Show, Don't Tell (Said the AI to the Designer)
Or How I Learned to Stop Worrying and Love the Imperfection
I opened the brand designer’s marketing email, not because the subject line was sticky and original and exciting. I clicked it because it wasn’t.
Not these exact words, but something akin to “show, don’t tell.” (It’s not wrong advice at all. It’s quite, quite right, actually.)
“Is she okay?” I wondered about the seasoned designer who sent that consciousness-exploding, cortex-shattering subject line.
It blinked at me like a faulty neon sign, begging to be walked away from.
But I clicked anyway, because I had a theory about who—or what—wrote it.
I suspected that subject line was the brainy-work of AI—not this experienced, talented, human designer’s.
I read through the email. It was, as I assumed it would be, very good! Very well written. Nice flow. Nice word choice. Clean, friendly, catchy, neutral. All very, very correct information on the basic importance of branding and even some inspiring pre-Covidian language about “stepping into your power and potential.”
(Didn’t we agree as a species we would stop “stepping into” figurative things? Right along with “unleashing our inner whatevers” and “manifesting destiny?” No?)
Anyhoo, this “good enough” email was so beige I felt like I was choking down a bowl of taste-free oatmeal that definitely boiled over and left a (somehow simultaneously sticky and papery?) film on my carefully polished stainless steel cooktop and didn’t even have any fun bonus textures in it like chia seeds to get stuck in my teeth.
It sounded pretty dang AI to me.
And even if I’m wrong, the point is that I could be right. Because we’re now in this space where AI can write so convincingly and so accurately that it’s so, so easy to have it write content for us. And you know what? You should take advantage of that, if it helps you accomplish more.
However, comma.
Your AI’s output is only going to be as good as your input.
If you prompt ChatGPT or Claude to “write me a post about the importance of brand design”—then it will vomit up the gray-matter-liquidating advice to show, not tell.
And?
Every brand designer on earth could send that exact email to their followers, word for word, and no one would think it was weird. It would seem like it “fit.” Because the message in it is true. It’s universal.
I really don’t have to tell you that “fitting” in with everyone else’s message is not what you’re aiming to do with your creative business.
But that’s exactly where we’re headed as more and more people discover the incredible capabilities of AI and start using it to do things for them like writing content.
It’s all starting to sound the same. It’s becoming possible to pick out AI-generated content. And not because it’s glitchy and unnatural sounding.
It’s the opposite. It’s because it’s too good. Too perfect! Not glitchy enough!
Isn’t that freaking terrifying?!?!
I miss the old days of 2023 when you could pick out AI-generated content because of it’s rampant hallucinations and for it’s complete lack of understanding of what a human hand should look like.
This makes me wonder...
Could we be headed for a future where making imperfect content is something we strive for? A sort of “humans made this” watermark that AI will struggle to mimic? Will imperfect be our new definition of “what good looks like”?
I’m going to admit...something about that feels exciting to me. Not just the permission to be flawed, but the desire to be.
In the meantime, the way to mitigate AI’s over-achieving bullshit isn’t to avoid using it. It’s to improve your inputs. And don’t you roll those eyes at me, because I’m not going to say “bEcOmE a GoOd PrOmPt EnGiNeEr!”
Okay, well, I am but just not in those words because those words are gross and discouraging (I have to become some new-age breed of ‘engineer’ to use this stuff?? Yeah no thank you and good day.)
I’m going to frame it a bit differently. Instead of you obsessing over learning AI, flip the script.
Make your AI obsessed with learning you.
The best way to do this is not necessarily to figure out the exact words to use in every single prompt, but rather to take advantage of features like Claude Projects, ChatGPT Projects, or even Custom ChatGPTs.
I’ll put together some tips for y’all on exactly what to do with these tools in a future post. But the gist of it is, dump a bunch of info about you and your work into these features. Make it become an expert on you, and instruct it to center you in all its responses.
Because these LLMs are already an expert on your field—be it branding, copywriting, graphic design, whatever. More than you are, in fact. Because AI’s expertise is literally a collection of all the expertise on the internet.
Yes, use that to your advantage. Learn something from it. But don’t forget that everyone else has access to that same information.
AI is ending the Information Age. And it’s kicking off an age where your Imagination, and Creativity, and Weird Human Uniqueness matters way, way more than it ever did before.
Instead of using precious email space to tell me to “show, don’t tell”—goddamit, take your own advice! Don’t tell me in the words of an AI why branding is important.
Show me in the way only a weird unique human like you can, what it looks like to write an email about brand design with your caffeine-jittery hands at 2am, rambling about that client who tried to convince you Comic Sans was "approachable." Show me how your brain spins metaphors between brand identity and dating profiles ("your logo can't be the equivalent of 'I like long walks on the beach' Karen, we need SPECIFICS"). Tell me about the time you almost rage-quit a project because the client wanted her new luxury spa brand to look "fun and whimsical" like a children's toothpaste commercial she loved in the 90’s.
Because THAT's the content that makes me stop mid-scroll, finger hovering over the delete button, and think "well shit, I need to read the rest of this."
🤔 Weekly Wonderings
Will 'imperfect' become the new marker of human-made content? (And how long until AI figures out how to fake that too?)
If AI doesn’t actually “understand” a concept, but can explain it perfectly and have discourse about it, then what’s the difference?
Could we gaslight AI into thinking Comic Sans is genius? Could it gaslight us back?
Has anyone ever successfully stepped into their power, and if so, is it like...a solid surface, or?
And Now, Your AI Afterparty 🤖🎉
📈 Major Platform Updates
Gemini Advanced launches with code execution and image generation
Integrates with Google Workspace, for when you want AI to judge your spelling mistakes across ALL platforms
🤪 Delightfully Chaotic Applications
"The Brutalist" film used AI to perfect Hungarian accents
Coming soon: AI fixes British actors attempting Southern accents (looking at you, House of Cards)
🎨 Tools Creatives Should Actually Care About
New AI design tools dropping faster than you can say "but what about my job security?"
Canva, Adobe Sensei, and others promising to make you obsolete (kidding!)
🎭 Industry Drama & Plot Twists
Chinese AI start-up DeepSeek releasing models to rival OpenAI
Tech leadership arms race getting spicier by the minute
🤮 "Thanks, I Hate It" Corner
Meta creating AI profiles that'll post alongside humans
Because what Facebook really needed was MORE fake accounts