As a man who has spent twenty-five years convincing editors that my unique blend of deep personal insecurity and profound laziness constitutes a “literary voice,” I am deeply offended by how cheap it has become to replace me. I recently discovered that a prominent cable news talking head with over 500,000 Substack subscribers uses artificial intelligence to generate every single one of his weekly posts. I won’t name him, mostly because I don’t want to look up how to spell his last name, but also because why should I do investigative journalism when he isn’t even doing writing? A year and a half ago, an investigation by Wired revealed that 10 percent of the top 100 Substacks were entirely AI-written. Given how rapidly humanity is sprinting away from intellectual accountability, I assume that number is now closer to ninety percent, with the remaining ten percent consisting entirely of recipe bloggers desperately trying to explain how their grandmother’s experience in the French Resistance relates to a gluten-free lemon bar. We have officially reached the era where content is no longer generated by human neuroses, but by an algorithm designed to sound exactly like a human who has never actually lived, loved, or suffered from lower back pain. And the truly terrifying part isn’t that the robots are getting better at mimicking my profession. It’s that you, the consumer, have lowered your standards so aggressively that you cannot tell the difference, or worse, you simply do not care as long as the paragraph breaks look neat on your phone. To understand exactly how low the bar has fallen, you only need to look at a recent weekend baseball tournament featuring my 14-year-old nephew, Matthew. His team won, which was exciting for everyone involved, particularly the parents who could finally stop pretending to enjoy sitting on uncomfortably cold aluminum bleachers. After the final game, the boys posed for a celebratory photo with their coach. Unfortunately, the mom tasked with taking the picture possessed the photography skills of a startled raccoon, resulting in an image so incredibly blurry that the children looked like cryptids emerging from a swamp. Rather than accepting this as a metaphor for the fleeting nature of youth, she ran the photo through an AI upscaler. She then distributed the glossy, high-definition result to the other mothers, and announced her intention to frame it as a formal gift for the coach. It was a beautiful gesture, right up until my sister looked at the image and noticed that something was deeply wrong. None of the children’s faces looked quite right, and, more notably, the AI had hallucinated an entirely extra child and placed him directly into the team lineup. When my sister pointed out that a ghost boy had joined the roster, the mom shrugged and said it was “close enough,” reasoning that the coach probably wouldn’t notice a surplus child. My sister, possessing a rare and stubborn commitment to physical reality, insisted they fix it. She ran the photo through a different AI model, managed to banish the phantom toddler, and delivered a slightly cleaner version. The kids still looked like uncanny valley androids cloned from a Sears catalog, but the other moms were absolutely thrilled and posted it immediately. “Close enough” was officially a victory. This “close enough” ethos is exactly how we treat everything now. Back when I used to regularly appear on cable news networks—whether it was CNN, Fox News, or MSNBC—the routine was always the same. I was required to arrive at the studio a full hour before my segment just to sit in a chair and have a professional slather layers of industrial-grade pancake makeup onto my face and blow-dry my thinning hair. This wasn’t because anyone believed my thoughts on American foreign policy were deeply profound; it was because the network executives understood a fundamental truth about television audiences. If you have a single stray hair sticking out of your head, or if your nose is slightly shiny under the studio lights, viewers will spend the entire four minutes focusing exclusively on that blemish. They won’t hear a single word you say about the federal deficit because they are completely consumed by your oily T-zone. Appearance dictates credibility. The exact same principle is now governing written content. AI makes everything read, look, and sound perfectly uniform, provided you don’t look too closely or think too hard. It formats paragraphs beautifully, uses impeccably correct punctuation, and ensures there are zero typos. The presentation is flawless. The actual substance, however, is a distant second. We are so mesmerized by the sleek, shiny packaging of information that we completely ignore the fact that the box itself is entirely empty. The core of the problem is that the more information we consume, and the more rapidly we swipe through it on our phones while pretending to go to the bathroom at work, the less time we spend contemplating any of it. When you are consuming three hundred pieces of content a day, depth and quality become active liabilities. You don’t want a deeply reported, nuanced essay that challenges your worldview; you want a fast, easily digestible text-nugget that smoothly slides into your brain without causing any friction. AI is succeeding wildly right now, not because the technology is achieving true sentience, but because we are enthusiastically getting dumber. We are actively demanding less from our writers, and the machines are more than happy to oblige our intellectual laziness. If you take a moment to slow down and actually analyze an AI-written column, you’ll notice a very specific pattern. It is incredibly confident, fiercely authoritative, and entirely constructed from generic platitudes. It offers a soothing, uninterrupted stream of confirmation with absolutely zero original insight. It tells you exactly what you want to hear, using the precise vocabulary you expect, wrapped in a warm blanket of safe, algorithmic consensus. It is the literary equivalent of a beige waiting room at a dental office. When I first started writing a column decades ago, I made a solemn promise to myself. I vowed that I would never write pieces that people liked merely because it confirmed their pre-existing political biases. It is incredibly easy to write a column where the entire thesis is “Yes! Trump sucks!” or “Yes! Elites are bad!” and watch the likes roll in from people who just want their own opinions read back to them by a professional. My goal was always to write things that made people think, or feel uncomfortable, or see an everyday situation from an entirely bizarre and unexpected angle. AI is fundamentally incapable of doing that. It cannot surprise you because it only knows what has already been written. It cannot make you feel anything because it doesn’t feel anything itself. It is just a highly advanced mirror reflecting our collective mediocrity. And you, dear reader, are perfectly fine with it. In fact, I am so confident in your apathy that I am willing to make a deal. I swear to you right now, if this particular column gets me a significant surge of new paid subscribers on this Substack, I am completely done working. Every single column published under the name The End of My Career from this week forward will be written entirely by an AI prompt while I spend my mornings lifting weights and my afternoons drinking an overly expensive bottle of Barolo in my kitchen. You won’t even notice the difference. And honestly, neither will I. Thank you for paying to read my column. Wait: This is for the people who didn’t pay? Then I owe you nothing. You are the ones contributing to the end of my career. If you want to pay an exorbitant amount of money to get one extra post a month – which often won’t even be that good – upgrade to a paid subscription here:
|
No comments:
Post a Comment