For @AshKoosha , the test of an AI-generated shot is whether it can hold its place in a feature film. He put Kling 4.0 to work with that in mind. Here’s what he found. #Kling4 #KlingAI https://x.com/i/article/2117527269984124928

Ash Koosha: Kling 4.0’s Long Takes Give Emotion Time to Unfold

For Ash Koosha, the test of an AI-generated shot is whether it can hold its place in a feature film. He put Kling 4.0 to work with that in mind. Here’s what he found.

Notes from the Creator

My name is Ash Koosha, co-founder and CEO of Fountain 0, an AI-powered filmmaking platform. We use Kling’s frontier video models as the video-generation foundation within Fountain Engine, our proprietary filmmaking system. In 2026, I wrote and directed two feature films: Dreams of Violets, accepted by and officially screened at Tribeca Film Festival, and ODYSSEUS: The Fall, a 150-minute epic I made as a solo filmmaker. Both were created with the ambition of delivering complete cinematic experiences, not simply demonstrations of a technology.

Kling 3.0 powered video generation for both films. Its versatility gave our engine a strong foundation for achieving realism, dynamic movement and carefully controlled action, even in complex scenes. My interest was never just in generating an impressive image. It was in whether that image could become a convincing shot, and whether that shot could hold its place within a feature-length narrative.

With Kling 4.0, the most important development for us is the prospect of substantially less unusable footage. Based on what we are seeing so far, we anticipate at least a 50% reduction in discarded footage within the Fountain Film Engine. This remains an early estimate, driven by more convincing physical interactions, better handling of spatial relationships and collisions, and less drift away from the action established at the beginning of a shot. At feature-film scale, that could make a substantial difference—not just to generation costs, but to the time and attention available for creative decisions.

The ability to generate up to 30 seconds in a single pass, alongside more precise control over key visual moments, is particularly relevant to how I approach filmmaking. A longer shot is not valuable simply because it lasts longer. It gives a character time to hesitate, react and complete an action. It allows tension or atmosphere to develop without forcing an edit because the generation has reached its limit.

Kling 4.0’s emphasis on more stable camera movement and more nuanced character expression also addresses something important to me. Realism is not just resolution or texture. It is the relationship between the camera, the person and the space. A small glance, a believable shift in weight or a camera movement that feels motivated can matter more than an elaborate visual effect. These are the qualities I want a filmmaking model to prioritize.

Multi-shot generation remains a highly useful creative feature. It can suggest framing and shot progression beyond the initial brief, opening up alternative ways of approaching a scene. Kling 4.0 handles multi-shot sequences more reliably, and its shot orchestration tools are worth exploring, especially for controlling shot direction and order. For me, the balance between interpretation and prompt adherence is essential: when I am exploring, I welcome unexpected ideas; when I am executing a precisely directed scene, I need the model to respect those decisions. The ideal tool offers creative possibilities without imposing them.

The targeted editing capabilities are another promising development. Being able to change a particular expression, movement or visual element while preserving the rest of a scene addresses a practical production problem. Once a take has the right performance and composition, I do not want to lose everything that works simply to correct one detail. In the original clip below, the young man throws the tape away. In the edited version, he places it on the table, while the background and the other people in the scene remain largely unchanged.

What excites me about Kling 4.0 is not simply the possibility of generating more footage. It is the possibility of making a greater proportion of that footage usable, intentional and consistent with the film I am trying to make. Through Fountain, our aim is to give filmmakers more control over the complete work, with Kling providing a powerful generation foundation within that process.

The director still has to make the decisions. A better model makes it easier to carry those decisions all the way to the screen.

Thanks to Ash Koosha for showing us how Kling 4.0’s longer takes give performances time to unfold and how targeted editing lets him refine a single action while preserving the rest of the shot. We’ll be back with more hands-on reviews from creators putting Kling 4.0 to work.