Silly Rabbit: AI Tricks are for Kids

Recently, the internet collectively gasped at a video so magical it felt impossibly sweet: a family of adorable bunnies bouncing on a backyard trampoline under the soft glow of night‑vision, captured by what seemed like a Ring camera. It racked up over 200 million views (up to 231M as of this writing) on TikTok before anyone blinked, and then we all learned it was fake (well, most of us were fooled at first anyway), completely AI‑generated, glitching bunnies merging into blobs, disappearing mid‑bounce, a surreal optical illusion masquerading as a fuzzy bedtime story.

That’s when the question hit me: Wait…weren’t we doing this way before generative AI showed up?

Say hello to the viral oddities of yesteryear: curious, uncanny, and unapologetically human. Those moments when we’d peer at our screens, blink, laugh, or scratch our heads, and think, “There’s no way this is real,” but not because of an algorithm, it was raw human weirdness in highest definition.

Scroll with me back in time: a weird dress that split public perception. A mysterious voice, a dancing baby, and a water-skiing squirrel. Each of them and others in their own gloriously odd way felt as uncanny as today’s AI, but all of them were peculiarly, beautifully human.

A simple photo that splintered the internet’s perception: optical confusion, no algorithmic trickery, but wow, did it feel like something an AI prompt built from neural mush.

This auditory illusion had people divided and baffled without a hint of machine learning. Earworms, not prompt worms.

AKA “Baby Cha‑Cha,” the first viral internet video from 1996. Created as a demo in Autodesk’s Character Studio by animators—including Michael Girard, Robert Lurye, and John Chadwick—then shared via email and GIF across the early web.

A guerrilla billboard campaign claimed thousands reported dreaming of the same anonymous face: no AI involved, but prompt-like familiarity delivered.

Gary Brolsma’s 2004 lip-sync to “Dragostea Din Tei” feels like a generative preview; in reality, it was just timeless enthusiasm.

This bizarre, photoshopped visage popped up through Photoshop experiments—not AI-generated, but uncanny all the same.

A dude navigating a store like a gummy blob. If you saw it today, you’d ask for the motion-capture files.

Since 1979, a rescued eastern gray squirrel named Twiggy has been trained (via remote-control boat, plastic-foam skis, and peanut butter) to water ski across pools and lakes for live performances. No CGI or AI, just serious training and a squirrel with high water safety IQ.

Tay Zonday’s 2007 baritone anthem was nothing more than a hallway mic, a creative PhD mind, and timing. A meme born human, co-opted by the internet.

Imagine the “Dancing Baby” boogies into your feed with its retro CGI charm. By minute two, you’d be sliding into your lawyer’s DMs whispering, “That’s not nostalgia, that’s NVIDIA’s baby model.” Or think of the Yanny/Laurel mystery resurfacing: one audio clip, two entirely different universes, you’d swear it’s prompt engineering gone auditory.

And if a squirrel casually water‑skied across your screen? Forget “aww”, you’d question if an entire neural network rendered it. That rubbery supermarket strut? Today, that walk wouldn’t be human-motion‑quirky; it’d be labeled motion-capture simulation. Tay Zonday’s velvet-voiced “Chocolate Rain” would be piped into court as “tone‑warped so real it must be AI‑prompt‑generated.” Even Meme Man, that gloriously uncanny 3D head, would get the “algorithm testing” mark stamped on it. And that “This Man” billboard, you’d bet your last prompt token it’s a face pulled from an AI training dataset.

First, we’ve become hypersensitive to the uncanny valley: things that look human, but just enough off to make your gut say “nope”. Then there’s sensory whiplash: whether visual or audio, when your brain tries to compute that glitch in the matrix, skepticism kicks in. Let’s not forget production‑quality paradoxes: lo-fi setups delivering hyper-viral moments feel like staged illusions (because maybe they are). Finally, it’s the power of narrative gaps: mysterious origins and odd backstories prime us to assume technology had a hand.

Let’s give it up for the pre‑AI viral weirdos: dancing baby loops, lip-synced internet heroes, and squirrels who ski. These moments remind us that before we blamed models, we marveled, and sometimes panicked, at pure human absurdity. In 2025, when someone screams, “It’s AI, obviously,” maybe we just need to ask: “Or was it just brilliant, playful chaos?”

Wait….nope, it’s definitely AI.