The essay is called “AI-generated posters don’t have to be horrible,” which is the most honest thing about it. A defense of AI imagery whose own title concedes the premise. It was published in June, and it is sitting on Hacker News this weekend at 1,055 points and 597 comments — several hundred people now deep into an argument about aesthetics.
Meanwhile, Adobe reported its fiscal third quarter this month: record revenue of $6.76 billion, up 13 percent year over year; total ending ARR of $27.50 billion, up 11.2 percent; non-GAAP earnings per share of $6.13 against roughly $6.08 consensus; GAAP net income of $1.83 billion. And tucked into the release, the figure that should end the argument outright: AI-first ARR exceeded $650 million, growing more than 150 percent year over year.
The Figure Both Camps Are Skipping
$650 million against $27.50 billion is 2.4 percent. That is what a feature looks like, not a business.
Both sides of the AI-art fight have built their case on the assumption that generative imagery is a tidal wave about to swallow commercial illustration. The boosters think they’re riding it. The holdouts think they’re drowning in it. Adobe’s own numbers say the wave, so far, is a wading pool — and the incumbent everyone assumed was roadkill just grew total ARR 11.2 percent, added subscription revenue of $6.56 billion (up 14 percent), and crossed a billion monthly active users across its creative and productivity tools.
That is not a company being disrupted. That is a company selling the disruption as an upsell. The AI-first ARR figure growing 150 percent off a small base is the tell: Adobe is not losing the poster market to a model. It is charging extra for the model.
If you want to know who’s actually winning the AI-image era, stop reading the 597 comments and read the segment disclosure.
The Poster Was Always a Grudge Purchase
The deeper error in the whole debate is the assumption that the bottom of the design market was ever protected by taste.
It wasn’t. Nobody commissioned a church rummage-sale flyer or a regional bank’s branch-opening poster because they wanted art. They wanted forty of them by Thursday and a line item that survives an audit. The design market’s floor was held up by friction — the cost of finding a freelancer, writing a brief, sitting through three rounds of revisions for a document that would be stapled to a telephone pole.
Generative imagery didn’t win that segment on aesthetics. It won by deleting the brief.
Which is why “AI posters are horrible” is the wrong hill to die on. The people buying bottom-tier design work never thought the posters were good either. One first-year marketing coordinator at a regional credit union, in a Slack message to a print vendor, put the demand side about as plainly as it can be put: “I don’t need a good poster. I need forty of them by Thursday and a budget that survives an audit.” She is not a philistine. She is the customer.
Adobe’s billion monthly active users are mostly her — people producing documents, not masterpieces. The old arrangement sold her a tool that took three weeks to learn. The new one sells her a prompt. Either way, she was never in the market for a masterpiece.
The Policy Fight Is About Labels, Not Prices
The legislative action confirms it. A September 4 roundup from the Transparency Coalition tracks New Jersey’s A 5088, the “AI Likeness Protection Act,” and A 5089, the “AI Image Disclosure Act,” both introduced May 14 and both still parked in the Science, Innovation and Technology Committee. New York’s S 8451, the FAIR Act, goes at news transparency instead.
Every one of those bills regulates disclosure, likeness, and labeling. Not one of them touches the price of a flyer. That is not an oversight — it is an admission. The symbolic layer is where the fight still feels winnable, because the economic layer has already resolved.
And here is the part that should make the right uncomfortable. If you genuinely wanted to protect working illustrators from AI, you would have to restrict the supply of images — licensing requirements, mandatory human-in-the-loop rules, a genuine labor cartel. That is the argument. It is the same argument organized labor has made for a century, and it is the argument a lot of people on my side spent forty years mocking as rent-seeking. You cannot be against occupational licensing for hairdressers and for it for illustrators because the illustrator has a nicer Instagram.
Who Has to Update Their Priors
The boosters first. You are not democratizing creativity. You are serving the one segment of the creative economy that always wanted to pay nothing, and you are doing it inside tools owned by the same two or three companies that owned the old tools. Adobe’s AI-first ARR is $650 million because Adobe is charging for the prompt. The liberation is a line item.
The illustrators second. Your moat was never aesthetics. It was procurement — the search costs, the negotiation, the revision cycle. Those fell, and no amount of beautifully rendered posters will rebuild them. The clients who left were never going to be argued back with craft.
And the arguers, most of all. A fight over whether AI posters are “horrible” is being conducted by people with no purchasing power in the category they are arguing about. Everyone is waiting to see whether the models get good enough. The market already answered: it doesn’t need them to be.