The Executive with an AI Subscription

An executive at a company we will not name spent an afternoon in Claude Code. They prompted their way to glory generating a handful of static pages and calling it a website. The pages went live. The pages are, against any comparison to their entire competitive landscape, the absolute ugliest marketing site in their category.

Six weeks of design iteration had been underway when the executive made this decision. Six weeks of real work. Informed by a real brief. Audited against the brand strategy. Refined across rounds. Costed properly. Paid for. And then, displaced by one afternoon of prompting, and the executive’s personal conviction that the prompting got the job done.

No, it didn’t. The design team knows it didn’t. The competitors will know someone fapped at the proverbial wheel the next time they audit the category. The executive is blissful and “victorious” in their imagination, because intellectual masturbation and the ego’s happy endings. The only people—who know full well that it didn’t, but still nod—are the executive’s board. Because they’ve been told a story. Just like everyone else. About AI-driven efficiency, and velocity unlocks, and modernised workflows. And so everyone nods. Because that is the story everyone is currently nodding at. Or at least that’s what keeps the salary coming in. Nothing wrong with it. Just symptomatic.

This is one specific story about one specific person. We are going to come back to it three more times in this piece. We are coming back to it because every other example we cite, and there will be several, is a variation on what this executive did. And the easiest way to understand the variations is to first understand the original.

The original is this: a person with authority and a tool, choosing the tool over the authority of the actual experts in the room who were hired and paid to do the work better. We are calling that pattern, throughout this piece, the Great Industrial Cowardice. It is the practice of shipping work you know is inferior, while telling a story that frames the inferiority as a choice rather than a failure. It is everywhere right now. It is, in our considered view, the actual story of the AI-era’s commercial output. And it has almost nothing to do with AI.

What Cowardice Looks Like in 2026

Three signals tell you, in any organization, that you are looking at the Great Industrial Cowardice rather than at genuine AI-driven transformation.

The first: the output is worse than what existed before the substitution. Not different. Not unfinished. Just worse. By any honest comparison the people inside the organization would make if they were asked privately. And by the comparison the customers are already making, whether the organization acknowledges it or not.

The second: the people who would normally object have been routed around, fired, or silenced. The senior writer who would have killed the article. The design lead who would have killed the layout. The marketer who would have killed the page, or the campaign. The product manager who would have killed the feature. No wait, the product manager wouldn’t have done anything but the opposite. Never mind. The point is: the experts were not consulted. Or they were consulted and overruled. Or they were consulted, overruled, and have since left, which is treated by the remaining leadership as a vindication of the decision rather than as the indictment it actually is.

The third: the PR is more confident than the work warrants. The press release uses the word transformation. The internal memo uses the word velocity. The LinkedIn post from the executive uses the phrase AI-native. (And let’s be honest: “AI-native” only tells you someone is trying to ride a pony they named AI while borrowing from the biggies whose models are really at play under the hood. Some business moat that is.) None of the language addresses the work itself. The language is, in fact, a substitute for the work itself. An attempt to rebadge the inferior output as the leading edge of a category that the rest of the industry has not yet caught up to.

When all three signals are present, you are not watching innovation. You are watching cowardice in cosplay.

This is not, to be clear, an AI problem. The Great Industrial Cowardice is not new. Companies have been shipping work they knew was inferior for as long as companies have existed. What is new, is the costume. AI has become, in 2026, the most popular outfit cowardice has ever worn. Because it is a costume that comes pre-fitted with its own defenders, its own glossary of self-justifying terminology, and its own narrative arc in which the cowardice gets to be cast as bravery.

We will return to why AI specifically has become this costume in a moment. But first, some examples.

Coca-Cola, the Christmas Ad, and the Defence of the Indefensible

In November 2024, Coca-Cola, the company that has spent the better part of a century training the world to feel something specific about Christmas, released an AI-generated reanimation of its iconic 1995 Holidays Are Coming spot. The ad was widely panned. NBC News reported that audiences described it as “soulless” and “devoid of any actual creativity.” (OMG! Really?) Gravity Falls creator Alex Hirsch mocked the campaign with a viral one-liner about Coca-Cola being made “from the blood of out-of-work artists.” The visual quality was uncanny in the precise sense. The trucks looked almost right. The snow looked almost right. The people moved with the strange off-rhythm motion that gives away every AI-generated video produced to date. Every focus group reading on this ad, internal or external, would have said the same thing. Don’t ship this!

And yet… Coca-Cola shipped it.

And then, this is the part that matters, Coca-Cola defended it. As Advertising Week observed in its post-mortem, “faced with this backlash, many brands in Coca-Cola’s situation would retreat. They’d issue a mild-mannered apology, including the usual platitudes like ‘we hear you’ and ‘we got this wrong,’ and then pull the ad. Not here.” The company’s public communications framed the ad as a creative experiment, an embrace of new tools, an evolution of the marketing function. They followed it with a second AI-generated spot in 2025, just to make sure nobody mistook the first one for an accident.

Let’s be precise here about what happened.

A company with effectively unlimited budget for advertising. A company whose Christmas advertising has, for thirty consecutive years, been the gold standard of the medium. A company that knew, with absolute certainty, that it could commission better work from any agency on Earth. That company looked at an AI-generated reanimation of its own legacy, recognized that the reanimation was worse than the original, and shipped the worse version anyway. Then it told the public that the worse version was the future.

Coca-Cola did not make a worse Christmas ad because AI made them worse. Coca-Cola made a worse Christmas ad because they wanted the cost savings of AI more, or the label of “AI-nativity” or “future-readiness”, or any variation of that you’d like to pick, than they wanted the quality of their own legacy. And they wanted the story about being an AI-forward company more than they wanted the work to actually be good. Who gives a fuck if a cola company is AI-forward? Unless the soda itself were infused to make the consumers of the beverage more AI-native instead of diabetic.

That is not innovation. That is cowardice. That is shipping inferior work, knowing it is inferior, and using the AI label as the alibi for the shipping.

We wrote about the human cost of decisions like this in another piece in this series. Here, we are concerned with the decisions themselves. The cost of Coca-Cola’s decision is not just the embarrassing ad. It is the precedent. The message to every other marketing department on Earth that: you can ship worse work, defend it as innovation, and survive. That cost compounds across every category.

The Executive with an AI Subscription, again

Remember the executive with an AI subscription? Now widen the lens.

Coca-Cola, with the resources to commission any creative work it wants from any agency on Earth, did the exact same thing the executive did. Different scale. Different industry. Identical mechanics. A tool was used to skip the work. The work suffered. The institution defended the inferior work rather than admit the work had been chosen for the wrong reasons.

The executive’s afternoon in Claude Code, and Coca-Cola’s AI Christmas ad, are not analogous in budget, or in stakes, or even in intellect. They are, however, analogous in posture. Both are the posture of a person, or an institution, who would rather ship something inferior with a confident explanation attached, than wait for something better and admit they had needed to wait.

That posture is what we are naming.

CNET, the Fake Bylines, and the Rebadge

In late 2022, CNET, the technology news publication founded in 1994, with one of the most established editorial brands on the consumer internet, began publishing AI-generated finance articles under the byline “CNET Money Staff.” The byline was a fiction. There was no staff. The articles were written by an AI tool with light human supervision. CNET did not disclose this to readers. The only acknowledgement was buried in an obscure byline description on the bot’s author page.

The practice was exposed in January 2023 by Futurism, which followed up its initial report by digging into the actual content of one of the articles. A piece about compound interest. The piece contained significant factual errors. Mathematical mistakes. Conceptual misunderstandings about how interest works. At least five distinct inaccuracies in a single explainer. CNET issued corrections to 41 of the 77 AI-published pieces. More than half. A correction rate that, in any other context, would have ended editorial careers.

Here is what CNET did next. Rather than acknowledge that the experiment had failed, rather than discontinue the practice, rather than restore the staff writers whose work had been displaced, CNET escalated into rebadging. They added “reviewed and edited by a human editor” disclaimers. They paused the explicit AI generation, while exploring how to continue the practice under different cover. Futurism observed that CNET even began adding AI disclaimers to human-written articles about AI topics. A labelling-as-confusion strategy that would make any consumer-protection lawyer flinch.

This is the cowardice in its most fully-developed form. Each individual choice CNET made was, in narrow operational terms, rational. Don’t admit the experiment. Don’t discontinue the practice. Don’t restore the writers. Add a disclaimer. Change the labelling. Wait for the news cycle to move on. Each step was defensible inside a meeting. Each step was indefensible in the aggregate.

CNET did not lose its credibility because AI replaced its writers. CNET lost its credibility because its leadership was more afraid of public discomfort than they were of the systematic erosion of a thirty-year-old brand. The AI was the proximate cause. The cowardice was the real one.

That distinction is, in our view, the entire story of the current moment. People keep writing think-pieces about how AI is reshaping work. The think-pieces are mostly wrong, because they are looking at the wrong layer. AI is not reshaping work. AI is providing cover for executives who wanted to reshape work in ways they previously couldn’t defend in public. The cowardice was already there. AI just gave it a wardrobe.

The Executive with an AI Subscription, again

Remember the executive with an AI subscription? Of course you do.

Now move from advertising and journalism to design. The cowardice has the same shape across every discipline. Because cowardice is not domain-specific. It is a posture. It is the posture of the person who would rather ship something they know is wrong than commission the patience to do it right. In Coca-Cola’s case, the patience was a real director and a real production team. In CNET’s case, the patience was the staff writers they had displaced. In the executive’s case, the patience was the six weeks of design iteration they were already paying for.

In every case, the patience was available. In every case, the patience was rejected. And in every case, the rejection was dressed up in the same borrowed language about velocity, efficiency, and the future.

The borrowed language is the cowardice’s favourite trick. It lets the person making the cowardly decision describe themselves as the one with vision. While the people they have overruled get described as legacy thinking, friction, or, our personal favourite: not embracing the change.

Why Cowardice Is Wearing the AI Costume in 2026

The honest question is, why now. Cowardice has always existed in commercial decision-making. Why has AI specifically become its uniform-of-choice in this particular two-year window?

Four conditions are doing the work, and they all have to be true at once for the cosplay to work.

One: AI output is plausibly impressive. The AI-generated Coca-Cola ad is bad, but it is bad at a level that requires you to know what you’re looking at to call it bad. To a casual viewer scrolling past on a phone, it looks expensive. The CNET-bylined articles read fluently. The errors are inside the substance, not on the surface. The executive’s prompted website looks, from twenty feet away, like a website. The cowardly executive needs plausibility, not excellence. AI delivers plausibility at unprecedented scale. That is the first condition.

Two: AI is currently celebrated. The discourse, the press, the LinkedIn algorithm, the average board meeting, all of these reward AI adoption regardless of outcome. Shipping bad AI work in 2026 costs less reputationally than shipping bad human work, because the AI version comes with a built-in story about progress and the human version comes with a built-in story about failure. The cowardly executive is not insensitive to incentives. They are responding rationally to the incentives that exist. The incentives exist because the discourse is broken. But broken or not, the discourse is real. And it is currently rewarding precisely the behaviour we are describing.

Three: AI is structurally deniable. “We’re experimenting with AI” is an alibi no executive can be punished for. The experiment cannot fail. It can only iterate. The Coca-Cola ad was widely panned, but the panning was absorbed by the company as “learnings” for future AI work. CNET was caught lying about its bylines, but the leadership rebranded the practice rather than ending it. The executive with the AI subscription will, when challenged about the ugly website, point to plans for the next iteration. There is no end-state in which the cowardice has to take responsibility. There is only the next experiment.

Four: AI externalises the cost. This is the most important of the four. The cost of bad AI work falls on the people who have to live with it. The customers using the worse product. The design teams maintaining the inferior site. The writers who lost their jobs to the bot. The engineers who have to debug the prompted code six months from now. The cost almost never falls on the executive who made the decision. They have moved on. They are at the next quarterly review, defending the next velocity initiative. The structural asymmetry is what allows the cowardice to compound. The decision-maker captures the upside (the story of being AI-forward), while distributing the downside (the actual work being worse) across people who never made the decision.

When all four conditions hold simultaneously, AI becomes the perfect costume for cowardice. Plausible enough to defend. Celebrated enough to reward. Deniable enough to escape consequence. Externalised enough to keep the executive’s personal cost at zero.

This is not, and we want to be particular about this, because the position will be misread otherwise, an anti-AI argument. Real intelligence is doing real work in the real world right now. AI tools are folding proteins, accelerating drug discovery, helping radiologists catch cancers earlier, helping engineers ship better code, helping writers think more clearly. The medicine is good. The science is good. The carefully-considered use of AI by people who know what good looks like, and use the tools to get there faster, that is good too.

What we are talking about here is the corporate communications layer of the AI moment. Where the technology has been cheapened into an alibi for decisions that would have been embarrassing to defend in any other costume.

The tech is not the cowardice. The cowardice is the cowardice. The tech is just the current uniform.

If AI did not exist, the same executives would be making the same cowardly decisions in different clothes. They would be calling it digital transformation. They would be calling it agile. They would be calling it lean. The clothes change. The body underneath doesn’t.

The Executive with an AI Subscription, one final time

Remember the executive with an AI subscription? Yeah, we’re only going to come back to this one final time.

The pages they shipped are still live. The competitors, the ones they were trying to outflank, the ones they were paying us to help them surpass, are still doing fine. The objective of the original engagement was for this company to be positioned as the strongest brand in its category. Instead, the executive’s afternoon of prompting produced a website that is, against any salt-worthy comparison to the rest of the field, the absolute ugliest in the category. Not the second-best. Not the middle of the pack. The bottom.

The objective was to be the best. The outcome, was being the worst. The executive defended it as efficiency, AI-fluency, and auto-fellatial supremacy.

The board has been told the velocity story. The design team has, mostly, been discarded. And none of this is hypothetical. It is very real. Right now. In the companies where your friends, ex-colleagues work. In the companies where you want to work next. None of this is recoverable. The cowardice produced an outcome that compounds for years. And the executive has already moved on to the next decision they will defend with the next round of borrowed language, borrowed money, and borrowed time.

That is the cost of the Great Industrial Cowardice. Not the bad ad. Not the embarrassing article. Not the ugly website. The compounding deficit produced by people who would rather ship something they know is wrong than commission the people and the patience to do it right.

Coca-Cola will lose more brand value this decade than it will save in production costs across every AI Christmas ad it ever ships. CNET will lose more editorial credibility than it will recover, even if it never publishes another AI-generated article. The executive with the AI subscription will, in eighteen months, be quietly looking for a new role at a different company, or a different, equally-mediocre epiphany to sell. Where they will defend a slightly different set of cowardly decisions in slightly updated language. Their next employer will be told that they are AI-forward. They will be hired. They will do it again.

The cowardice is, in the short term, fucking working. That is the part that needs to be said plainly. It is not failing in the marketplace. It is being rewarded in the marketplace, because the marketplace is currently mispricing the cost of cowardice and overpricing the cost of patience. This will not last. It cannot last. The brand deficits compound. The customers leave. The talent leaves. The boards eventually notice. But while the mispricing holds, and it has held for at least two years now, and may hold for two more, the cowards will look like winners and the patient will look like dinosaurs.

To the People in the Cowardice’s Path

A direct word to the design teams, the writers, the engineers, the strategists, the editors, the brand leads, the producers, the directors, the people whose paid expertise has been overruled this year by an executive with an AI subscription and a velocity story.

You are not crazy. Your work was better. Their decision was cowardly. The fact that the decision was rewarded does not make it less cowardly. The fact that the cowardly decision-maker is currently being congratulated does not change what was decided. You saw what you saw. You knew what you knew. You were right.

The cowardice will not last. It is too expensive to last. Patience is a longer game than cowardice, and the longer game is the one that wins. The AI costume is going to fall out of fashion the way every costume eventually does. When it does, the people who kept doing the work properly, quietly, expensively, patiently, against the grain of the moment, will be the ones who own the next decade.

That’s you. We will see you on the other side.

Stay sharp.

Sources and references

On the Coca-Cola AI Christmas ad. NBC News’s November 2024 coverage and Advertising Week’s January 2025 post-mortem, both linked inline.

On the CNET AI articles, the errors, and the rebadge. CNN Business’s January 2023 reporting, Futurism’s exposé of the practice and follow-up on the labelling escalation, and Gizmodo’s January 2023 piece on the corrections, all linked inline.

From the Methodborne archive. Taste is Judgement: Why AI Can’t Replace Earned Discernment. The 7 Dimensions of Meaningful Differentiation. The Violence of Hype and the Slow Invisibling, part one of this series, linked inline.