A Film Is Not A Sentence

I keep coming back to one stubborn fact about filmmaking: images think faster than sentences.

A storyboard is not “drawing.” It is decision making at the speed of instinct. A few ugly lines can lock the camera height, the distance, the power dynamics, the left to right flow, the reveal, the rhythm of the cut. It is not trying to be pretty. It is trying to be clear. It is a tool for sequence, not a tool for taste.

That is why writing a screenplay has always felt a bit unnatural to me. Not because scripts are useless. They are essential. But because they are a translation of a visual medium into an abstract one. The script is a social document. It coordinates humans. It secures money. It gives departments a shared reference. Yet the moment you want to answer the real questions, where is the camera, what do we see first, what do we hide, when do we cut, what is the emotional geometry of the scene, you escape the sentence and you go back to images.

Studios learned that early. Disney is often credited with pushing storyboarding into a formal production method, but I am not going to pretend there is one clean origin story. Film history rarely works like that. What matters is the principle: the moment you pin a sequence on a wall, you stop arguing in language and start thinking in shots. You move scenes around. You feel timing. You see cause and effect. The story becomes physical. Words stop being the bottleneck.

Now we arrive in 2026 and suddenly we are back at the written word again, except now it is called a prompt.

And yes, at first glance it looks like a step back.

Because the prompt is the script stripped down to a steering sentence. And the steering sentence is a low bandwidth interface for something that lives in space and time. If I write “close up, anxious, backlight, rain, handheld,” I have not actually authored the shot. I have described vibes and hoped the system guesses the rest. That is not storyboarding. That is negotiation.

The practical problem is even more brutal. Storyboarding is fast because it is deterministic. I draw the character on the left, they are on the left. I draw the camera high, it is high. I want a brutal silhouette, I scribble a brutal silhouette. If the idea is wrong, I scrap it and redraw it in two seconds. That loop is the whole point. Rough is not a limitation. Rough is the engine.

Generative images are not built for that loop. They are inherently stochastic. Even when you do everything “right,” you still get outputs that are variations of an interpretation. That is great for exploration, but it is poison for the kind of tight iteration that storyboards are made of. You spend time prompting, waiting, sifting, and retrying. You are no longer sketching choices. You are curating guesses. That is why it feels hit or miss. The system is doing what it is designed to do. It is sampling. Storyboarding is deciding.

So yes, if someone tells me prompting is the new storyboarding, my instinctive answer is no. It reverses the original win. Storyboards were invented, adopted, and loved because they replaced fuzzy language with direct visual thought. Prompting reintroduces language as the primary interface and then asks a machine to translate it back into images. That is two translation steps, not one. There is loss in each step.

But I also do not buy the easy conclusion that it is simply regression.

Because the real question is not “text versus image.” The real question is “where does authorship live.”

In my view, AI is not a magic creativity button and not a simple calculator. It is a medium. A medium becomes meaningful through human intention, taste, interpretation, cultural context, craft, and responsibility. In other words, authorship is not the act of pushing the generate button. Authorship is the chain of decisions that carries meaning and takes accountability for the result.

AI does not replace storyboarding, because storyboarding is the fastest way to make visual decisions in sequence. That was the whole point of the Disney era shift to boards in the first place: stop describing the movie and start seeing it. If AI pushes you back into typing descriptions and rerolling images, it is fighting the job the storyboard is meant to do.

Where AI can fit, in a way that actually matches a real film process, is at the handoff points where your rough boards need to become usable for other people. Not prettier. Usable. You still do the panels fast and ugly, because that is how you iterate. Then AI can help turn that sequence into practical artifacts that production already relies on anyway, like a readable shot list and a first animatic timing pass. This is not about “finding the look.” It is about taking what you already decided and getting it into a form that can travel through the pipeline without you spending hours retyping, formatting, and rebuilding the same information in three different documents.

Important caveat, and this is inference based on typical production realities, not a guaranteed promise: this only works when the inputs are structured and the expectations are modest. AI is decent at translating and assembling. It is not reliable as the thing that decides geography, screen direction, and cut logic. Those decisions belong to the boards and the edit, because that is where filmmaking actually happens.

There is another uncomfortable truth here. Prompts can seduce people into skipping the hard part. A generator gives you finished looking frames. That feels like progress. But film is not frames. Film is cause and effect over time. When the tool rewards you with instant polish, it is easy to mistake polish for clarity. You can end up with a pile of impressive images that do not cut together, do not maintain screen direction, and do not carry performance logic. That is not a technical failure. It is a workflow mistake. The tool is optimized for single outputs, while storyboarding is optimized for sequence thinking.

At the same time, I do not want to dismiss what prompts can do for people who do not draw. If someone cannot sketch, they still need a way to explore ideas. For them, prompting can be a bridge into visual thinking. The danger is when the bridge becomes the destination. When you stop learning how to decide shots because the machine keeps handing you options, you can become dependent on randomness. That is not liberation. That is just outsourcing taste to a slot machine.

So is this a step back?

Sometimes, yes. In the specific craft of storyboarding as rapid rough iteration, prompting is usually slower, less controllable, and less aligned with what the tool is for.

Sometimes, no. In the broader creative pipeline, AI can be an amplifier if it is kept in the right role, as a medium that accelerates exploration and reduces overhead while leaving authorship intact.

The simplest way I can say it is this.

Storyboards exist to turn filmmaking back into seeing. Prompts risk turning it back into describing. If I use AI, I do not want it to replace seeing. I want it to serve seeing.

The moment the prompt becomes the center of the process, I am not storyboarding anymore. I am writing requests and curating accidents. And I did not get into cinema to become a professional requester.

I got into cinema to decide what the audience sees, and when they see it, and why it matters