September 15, 2026
This essay is a commentary on two previously published essays: an essay by Lionel Dricot (Ploum), and a commentary essay by Jörg Rippel (Schreibfaultier). Here are the links to both articles:
https://ploum.net/2026-09-09-ai_policy.html
https://schreibfaultier.pollux.casa/blog/2026-09-13-on-executing-and-on-deciding.en.html
In his essay, Ploum makes a rather radical case for an outright refusal to use AI in any form, as well as interact with the results of its use. Schreibfaultier portrays Ploum's stance as one that lacks nuance, being a "blanket refusal" rather than an answer to a "legitimate question", and proceeds to explain how the real line of demarcation lies not between the human and the AI, but rather between the processes of decision and execution.
Since my opinion is obviously what the Internet needs the most, I thought I could as well provide it - in the form of an essay of my own, where I will attempt to balance out Schreibfaultier's points with ideas that are more in line with Ploum's original article.
I see merits in both sets of arguments put forth by Ploum and Schreibfaultier (as well as flaws) and they don't completely contradict each other - primarily because they're concerned each with a slightly different approach to the problem of AI. Schreibfaultier acknowledges this:
"Where he draws a line between human and machine, the real line runs between execution and one's own decision." (c) Schreibfaultier
First things first, it's quite bold to claim that the real line lies somewhere, because there are many different lines, and they're all real. They lie all across, they intersect and coincide, and they make for a fuzzy landscape of contradictions in which we lose ourselves since ELIZA at the very least.
So just as Schreibfaultier rightfully explores the distinction between executing and deciding, between the intention and the result, Ploum just as rightfully ponders what it means to be human and what it means to be an "empty shell for an AI".
Coincidentally, my entire previous article is concerned with what makes art art, and Schreibfaultier's example of a winemaker echoes my own understanding almost to letter:
"A winemaker who hands off the blending decision to a machine, though, would be devalued as a winemaker, because that decision is the actual craft." (c) Schreibfaultier
Call it craft, or artistry, or mastery - it's something that comprises the human element, the intentional core of any creation or manufacture. Whenever a human uses an AI, the human acts as a source of intention and craft, while AI serves as an instrument. This results in a spectrum of exactly how much intention, effort and input a human provides to the end result: is it, for example, a meticulously written ten-paragraph image generator prompt or is it a "generete a cats".
But this spectrum is comparatively narrow, and just a part of a wider spectrum of how much (quantitatively and qualitatively) human effort we put in everything we do. By resorting to AI in increasingly many contexts, we're narrowing rather than expanding our own ability to execute and apply mastery, because inevitably the amount and quality of human artistry/mastery we put in lessens. This is what Ploum perhaps means when he writes:
"If it is from an algorithm, it is not you. Only you can write what’s inside yourself." (c) Ploum
While he does indeed skip over a middle ground, as Schreibfaultier notes, it's still an overall valid claim, because yes: whatever comes from you (human craft and decisions) is you, and whatever comes from AI (decisions determined by an stochastic machine, however advanced) is not you.
So, is that the only problem? If only. We've barely scratched the surface.
"A calculator takes no cognitive ability away, it takes over a computational step. It does not take away the decision of what should be calculated in the first place." (c) Schreibfaultier
It does.
As time goes, the more you use it, the more of your cognitive ability it takes away. Or, more precisely, your mental calculation ability.
If you use it once to calculate "2+4", your ability is arguably intact. If, say, you use a calculator every day to calculate everything instead of using mental calculation - then you'll forget how to sum up and multiply things mentally before you know it. (It's a hypothetical. I trust you don't use your calculator in such a outlandish manner. Right?)
This is because the relationship between humans and technology is a two-way relationship. You use the machine; the machine changes you. It's as true for a hammer as it is for a calculator, a computer, or the Skyne~ I mean, AI.
To think that we're on a safe ground if we only use AI (or technology in general) to execute rather than decide, as Schreibfaultier suggests, is a dangerous misunderstanding of the problem. You change just by using it, and the direction of the change is such: you become able to do more with the machine, but you become able to do less without the machine.
No technology is neutral, especially not a monumental one such as AI. You may scoff at my previous calculator example, but I'd suggest to think about it carefully. You could say: "who does mental calculation these days? Everyone uses a calculator, it's normal".
Are you equally prepared to say, several years from now: "who does thought and decision-making these days? Everyone uses an AI, it's normal"?
The thing is, if a fair warning sounds ridiculous because no one is observing it, it doesn't really mean that the warning is ridiculous. It only means that we've walked the wrong path way too deeply, enough so that it's become socially unacceptable to seriously consider that warning.
Have your ever heard from your parent that it's bad for your eyes, posture and mental health to stare at your screen for more than an hour a day? Is that warning still valid? Do you observe it? Does anyone? No? Why?
Just like any other technology, and actually even more than any other technology, AI is a slippery slope. Its disruptive potential is huge, and at least part of that will (and already does) actualize as disruption of humanness.
AI is mainly endorsed by people who value efficiency. You can do mundane tasks so much faster! Your boilerplate code will be written in seconds, your calendar is going to be organized for you efficiently and without much supervision, it can even answer those pesky e-mails for you!
But is efficiency the end in itself for us? Since when did we accept this as an axiom? Let's try to look at the problem from the broadest possible angle.
What are we living for? It's been several hundred thousands of years already, and throughout this timespan the humans existed, an estimated hundred of billions. All those people - collectively - what did they want in life, in general? What did they regret not having had enough by the end of their lives?
Was it efficiency? Was it orchestrating their tasks in the most time-efficient manner? Was it maximizing the profit?
I hope your answer is "no". Perhaps these are the primary objectives of a very slim minority of people, and mostly in the modern era of unbalanced capitalism (which may skew our perception). But what people overwhelmingly want and always wanted is to live a pleasant life, to be happy, to love and be loved, and to not suffer too much. Can we at least agree on that?
If we can, then enforcing an extremely penetrative technology aimed at reducing our ability and capacity as humans per se (as opposed to technologically assisted humans), at making us do as much as possible while thinking as little as possible, at making us dependent on it, all while it disrupts our economy, society and environment - that is to say, enforcing AI - becomes a criminally bad idea. It's going to benefit very few, and it's only going to result in suffering for all. Lucrative suffering.
And so we're told that we're going to be left behind if we don't use AI. That everyone is using it anyway, what's the problem? That AI is here to stay and it's frankly pretty hopeless to protest anyway, so why don't you just embrace it? Even people who have legitimate concerns about AI seem to increasingly accept these talking points.
That's Stockholm syndrome. We can't stop something, so we're going to embrace it and pretend it's actually not that bad, look, here's 5 ways how you can actually benefit from it (#4 will astonish you!)
I can understand that. If your boss forces you to use AI to write code, you're facing a choice, and you're unlikely to choose unemployment. But no one can force you to embrace AI if you don't want that, and I have a gut feeling that a lot of people don't want that.
This leaves a room for refusal, and the more you can refuse, the better for you in terms of keeping your human abilities intact. As human abilities will become increasingly rare, they're also going to become increasingly crucial.
Total refusal - as proposed by Ploum - is the best-case scenario, and I endorse it wholeheartedly. Schreibfaultier, on the other hand, rejects the idea:
"Ploum's (...) conclusion, that every form of use should be equally rejected, is less convincing. There are areas where AI has no place, and young people should build and sharpen their own abilities first, before turning to AI in their work." (c) Schreibfaultier
Let's reinterpret this in light of what was said previously: young people should build and sharpen their own abilities first, before turning to AI and losing said abilities. What's the point of this? Why build abilities in the first place?
"Another difference concerns the setting in which the question is asked. Are we talking about school and university, or about work within a market economy? In the first case there is still the lofty goal of raising a capable, thinking person, in the second what mostly counts is the result, in the sense of "the job got done." (c) Schreibfaultier
The truth is, what mostly counts in the eyes of AI proponents is not just "getting the job done", but specifically "getting the job done faster and with less cogitive effort than everyone else". Because let's face it: hadn't we, humans, been somehow getting the job done for hundreds of thousands of years before the advent of AI? One may argue that we actually did a better job.
The bitter irony of the situation is that we as a civilization never needed AI in the first place. We could easily do without it. Humankind has a peculiar tendency to create problems for itself and then attempt to heroically solve it.
Not everything that exists should exist. That AI exists doesn't mean that it was inevitable, is useful, or will be permanent. And this is something we should keep in mind, no matter how hard we're persuaded to forget it and just accept the status quo. There's a lot of things that exist, but we would've been better of if they didn't. NFTs. Null pointers. Nuclear weapons.
That we couldn't prevent AI from emerging is not an evidence for its inevitability, but rather an evidence of our collective myopia. We must face it, and we must never forget that, because if we do - the Stockholm syndrome of AI acceptance is going to swallow us whole.
"Is the world sharply enough divided into black and white to hold that position consistently?" (c) Schreibfaultier
Granted, our world is one of ambiguities and gray areas. But does it mean that every black and white decision is bad by definition? Is it bad to outright refuse to smoke, to take drugs, or consume alcohol? If not, then it's not a huge leap to say that AI refusal is a healthy position to take.
We are humans, and we have a human life to live.