If you've spent more than five minutes inside the AI-creative corner of the internet, you've probably bumped into Zack London — the guy behind Gossip Goblin, the channel that's quietly turning out neo-feudal cyborg goblin civilizations and other oddly specific worlds that shouldn't be possible to make alone. He's one of the more obvious answers to the question "is anyone actually doing serious work with these tools."
So it's a little jarring when he says, in plain English, that AI has zero skill.
He said it. More than once. "There is zero skill involved in generating AI images."
If you're someone who's been defending your AI workflow to your friends and family for the last two years, that line probably landed weird. It sounds like the gatekeepers were right the whole time. Like the guy who's clearly the best at this just admitted the haters had a point.
He didn't. But you have to actually listen to what he's saying.
The button press is not the work
Here's what he's not saying: "I don't work hard." Anyone who's looked at how Gossip Goblin actually makes a video knows he runs around 400 prompts to land a single set of characters. He calls his lip-sync workflow "a colossal headache." He talks about generating "a FUCKload of background shots" just to find one he can force his characters into. End-to-end, a single piece of his work takes 12 to 14 hours.
That's craftsman-level patience.
What he's actually saying — and this is the part worth chewing on — is that the act of generating an image isn't where the skill lives. The button press is not the work. Typing a prompt and hitting enter is the easiest part. A 12-year-old can do it. Your mom could do it on her first try. The thing that comes out of the model on a one-line prompt is, statistically, going to be slop.
That's not a slight against people doing it. I started there too. We all did.
So where does the skill actually live
Watching how Zack works, the skill is in everything around the generation:
- The script. He says it himself: "Every video starts with a script." Not a vibe. Not a mood board. A written, structural piece of writing that tells him what worlds, characters, and scenes he needs to build.
- The taste. Knowing which of those 1,600 images is the right one. Knowing when a face looks "off" in a way that will read wrong on screen. Knowing when the lighting in a Seedream blend is fighting the character instead of helping.
- The judgment to throw work away. 150 generations for a 90-second scene means 149 things he made and rejected. That's discernment, not button-pressing.
- The world-building. Goblin civilizations and cyborg aristocrats don't come from prompts. They come from a writer's head. The model is a paintbrush; he had to know what to paint.
Compare it to a creator like Voidstomper, the AI horror artist with 3 million Instagram followers. Totally different aesthetic, totally different approach — Voidstomper embraces the model's errors as the art. But the skill lives in the same place: knowing what to keep, what to trash, what aesthetic territory you're staking out, and how to develop a body of work that holds together. Both of them are doing real artistic labor. Neither of them is doing it with the prompt itself.
Why it matters that he said it like that
If Zack had said "AI image generation requires patience and taste," nobody would have argued. It would have been a forgettable LinkedIn carousel post. Instead he picked the most provocative version — "zero skill" — and let everyone fill in the rest.
That phrasing is doing two jobs at once.
One: it's a flare for the people who are just prompting and calling it art. He's not insulting beginners. He's calling out the version of "AI artist" that thinks one good output is a portfolio. And I get it. That used to be me. I'd hit a banger in ChatGPT and feel like a wizard. Then I'd try to make a second one that matched it and realize I had no idea what I was doing.
Two: it's an honest description of his own workflow. The image generation step really is the easy part. The script, the iteration, the editing, the world-building — that's where his time goes. He's just being accurate.
What the haters get wrong
The "AI takes no skill" line is also the one critics reach for to dismiss the whole medium. I'll steelman it. They're not entirely wrong about the first generation being trivial. It is. The model does most of the heavy lifting on that first output.
Where they're wrong is treating that first output as the finished work.
A photographer doesn't shoot one frame and call it a portfolio. A writer doesn't keep the first draft. A painter mixes a color twenty times before the one that goes on the canvas. Nobody looks at those mediums and says they take "zero skill" — even though the individual button-press, brush-stroke, or shutter-click is also pretty mechanical.
When Zack London says "zero skill," he isn't conceding the argument to the critics — he's just describing how he actually works. The mechanical part is mechanical. The art is somewhere else, and most of the people complaining about AI haven't bothered to look at the somewhere else yet.
If you're learning right now and that quote made you feel small, don't let it. He's not talking about you. He's pointing at the same thing you're already figuring out, which is that the easy part — typing the prompt — was never going to be the part that mattered. What matters is what you do for the next 14 hours.
Generated Images
Seven variants below — three standard compositions, one documentary (foreground bokeh), and three dynamic-angle "spatial" compositions for parallax video. To request a fix on any one, add a checkbox under## Image Touch-upslike:- [ ] spatial-square: remove the random hand on the right
landscape — 1920×1080
