Is good writing cheap?

An essay on AI slop, F1 driving, and the price of actually having a thought

Reader note: @themgmtconsult Maurizio & I had a chat asking the question - is good writing cheap? This essay follows my thoughts after it.

F1 driving can’t be that difficult, right? Why are those guys in those little highlighter color Hot Wheels-esque cars paid millions of dollars to go zoom, zoom, zoom in circles? Isn’t it just driving? I have a driver’s license, so why couldn’t I theoretically become an F1 driver? Similarly, writing well can’t be that difficult, especially now with AI, right?

Isn’t writing just putting words together in a row? I have a language facility, as do you as a reader, so why can’t anyone theoretically write well? When everyone thinks they can write now that we’re in the AI copy-paste thought leadership this, slop that era, what is good writing? What is the value of it? Is good writing cheap nowadays?

Writing well and F1 driving are the same thing to me - to ask if good writing is cheap means to ask what it means to be a good driver. Asking if good writing is cheap to me is sort of like saying, well, driving can’t be difficult ultimately, you get the hang of it, so F1 can’t be that difficult, can it?

A regular driver and an F1 driver are both driving as actions, but they’re performing completely different activities. They are driving, yes, as they are participating in entirely different worlds with different skillsets and different reasons for different means, but it seems easier to group them under the same umbrella because from the outside looking at it, well, aren’t they just doing the same thing? Zoom, zoom, zoom, right, with driving?

Of course not, they are absolutely not doing the same thing, in fact they are radically different modes of operating, but most people have never seen the gap between who they are driving a car and who they could be as an F1 driver because most people have never been on the other side of the track inside a highlighter colored tiny toy like car, sitting on the driver’s seat helmet weighing heavy with hands gripping the wheel before they face the possibility of their existence ending in a televised crash. Zoom, zoom, zoom, you’re dead.

What that gap actually looks like when you put a number on it is something like this: typically, an F1 driver starts learning in early childhood (i.e. between the ages of 4-8), and like any elite sport, it must be financed to the tune of tens of thousands a year on equipment and travel and coaching before the kid can do long division.

By the time they reach Formula 2, which is the last waiting room before F1, the family or sponsor group has by some estimates usually spent somewhere between hundreds of thousands if not millions getting them to the door. Now, within the sport? There are twenty-two seats in Formula 1. That means there’s twenty-two in the entire world. Like in investment banking hiring, the F1 pipeline eliminates virtually everyone who enters it, and the ones who wash out at Formula 3 are still more talented at driving than virtually every person who has ever touched a steering wheel, including you, me, and everyone we know. Yet none of this is visible when you watch the race. You just see the corner. You see a car brake. You see a car turn and accelerate. You think you understand what you’re looking at because you also brake and turn and accelerate on your way to the grocery store. You do not understand what you are looking at. The gap between what an F1 driver is doing and what you are doing when you drive is so large it is functionally invisible to anyone who hasn’t been told it exists.

Now, I think it’s fair to say that there’s some people who legitimately think their daily commute and Max Verstappen’s equivalent of a commute on a Sunday afternoon going around in circles are points on the same continuum. If you think this is a ridiculous proposition, consider that a YouGov poll of 1,732 British adults found that one in eight men think they could score a point off Serena Williams. Another 14% weren’t sure while only 3% of women thought the same. People genuinely believe they are operating on the same continuum as the best in the world, and they are not even close. One of them is transportation and the other is the most expensive, physically brutal, cognitively saturated life or death in a flashbang moment performance in professional sports, and the only thing they share is that both involve vehicles. Of course, you could also die while driving, but you get my point, right?

Writing has that same gap between what people think they can do and what they actually do. Similarly, the function of actual good writing, writing that deserves the word, is thinking in public. It takes raw material, which is reading and research and experience and domain knowledge and the specific texture of a life that was actually lived, and it processes that material through a specific mind into something that did not exist before. When I say writing well through thinking like this and processing it, I mean no, you shouldn’t be approaching writing with something like ‘can’t I just plug something into an LLM and copy paste that out?’ Yet a lot of people think this, they produce this, they assault our eyes with it, so they think this is good writing just like the delusion of thinking I can be an F1 driver because I can drive according to my driver’s license. So consider the ways driving and writing seem similar as abilities most people have some grasp of to some extent to some means, but what about doing it par excellence? Hmmm?

If you want to dig deeper at the question, consider the premise. What do you value? What do you find cheap? And how do you tell the difference? I think the key here is to consider the role of process. Initially I would have said curiosity was the way forward into thinking about value, but this is not sufficient. That’s because when Maurizio and I first spoke, he left me with a powerful metaphor to consider as he mentioned the Italian word discernere, and then connected this concept of curiosity in the process of thinking by using the analogy of sifting flour. Curiosity is part of what feeds and composes the flour, and then you put everything through a fine mesh and what comes through is usable.

You accumulate widely, you follow rabbit holes, you keep sixty tabs open like a person who has lost control of their browser and possibly their relationship to time (me, my goodness). But curiosity as an action alone is not enough. Discernment is the activity that builds on the action to do the filtering. It’s the moment where you sit with everything to figure out what actually holds, matters, is relevant, and necessary versus what was just noise that felt interesting at 2am. This is where the thinking happens that differentiates, in my view, the start of good writing. Discernment is also what F1 drivers do at five to six Gs through the corner, which is to say it’s the cognitive version of taking a turn at a speed that would kill most people, except they’ve been doing it since they were a child so their body knows something yours doesn’t because yours never had to learn. That’s the same, drumroll please, with writing well. Discernment, discernment, discernment, it’s the new location, location, location in the cognitive real estate section of our brains, I guess.

Don’t even get me started on the topic of taste in the role of discernment as that is a whole other beast. But yes ciao grazie duh, the part that costs cognitive resources, effort, all the words we can use to simply say brain power, well, it isn’t cheap. And AI skips it entirely. It pattern-matches without curiosity, it produces without discernment, and it outputs text that looks like writing the way a commuter and an F1 driver both look like they’re just doing the same action of driving when the reality is their activity is completely different.

Photo by Jonathan Ikemura on Unsplash

So if the thinking is what costs, and AI doesn’t do the thinking, then what AI actually replaced was never writing. It was the assembly of words for an aim in content production, and confounding these two things is how you end up believing that writing is cheap when what is actually cheap is formatting other people’s ideas into an expected structure and calling it work. A peer-reviewed study published in Management Science found that automation-prone freelance jobs dropped 21% within eight months of ChatGPT’s launch while a separate study using actual firm-level spending data from Ramp’s expense management platform found that more than half the businesses paying for freelancers in 2022 stopped entirely by 2025. For every dollar they cut from freelancers, they spent three cents on AI. The speed and totality of that replacement is itself the proof. If this work had required original thought, it could not have been automated this quickly. You cannot automate what you cannot formalize, and you can only formalize what was already formulaic.

But the collapse of assembly was only half of what happened, because the market for thinking didn’t collapse alongside it, and instead it grew. The evidence for the distinction between good writing and assembly, I believe, is in the market itself, since it’s signaling something that neither doomers nor the naive are grasping. More than fifty authors earn over a million dollars a year through subscriptions on Substack alone, the New York Times added 1.4 million digital subscribers in 2025 and posted record revenue of $2.825 billion while the Financial Times hit a record 1.4 million subscribers with revenue topping half a billion pounds. If AI had made good writing cheap, all of these numbers would be falling, right?

Yet they are rising, and it is their thinking, their voice, their production of writing as thought that people pay for which serves as proof that words have the power to move people. Good writing isn’t cheap because it does things that AI cannot (at least) quite yet capture, and this is where it gets genuinely weird, because the irony, in my view, is that as hysteria over the value of writing proliferates as AI slop content is equally profligate, it turns out that we don’t trust AI-generated content. How funny it is then to see that a Raptive study of 3,000 U.S. adults found that when people suspect content is AI-generated, their trust drops by nearly 50%. This held true even when the content was actually human-written. The perception of it being AI-generated alone was enough to drive people away. Zoom, zoom, zoom quickly becomes disdain when @ pangramlabs asking is this slop.

By the way, roughly half of new web articles/content posing as articles is now AI-generated. That feeling you may have of everything getting lower quality? Yes, it is real, it is nonsensical, and it is everywhere. Speaking of nonsensical, just to recap so far: we began with the question of if good writing is cheap, we explored the possibility that we don’t think writing is valuable now that AI can generate it, but we’re surrounded by slop and the moment we think something is AI-generated slop, we don’t trust it, and also, by the way, we’re not accurate at detecting slop. Make it make sense!

The paradox is that people cannot actually tell the difference but they don’t trust it when they think it is, as a Stanford study found that humans distinguish AI from human text at roughly coin-flip accuracy. So what we actually have is a market that is paying a premium for something it cannot verify, which is the belief that someone on the other side of these words actually lived a life and had the thought and put their name on it knowing full well that people with agendas would come for them. That is what the market is paying for. Not quality in any technical sense but belief in thought, in production, in voice, in prowess, in the belief that the words you’re reading right now haven’t been hammered out by Claude or GPT or Gemini or (god forbid) Grok, or really any of the multitude of other options. Pattern recognition by AI is not thinking, which is just as true as the fact that the market does not reward merely thinking. The market rewards visible thinking. Nowadays, visibility is a function at the intersection of distribution, timing, nerve, sheer luck, and whatever parts of an algorithm seem to be triggered that day depending on the stability of Nikita Bier’s employment prospects.

But let’s be fair, because even if you sift the flour, the system that distributes your writing does not care about any of it. Remember the metaphor of sifting flour? The algorithm does not sift flour, it sifts engagement, which is why it pushes slop via all the generic generalized garbage that you’re likely feeling like you’re often reading. The thing we all depend on to find good writing is running on the exact opposite incentive structure from what makes writing good in the first place, and nobody is going to fix this for you because the people who could fix it are the same people who benefit from you not understanding the gap. I don’t have solutions for that right now and I don’t think anyone else does (to my knowledge).

And this brings me to one more thing that Maurizio and I spoke about that has been on my mind. Powerful people have an incentive to make you believe writing is cheap. Consider how some CEOs who over-hired during 2020 and 2021 need a story for the layoffs and they have two options. Option one is admitting they miscalculated, that they hired based on a growth curve that was obviously unsustainable, that they did not see the correction coming because they were too busy being congratulated for scaling and that this is how you fix it under an umbrella of a response. On the other hand, option two is saying they are embracing cutting-edge AI and unfortunately this magnificent technology means they need fewer people because their role can be automated. Consider which story protects their board seat, which story gets a positive headline in Bloomberg, and which story leads to a stock price bump. Then go watch some F1 racing and think about what goes into it.

Disclosure: Definitely not military, financial, or legal advice or allegations, just vibes. Contact me at ani@anibruna.com

Here is the obligatory message to like and subscribe for updates or share with everyone if you hated it.

Like and subscribe for updates - www.notyourtypicalfinancebro.com