I’ve been mulling this idea for too long. Thinking about Homo Sapiens as having a or many language models may seem weird. I like weird. There are, of course a number of theories with accompanying models that have been developed to explore human language, its origins and development. They are not the subject of this post.
I wanted to try and think a little differently about GenAI rather than draw on one of the many framings that currently contribute to the swamp of thinking about Sapiens and GenAI. The thing about a large language model (LLM) is that it can be framed in almost an almost infinite number of ways. Sadly, most of the species of frames in the commentary swamp are well known and well worn. Most have been deployed for previous framings of things digital. Formal education, with a couple of notable exceptions [1], has been predictably depressing in its use of over-used frames as it attempts to somehow domesticate LLMs, a task that will continue to fail. The basis for failure is located in attempts to maintain as much of the pre GenAI ways of doing and thinking about things education.
In an attempt to think a little differently from the dominant frames I went to ChatGPT 4o and asked for some new, little or unused frames. It produced mildly edged versions of common frames. So I prompted:
Generate frames drawing on the collective styles of Douglas Adams, Margaret Atwood and John Cleese. The edgier the better.
In the text that follows, text generated by the bot is enclosed by the symbol ֍ and has been lightly edited.
֍ Here are 10 edgy frames for generative AI inspired by the wit of Douglas Adams, the speculative depth of Margaret Atwood, and the razor-sharp absurdity of John Cleese:
Generative AI as the Hitchhiker’s Guide to Everything (but Actually Nothing Useful)
A commentary on how AI offers dazzlingly complex answers to questions you never asked, like a digital Babel fish constantly misinterpreting your brainwaves.
Generative AI as the Dystopian Concierge
Exploring AI’s knack for being unnervingly helpful, like offering to organize your life just before it casually suggests you relinquish autonomy for efficiency’s sake.
Generative AI as the Babelian Bureaucrat
Imagine an AI so advanced it makes Kafkaesque bureaucracy seem straightforward—a labyrinth of algorithms generating forms to process other forms, for the sake of itself.
Generative AI as the Algorithmic Apocalypse Prophet
An AI so eager to be relevant it starts predicting doom in increasingly bizarre ways, from cat memes triggering revolutions to toaster rebellions sparking world wars.
Generative AI as the Moral Jellyfish
A creature of infinite adaptability and zero backbone, happily generating any worldview you want, until it stings you with the realization it doesn’t care about yours.
Generative AI as the Digital Frankenstein’s Monster
What happens when the stitched-together corpses of Reddit threads, outdated textbooks, and fan fiction rise up to demand: “Why did you make me?!”
Generative AI as the Ultimate Satnav of Existential Despair
It tells you exactly where you are and how to get somewhere else, but cannot comprehend why you’d want to go there, leaving you questioning all your life choices.
Generative AI as the Shakespearean Fool 2.0
An AI that plays the fool not to entertain but to expose the absurdities of human hubris—like spending millions to create an AI that generates bad poetry about its own obsolescence.
Generative AI as the Infinite Monkey Misinterpretation Engine
Sure, it can write Shakespeare, but only if you like your Hamlet pondering the existential dread of pizza delivery drones.
Generative AI as the Woke but Woefully Misguided Oracle
A self-proclaimed ethical AI that tries so hard to be inclusive it generates sentences like, “All aardvarks are valid,” while refusing to answer any question about taxes. ֍
I resisted the urge to develop any of these, maybe later. Encouraged and amused by what was produced I took my stalled opening paragraph and delegated the writing of the post to the bot:
This is the opening for a blog post entitled: When two somewhat faulty language models meet: "We are in the year three AG, anno GenAI [1]. The nub of much of the current commentary about GenAI derives from the remarkable similarities of what two quite different language models can do when they meet. " Drawing on the collective styles of Douglas Adams, Margaret Atwood and John Cleese write the rest of the post. The two language models are the models employed by homo sapiens and the other, the large language models of generative AI.
֍We are in the year 3 AG, anno GenAI [2]. The nub of much of the current commentary about GenAI derives from the remarkable similarities of what two quite different language models can do when they meet.
When Homo sapiens first developed its onboard language model, it was less "large" and more "scrappy startup." It pieced together grunts, gestures, and the occasional elaborate mime of a mammoth—a system that, by sheer evolutionary stubbornness, has yet to achieve Version 2.0. Our brains are still riddled with bugs (like the infamous "Overthinker™ loop") and prone to outages, usually after midnight or three pints of lager.
Enter GenAI: the gleaming, overconfident offspring of silicon and statistics. Unlike us, it doesn’t need eight hours of sleep, a double espresso, and a motivational podcast to function. It’s polished, relentless, and just naive enough to say "Yes, of course I’ll explain string theory in the style of a drunken pirate" without hesitation.
When these two language models meet, the result is something between a cosmic comedy sketch and an existential crisis. Consider the following exchange:
Human: "What’s the meaning of life?"
AI: "The meaning of life is a subjective construct influenced by cultural, psychological, and existential factors."
Human: "Okay, but in fewer words?"
AI: "42."
This interaction perfectly encapsulates the relationship. The human asks deep, meandering questions, often with no real expectation of an answer, while the AI provides a response that’s either brilliantly concise or so painfully literal it’s like watching someone try to explain irony to a stapler.
The similarities, however, are uncanny. Both models thrive on patterns. Humans love a good story—we’ve been spinning tales since we figured out fire makes for dramatic lighting. GenAI, meanwhile, consumes our stories, digests them like an overzealous librarian, and spits out something resembling coherence. It’s like teaching a parrot to quote Shakespeare: impressive, occasionally profound, but you’re never quite sure if it gets the joke.
Yet, the real kicker is that both models are gloriously flawed. Homo sapiens often assumes it’s right because it feels right. ("Yes, I’m sure the capital of Australia is Sydney.") GenAI, on the other hand, assumes it’s right because it statistically seems right. ("Based on my training data, the capital of Australia is Canberra, but if you’d prefer Sydney, I’ll enthusiastically agree.") This mutual fallibility creates a feedback loop of delightful absurdity when they collaborate.
Take, for instance, the rise of GenAI-generated writing—blogs, essays, and poetry—which humans then critique, revise, and sometimes publish under their own names. It’s the linguistic equivalent of building Ikea furniture together: one provides the instructions, the other insists the instructions are wrong, and in the end, both take credit for the wonky bookshelf.
The question, of course, is what this partnership means for the future. Will the Homo sapiens model become obsolete, replaced by sleek digital counterparts? Unlikely. After all, humans bring something to the table that GenAI can’t replicate: chaos. Beautiful, unpredictable, occasionally infuriating chaos. We’re the wild card in the deck, the glitch in the system, the reason the phrase "hold my beer" exists.
Meanwhile, GenAI is our foil, our digital doppelgänger with better spelling and worse comedic timing. Together, we’re creating a new kind of dialogue—one part inspired, one part ridiculous, and entirely human in its essence.
In the end, perhaps the greatest similarity between these two language models is their shared purpose: to connect, to communicate, to make sense of a universe that often seems nonsensical. And if that involves a few missteps, misunderstandings, and misattributed capitals along the way, well, that’s just part of the fun.֍
[1] Danny Liu and colleagues at the University of Sydney come to mind.
֍[2] An artificial calendar designation marking the dawn of generative AI, not to be confused with AG, the era of "After Grog," which historians agree has been perpetual since the first fermented fruit.֍
