Planet Glitch vs. Inside: Quality Control, AI Costs, and Traditional 3D Development

At Mad Saturn Interactive, every new project doubles as a production experiment. This post is a behind-the-scenes look at two very different workflows we've run side by side this year: Planet Glitch, our first AI-generated series, and Inside, our upcoming project built the traditional way, in Blender and 3ds Max.

Inside moves away from the science-fiction territory we've explored in Planet Glitch and Nodisea 2000. Instead, it will explain how everyday environments and objects actually work — the kind of "how does this thing function" curiosity that doesn't need a sci-fi premise to be interesting. We wanted to share what pushed us toward building it with full manual 3D control instead of repeating the AI-generation approach.

🛠️ 1. Quality Control: The Challenge of AI 

Planet Glitch wasn't built in Blender. For that series, we deliberately stepped outside our usual pipeline and worked with generative video tools — Veo 3 and Kling AI — to see how far AI-first production could take an independent studio. It required paid subscriptions to both platforms, and the credits ran out faster than we expected, episode after episode.

The core problem wasn't the quality of our prompts. Even a carefully written, highly detailed prompt is no guarantee that the model will interpret a scene the way you intended. We spent a large share of our credit budget just correcting recurring visual errors rather than producing new footage:

• Anatomical anomalies — extra limbs, or hands with the wrong number of fingers.
• Temporal artifacts — objects popping in or out of existence between frames.
• Model hallucinations — the generator quietly ignoring explicit instructions in the prompt.

Every one of those fixes meant another generation attempt, and another chunk of credits gone. With Inside, we're back in Blender and 3ds Max, where every light, camera, and object in the scene is something we place and control directly. There's no generation lottery — if a shot looks right in the viewport, it looks right in the render.

📐 2. Credit Consumption vs. Format and Render Resolution

One advantage of owning the full 3D scene is flexibility of output. Because we control the camera and composition directly in Inside's project files, we can render the exact same scene as a vertical clip for YouTube Shorts or TikTok, or as a horizontal Full HD/4K long-form video, without touching the underlying quality or style.

That flexibility isn't free when you're working with AI generation. Reformatting an AI-generated scene into a high-fidelity horizontal cut often means generating it again at a different aspect ratio, and each attempt burns more credits. For a small studio operating on tight deadlines, that turns "resize the video" into "pay for it again" — and it's one of the quieter costs that doesn't show up until the invoice does.

⚖️ 3. Quality vs. Quantity: The Dilemma of an Emerging Studio

None of this means we're writing off AI tools. As a studio focused on 3D animation and game development, we still see real value in using AI to lighten the load in pre-production — concept art passes, rough voice-over drafts, early ideation. Where it gets harder to justify is high-volume, finished content production, and that's where the economics start working against you:

• AI infrastructure costs — a faster turnaround from AI generation can be offset, or entirely erased, by recurring, unpredictable processing costs.
• Performance risk — no generation tool can guarantee that a finished Short or video will get enough views to earn back what was spent rendering it.

💡 Conclusion: What Should We Optimize?

Running Planet Glitch's production this way left us agreeing with something we've seen echoed elsewhere in the animation industry: AI video generation can end up more expensive than expected the moment you hold it to a real production standard, rather than a quick draft.

That leaves every independent creator with the same underlying trade-off to make: in your own pipeline, is it more valuable to optimize for development time, or for production budget?

💬 We Want to Hear Your Opinion!

If you're an animator, 3D artist, or content creator — which side of that trade-off do you lean toward? Would you rather save time by leaning on automated tools, or keep full technical control over your production budget, even if it takes longer? Tell us about your own experience in the comments below.

---

Support Mad Saturn

If you enjoy following projects like Planet Glitch and Inside as we build them, you can support the studio directly:

• 🎨 Join us on Patreon for behind-the-scenes updates and early access to new episodes.
• ☕ Prefer a one-time contribution? You can send support via PayPal.

Every bit of support goes straight back into production — renders, tools, and the time it takes to build these worlds.

☕ Enjoyed what you saw? We're a small independent team building this universe project by project — if you'd like to help fuel the next one, buy us a coffee!

☕ Buy Us a Coffee