AI is Finally Letting the Weirdos Take Over (And It’s Brilliant)

I’ve been mucking around with AI tools for the past year or so, mostly making absolutely pointless stuff like getting StyleGAN to generate endless variations of my daughter’s stuffed penguin (sorry, Mr. Waddles). And you know what? It hit me the other day while I was using Midjourney to create “medieval manuscripts about robot ducks” at 3 AM - this might be the first time in tech history where the weirdos are actually leading the charge.

A grid of AI-generated variations of a stuffed penguin, each increasingly abstract and strange

Here’s the thing about AI right now: nobody’s written the rulebook yet. The suits haven’t figured out how to properly monetize it, the academics are still arguing about methodologies, and meanwhile, people like us - the ones who’ve always had too many tabs open in our brains - are just… making stuff. Really really good stuff, actually.

Your Brain on AI (It’s Different Here)

I know, right? For those of us who’ve always thought a bit sideways, who’ve had to explain why we’re suddenly deep-diving into Victorian-era typefaces at midnight or creating elaborate systems to categorize memes, AI feels like finally finding the right adapter for your weird brain-cable.

Take my mate Steve (not his real name, protecting the gloriously weird here). He’s got ADHD and has always struggled with traditional creative tools because they’re too linear, too structured. But now? He’s using GPT-4 as a bouncing board for his ideas, letting it catch his thought-fragments and spin them into new directions. Then he feeds those into Stable Diffusion, which somehow understands exactly what “make it more cronenberg-meets-cottagecore” means.

Split-screen comparison of traditional creative workflow vs. AI-assisted creative chaos flow chart

The Beautiful Chaos of No Rules

But here’s where it gets properly brilliant - because there’s no established “right way” to use these tools, every neurodivergent approach is equally valid. Want to use DALL-E to design furniture based on dreams you’ve had? Brilliant. Using AudioGen to create music from the sound of your cat purring? Wicked.

I’ve been futzing with this absolutely crappy hack where I:

  1. Generate abstract shapes in Midjourney
  2. Feed them into ControlNet as control images
  3. Tell it to “see” Pokemon in them
  4. Use the results as album covers for AI-generated music based on the “mood” of each Pokemon

Is it pointless? Absolutely. Has it led to some genuinely amazing bits of internet? You bet your bytes it has.

Breaking Tools in the Best Way

The really incredible thing is how these tools seem to reward divergent thinking. The more you break them, the more interesting the results get. I was mucking around with language models the other day, deliberately feeding them contradictory prompts and roleplaying as a time-traveling Victorian botanist discovering modern memes. The output was… well, it was something else.

Screenshots of increasingly absurd AI conversations about Victorian meme botany

And god forbid you follow the “intended use” of any of these tools. The most interesting stuff happens when you:

  • Stack different AI models together in ways they weren’t meant to work
  • Feed the output of one system into another until something breaks
  • Create feedback loops that shouldn’t work but somehow do
  • Mix analog inputs with digital outputs and back again

The Human-AI Jam Session

It’s a bit odd, because for years we’ve been told that AI would make creativity more efficient, more streamlined. But instead, it’s making it more chaotic, more experimental, and infinitely more interesting. It’s not replacing human creativity - it’s becoming the world’s weirdest jam session partner.

I’ve seen people:

  • Using AI to generate impossible recipes, then actually trying to cook them
  • Creating entire fictional civilizations by iteratively building on AI outputs
  • Making music by feeding train schedules into audio models
  • Building virtual galleries where every artwork is based on misheard song lyrics

And the best part? None of these people are “AI experts” or “professional creators” - they’re just weirdos with ideas and access to tools that finally speak their language.

So What Now?

Here’s what I reckon: we’ve got this brief, brilliant window where the weirdos can help shape what AI creativity becomes. Before the templates settle, before the best practices calcify, before someone decides there’s a “proper way” to use these tools.

The call to action is simple (and I know, right? Who am I to make calls to action?): Make the weird stuff. Make the things only your brain could think of. Use the tools wrong. Break them in new ways. Share your strange processes. Build communities around the unusual approaches.

Because right now, we’ve got this chance to make AI creativity stay weird, stay open, stay human in all our gloriously irregular ways. And that’s properly exciting.

A collage of various weird AI-generated projects from the community, showcasing the diversity of approaches

So go on then - what’s the weirdest thing your brain’s been wanting to make? The tools are there, they’re getting better every day, and there’s absolutely no one who can tell you you’re using them wrong.

(Just maybe don’t generate endless variations of your kid’s favorite toy. Mr. Waddles still gives me oddly accusatory looks from the toy shelf after his brief stint as an AI muse…)