can change that. It can amplify your creativity and imagination beyond anything we’ve ever seen. But first, we need to stop letting machines be “creative” for us and start using them to enhance our own creativity. Otherwise, we might just end up living in the Matrix after all.

So, weary and now inspired adventurer, what will you do next?

Project Shine Post

That marks the formal end of this article, or at this point you might say my plea. I want to shake things up because I believe we’re in need of a paradigm shift. And I’m not willing to merely sit on my hands and wait for someone else to come to the same realization and do it for me. Which is why I’ve launched Project Shine Post: A group of individuals focused on creating tools, and entire AI models, that encourage and value the creativity, imagination, and curiosity of its users.

Right now we’re few in number, and have some exciting projects we will be releasing in the very near future. But we’re always searching for individuals who:

1. Said “damn right” at least once in this article
2. Thrive on figuring things out themselves. You don’t need to be told what to do.
3. Don’t mind wearing multiple hats at once. We’re small after all!
4. Have unique experience they’re ready to bring to the table

DMs and email are open. And we’re not just looking for technical talent—if you have a network of exceptional people or interest in funding projects like these, let’s talk.

We are extremely selective, and adore Steve Jobs’ recruitment principles. So just make sure you’re ready to answer this: where is the intersection of what you’re uniquely capable of doing, and find most interesting in what we’re doing?

Stay curious out there!