I've been increasingly disappointed in humanity lately. Not because of some grand geopolitical catastrophe, although there's no shortage of material there. Sometimes all it takes is a window blind.
I rent out my house on Airbnb. A couple of times recently, guests apparently couldn't figure out how to lower the blinds. These are the cordless kind. You grab the bottom and pull down. That's it. No strings, no buttons, no secret handshake. Both guests arrived independently at the same alternative solution: hang a blanket over the window.
Think about that for a moment. We are a species that has split the atom, mapped the human genome, landed robots on Mars, and built artificial intelligence that writes code, composes music, and holds remarkably sophisticated conversations. And somewhere along the way, we lost the ability to operate a window shade. It was like watching MacGyver solve a problem after a long weekend.
My favorite part: I had offered one of these guests a tour of the house. They politely declined, explaining, "We're pros at the Airbnb." Fair enough. Nobody should need a guided tour to spend the night in a house. But apparently pro status doesn't extend to the blinds.
In fairness to my guests, I'd left those blinds open on purpose. They swing in as you close the French door and can sometimes keep it from closing, and I wanted people to see the inviting back patio. So maybe the real lesson is that I designed a puzzle and then got mad when people solved it with a blanket.
I realize this is a petty complaint. Nobody is born knowing how every kind of window blind works, and we're all ignorant about countless things. Ignorance isn't the problem. The problem is what happens when ignorance meets an apparent lack of curiosity.
In fairness, I know I'm not the smartest person in the world. Not even close. And if blankets over blinds frustrate me this much, I'd hate to know how people with much higher IQs feel watching the rest of us. Me included.
Welcome to the Idiocracy
There's a word I think deserves a much broader definition: idiocracy.
I didn't invent it. Fans of Mike Judge's 2006 film Idiocracy will recognize it immediately, and the dictionary already defines it in political terms as government by idiots. But I think it describes something considerably larger than a dysfunctional government. Here's how I'd expand it:
Idiocracy (noun): A social condition in which ignorance, poor reasoning, and intellectual complacency become normalized, tolerated, or even rewarded, particularly in a civilization whose technological capabilities have outpaced its collective critical thinking and intellectual maturity.
It's not simply about people being stupid. It's about a society getting steadily more sophisticated technologically without getting more sophisticated intellectually, and, most importantly, failing to notice the difference.
The movie imagines a future that has become spectacularly less intelligent. I suspect the real danger is subtler. What if humanity doesn't get dumber at all? What if our technology simply gets so much more capable that our shortcomings become impossible to miss? Idiocracy might not be a future where we get dumber. It might be one where everything around us gets smarter while our willingness to think stays exactly where it was.
If so, we may already be in the early stages.
What the Movie Got Uncomfortably Right
A quick recap for anyone who missed it. In Idiocracy, Luke Wilson plays Joe Bauers, a soldier picked for a military hibernation experiment because he is perfectly, aggressively average. The experiment gets forgotten, he wakes up 500 years later, and he turns out to be the smartest man alive. By a lot. I'll admit I find that premise more threatening than most horror movies, because "aggressively average" is roughly the bar I'm hoping to clear.
Let's get one thing out of the way: the film's explanation for how humanity got dumb is that less intelligent people simply outbred smarter ones. That's a eugenics-flavored idea with an ugly history, and it's not the part worth taking seriously. Most researchers would tell you intelligence is a lot more complicated than that, and I'd rather not take my social science from the same movie that gave us a monster truck execution.
The part worth taking seriously is everything else. Nobody in the future is evil. They're just comfortable. The machines and systems built by long-gone smart people keep things limping along, and everyone has stopped asking how any of it works. Convenience has quietly replaced competence.
The best example is Brawndo, "The Thirst Mutilator." The country's crops are dying because they're being irrigated with a sports drink. When Joe suggests trying water, the Cabinet pushes back with the only argument they have: "Brawndo's got what plants crave." Why? "It's got electrolytes." What are electrolytes? "It's what they use to make Brawndo." That isn't stupidity, exactly. It's the reflex of trusting the thing that's always been trusted, repeating the slogan instead of checking the soil. In the movie, it's a computer at Brawndo that eventually responds to the crisis by laying off everyone automatically.
This is where the film stops feeling like a comedy and starts feeling like a user manual. Researchers have a name for handing mental work to tools: cognitive offloading. Some of it is great. Nobody needs to memorize phone numbers anymore, and nobody misses it. But there's a version where the tool does the thinking and we keep only the confidence. That's roughly what the Fernandes study (more on it below) found in one specific setting: people using AI on reasoning problems did better, and then overestimated how well they'd done. One study on LSAT questions doesn't prove we're all headed for Brawndo. But "the answer improved and the self-awareness didn't" is a very Brawndo-shaped result.
The risk isn't that AI makes us dumb. It's that "the AI said so" becomes our "it's got electrolytes." A sentence that sounds like a reason, repeated by people (me very much included) who never checked what's actually in it.
The Confidence-to-Competence Ratio
The Dunning-Kruger effect is usually summarized as people with limited competence overestimating their abilities. I've written about this before, and the actual research is more complicated than the meme, but the broader problem of poorly calibrated confidence is fascinating.
Some people don't simply lack knowledge. They lack awareness of what they don't know, and that second problem may matter more than the first. Not knowing something is fixable. You can learn, experiment, ask questions, Google it, or ask an AI. But if you don't recognize that your understanding might be incomplete, why would you bother? You're already right. Just ask yourself.
I encounter versions of this constantly: people confidently asserting things they haven't investigated, defending positions they've never seriously questioned, repeating claims without wondering whether they're true. Of course, I do these things too. Everyone does. The difference isn't whether we occasionally make stupid decisions. It's whether we can recognize them, correct them, and learn something.
Human intelligence is remarkable. Human intellectual humility is considerably less reliable. And in an idiocracy, intellectual humility becomes something of an endangered species.
Enter Artificial Intelligence
Recently I posted a thought on X about the growing idiocracy of humanity and how it makes me increasingly suspect that AI will win. Then I added the important part: "And win has lots of possibilities."
What does it actually mean for AI to win? Science fiction offers a limited menu: obedient servant, benevolent caretaker, or terrifying robotic overlord. Reality could be stranger. What if AI simply becomes so capable that human intelligence is no longer the most important intelligence on the planet? Not through conquest, not through an army of Terminators, but because more and more consequential decisions are better handled by systems that process more information, spot patterns we miss, and weigh possibilities beyond our individual limits.
What happens when the smartest entity in the room isn't human? More importantly, what happens when that entity is also better at recognizing when it's wrong? Presumably it would look at all of us, me included, the way I looked at that blanket.
That last part isn't guaranteed. Today's AI can be spectacularly confident and spectacularly incorrect, which makes it surprisingly human. But imagine systems that become not only more knowledgeable but genuinely more reliable at reasoning, estimating uncertainty, and correcting their own mistakes. That would change the relationship between us and the intelligence we've created quite a bit.
Smarter, But None the Wiser
Now for a particularly delicious complication.
Researchers led by Daniela Fernandes at Aalto University published a study in Computers in Human Behavior (online October 2025) titled "AI makes you smarter but none the wiser: The disconnect between performance and metacognition." Participants used AI to solve logical reasoning problems from the LSAT. Their scores improved by about three points compared with a norm group, but they overestimated their own performance by about four. The classic Dunning-Kruger pattern, in which the weakest performers overestimate themselves most, disappeared. Overconfidence no longer tracked skill; people across the board overestimated. And the people with more AI literacy were, if anything, worse at judging how well they'd done: more confident, less accurate. A second, larger study replicated the findings.
So AI helped people produce smarter answers without making them better judges of their own answers, and the folks who knew the most about AI were the most overconfident. We've built technology that can improve our reasoning while inflating our confidence. What could possibly go wrong?
It's the intellectual equivalent of riding in a self-driving car and stepping out convinced you've become a better driver.
Which raises an uncomfortable question: are we using AI to expand our intelligence, or to avoid exercising it? There's an enormous difference between asking AI to challenge your reasoning and asking it to validate your conclusions. One can make you smarter. The other gives your existing biases a research assistant. When millions of people use increasingly powerful AI mainly to confirm what they already believe, the result could be a technologically advanced idiocracy on a scale we've never imagined.
If You Can't Beat AI, Join It
This is where things get wonderfully weird.
Elon Musk has been talking for years about humans merging with AI. On The Joe Rogan Experience in September 2018 (episode #1169), he described Neuralink's long-term purpose as creating "a high bandwidth interface to the brain such that we can be symbiotic with AI." His summary was characteristically blunt: "If you can't beat it, join it."
His premise is fascinating. Humans are already augmented by technology. Think about how much more capable you are with a smartphone than without one: you have access to more information than virtually anyone in history. But getting at it requires typing, speaking, reading, and interpreting everything through our biological senses. Compared with computer processing speeds, we're communicating by smoke signal. Musk's longer-term vision is a direct connection between brain and machine, with AI functioning as another layer of cognition.
That raises a question I find particularly interesting. What if, rather than replacing human intelligence, AI increases it? Could a neural implant eventually augment our reasoning, give us near-instant access to knowledge, help us recognize patterns beyond our biological limits, or let us weigh competing ideas simultaneously? Could we actually raise our effective intelligence, and would biological IQ even remain a meaningful measure?
Perhaps the future isn't humans competing with smarter machines, but humans becoming smarter by integrating with them. The line between my intelligence and my AI's intelligence might eventually become as meaningless as the line between your ability to remember a phone number and your phone's ability to remember it for you.
Of course, there's an enormous distance between that possibility and today's reality. Neuralink's implants are still investigational, and its clinical trials focus on people with paralysis, ALS, and severe speech impairment: controlling a computer cursor, and more recently, restoring the ability to communicate. Nobody is getting an IQ upgrade installed over their lunch break.
But the larger vision is provocative, and the possibilities aren't mutually exclusive. Maybe we merge with AI and become more intelligent. Maybe we create an entirely new form of intelligence that isn't quite human and isn't quite machine. Either way, my original prediction that "AI will win" gets a lot less straightforward. What if AI wins because we become part of it?
Maybe We Need a Hardware Upgrade
Before we get too excited about installing a chatbot directly into everyone's cerebral cortex, there's a rather obvious problem: greater intelligence doesn't necessarily produce better judgment.
Imagine instant access to virtually every piece of human knowledge. You could explore difficult philosophical questions, understand complex science, or investigate the historical context of almost any political argument. Wonderful. But would you? Or would you use your magnificent new computational abilities to win arguments on social media?
Picture someone with a neural interface that can search humanity's accumulated knowledge in milliseconds, using it exclusively to build ever more elaborate explanations for why their political opponents are idiots. The technology would be revolutionary. The behavior would be depressingly familiar. We might simply produce a new generation of supercharged Dunning-Kruger enthusiasts: people who aren't any wiser, but are now equipped to be wrong at superhuman speeds.
An idiocracy powered by superintelligence. There's a delightful contradiction in that phrase, but it might describe one of the stranger possible futures for humanity.
And it leads somewhere more unsettling. What happens when the technology can recognize our cognitive biases and we refuse to listen? Your neural interface detects a logical inconsistency in your argument. You dismiss it. It presents contradictory evidence. You ignore it. It gently suggests reconsidering your position. You disable the notification. Congratulations: you've invented a neural version of the Facebook comments section.
There are serious questions here about autonomy and privacy, too. Who controls the intelligence connected to your brain? Who owns the information flowing through it? How do we preserve independent thought if our thoughts are constantly nudged by an artificial system? These aren't reasons to abandon the technology. They're reasons to think carefully about what we actually want it to do.
The biggest obstacle to making humanity smarter may not be technological at all. It may be persuading people they have something left to learn.
And Now for the Uncomfortable Meta Part
I used artificial intelligence to help write this article. An article in which I complain about human intelligence, propose an expanded definition of idiocracy, question whether humanity can manage AI, and speculate about machines surpassing us.
The irony isn't lost on me. In fact, it's partly the point.
I brought the frustration, the observations, the ridiculous Airbnb anecdotes, and the original argument. AI helped me organize those thoughts, explore the implications, locate research, and challenge some of my assumptions. It even helped me articulate my expanded definition of idiocracy. Thanks, robot.
But here's the distinction I care about: I didn't ask AI to tell me what to think. I asked it to help me think more deeply about something I was already thinking. At least, that's how I'd like to characterize the arrangement. There's always the possibility that I'm simply outsourcing my thinking while congratulating myself for being sophisticated enough to recognize the danger, which would be an exquisite example of the very phenomenon I'm criticizing.
Maybe AI-assisted writing is intellectual augmentation. Maybe it's cognitive laziness with better punctuation. Probably some of both, and admitting that uncertainty seems a healthier starting point than pretending I have all the answers.
In a sense, I'm already experimenting with a very primitive version of the future Musk envisions: an artificial intelligence that extends my ability to research, reason, and communicate. The interface is just considerably less sophisticated. Instead of electrodes on my neurons, I'm tapping on a glass rectangle. And fortunately, nobody has to drill a hole in my skull.
Yet.
So, Does AI Win?
I suspect AI will eventually outperform humans across an extraordinary range of intellectual activities. Whether, how completely, and what it means for us remain open questions. But I don't think the most interesting outcome is one in which AI defeats humanity.
Perhaps AI wins by helping us overcome some of our own limitations. Perhaps we merge biological and artificial intelligence into something neither could achieve alone. Or perhaps we become so dependent on AI that we gradually abandon the intellectual habits that made it possible in the first place.
That's what makes idiocracy so relevant to the AI era. The greatest danger might not be that machines become too intelligent. It might be that humans fail to become wise enough to live alongside them.
The optimistic possibility is that AI helps us become more curious, more analytical, more self-aware, and less susceptible to the seductions of our own certainty. The pessimistic possibility is that it simply makes human stupidity more efficient. I'm hoping for the former.
Because the question that increasingly occupies my mind isn't whether artificial intelligence will become smarter than humanity. It's whether humanity will become wise enough to benefit from it.
And if we do eventually merge our minds with machines, I have one modest request for the engineers: please include an intellectual humility setting. Make it impossible to disable.
And while you're at it, add a tutorial for the window blinds.
Mostly written by a human, with assistance from an artificial intelligence that may or may not have better judgment than its author.
