The thought hit me on a morning walk, and it hit hard. I know I am not the first person to have it, but I could not shake it: artificial intelligence is not just a new tool. It is a disruptor, and it keeps disrupting one field after another in almost exactly the same way.
A Pattern With a Name
Business schools have a term for what I was noticing. In 1997, Harvard professor Clayton Christensen described what he called disruptive innovation, and the shape of it is worth knowing. A disruptive technology almost always arrives looking worse than what it challenges. It is cruder, less polished, and easy for established professionals to dismiss. But it is cheaper and easier to reach, so it finds people the incumbents were not serving well. Then it quietly improves, year after year, until one day it is good enough for most people. By the time the established players take it seriously, it is usually too late to respond.1
Seen through that lens, the word "slop" is not really a criticism of AI. It is what practitioners have always called a disruptor in its early stages. Painters said much the same about photography, and live musicians said it about recorded sound.
One Field After Another
Written text went first. When OpenAI released ChatGPT in November 2022, writers and teachers were quick to point out how bland, repetitive, and confidently wrong it could be.2 They were right, at the time. Four years later, AI drafts a large share of the routine writing done in offices, classrooms, and nonprofits, including grant narratives and donor letters, and the conversation has moved from whether it can write to how to use it responsibly.
Visual art followed close behind. Image generators like DALL-E 2 and Midjourney arrived in 2022, and by December of that year artists were flooding ArtStation, the leading online portfolio site for professional illustrators, with a protest image reading "No to AI Generated Images."3 The early results really did have six-fingered hands and melted faces. Today the hands are fine, and AI images are everywhere.
Music came next. In April 2024, more than 200 musicians, including Billie Eilish, Katy Perry, and Pearl Jam, signed an open letter warning against AI that would "destroy the music ecosystem."4 Performers were not far behind. When a production company unveiled an AI-generated "actress" named Tilly Norwood in 2025, the actors' union SAG-AFTRA responded that she was "not an actor" but a computer program "trained on the work of countless professional performers without permission or compensation."5
Each time, the objections were sincere, and many of them were fair. Each time, the technology kept improving anyway.
The Latest Example: Video Games
This week brought a striking example from the game industry. Bloomberg reported that a director at Epic Games, the company behind Fortnite, told the team building its Disney-backed Marvel and Star Wars shooter that the project was going "AI first." According to the report, the instruction was "no more hand-written code," with all art "generated with full AI workflows." Some staff reportedly hoped the directive was not serious. Epic confirmed the direction, saying it is "embracing AI-accelerated development," led by its own developers. The game, originally aimed at November, had already slipped to March before the mandate.6,7
Here is the detail that stopped me. ArtStation, the site where illustrators staged that 2022 protest against AI images, is owned by Epic Games. In less than four years, the same company went from hosting the artists' revolt to ordering its own teams to generate their art with AI.
Rank-and-file sentiment has moved in the opposite direction. In the Game Developers Conference's 2026 industry survey of more than 3,000 workers, 52 percent said generative AI is having a negative impact on the industry, up from 30 percent a year earlier and 18 percent the year before that. The most negative views came from the people whose work is most directly affected: visual artists, designers and writers, and programmers.8 The pushback is real, and it is growing. So is the adoption. Christensen would not be surprised.
Is Software Next?
That brings me to the part of my walk I keep coming back to. I recently came across ArtCraft, a small open-source project, written in the Rust programming language, that is taking a swing at Adobe's creative suite. It is very early and describes itself as "not a pixel-perfect clone," and its components carry names like PhotoCraft and VectorCraft, with a WordCraft piece hinting at word processing as well.9 It is not going to worry Adobe this year.
But a free alternative to Photoshop is nothing new. GIMP has existed for more than 25 years, and I use it myself. Adobe has survived free competitors for a long time, because a free program that does everything Photoshop does, almost as well, is still a hard sell to people who already know Photoshop.
What has changed is something different. With a tool like Claude Code, it is now realistic to build a piece of software for an audience of one. Not a clone of a commercial product with a thousand features, but a small program that does exactly the five things you need and nothing else. Suppose I wanted a PDF editor like SodaPDF. I would not need all of SodaPDF. I need to merge files, fill in forms, add a signature, and occasionally redact a page. An AI coding tool, building on mature open-source libraries that already handle the hard parts, can produce that program in an afternoon, and it runs on my own computer at no ongoing cost.
I know this works because I have done it. Over the past year I have built several small apps for my own Android phone: a timer that silences my phone for a set period, a podcast player that works the way I want a podcast player to work, a tool that turns a voice note into an email, and a word game. Each of them replaced something I would otherwise have bought, subscribed to, or put up with. None of them would interest a software company. Each of them fits me exactly.
One custom app is trivial. A few million people each quietly declining to renew a subscription because they built something that fits them better is a disruption. I do not know whether that is where software is headed, but the pattern from text, art, music, performance, and games suggests it would be unwise to assume it is not.
What This Means for Your Organization
For a nonprofit or a school, this is not an abstract question about the future of Adobe. It is a budget question. Small organizations often carry a long list of software subscriptions, each one modest on its own, that together add up to real money. Before the next renewal comes around, it is worth asking a simple question about each one: what do we actually use this for? Sometimes the honest answer is "three features," and three features may be something you can replace.
That said, building your own tool comes with responsibilities a subscription handles for you. Someone has to fix it when it breaks, update it when your computers change, and make sure it handles sensitive information, like donor or student records, safely. And if the volunteer who built it moves away, the organization needs to know how it works and where it lives. A custom tool that only one person understands is a risk, not a savings. The right candidates are tools that are simple, low-stakes, and well documented, not the systems your organization depends on every day.
The disruptor pattern does not mean every subscription is doomed, or that every organization should start writing its own software. It means the cost of building something tailored to your needs has dropped dramatically, and that is worth knowing as you plan. If you would like help sorting out which of your tools might be candidates and which are worth keeping, the contact form is the best place to start.
