My boss thought an AI video takes 5 minutes — the reality is endless failed generations
Sensitive_Signal_339 · reddit · 2026-09-22
After seeing impressive AI clips online, the author's boss assumed making a video was: idea → prompt → wait a few minutes. Reality: once you need an actual story — characters, scenes, pacing — complexity explodes, and even a simple-looking video takes many attempts.
The author used PixVerse to turn prepared ideas and images into short clips, but the tool was only part of the process: deciding what to create, what to regenerate, and how to assemble the final cut took most of the effort. His takeaway: there's a big gap between "AI can generate a video" and "AI can make the exact video you imagined." The button is getting easier to press, but good output still requires someone behind it.
More from Multimodal
- SVEET: streaming video editing with a diffusion model hits 15 FPS on a single H100 — SJTU · 2026-09-22
- Qwen rumored to push 5-10T params with recursive self-improvement, plus Yunqi lineup — lxfater · 2026-09-22
- tldraw flash: How video game rigging inspired a keyframe-free animation tool — max__drake · 2026-09-22
- MiniMax video model warps faces in 5 seconds, Reddit user seeks fix — ElatedPyroHippo · 2026-09-22
- Qwen Image 2.1 quantized to 2x speed — but INT4 quantization loses to FP4 — mesmerlord · 2026-09-22
- Tencent Hunyuan confirms its image model supports editing, not just generation — TencentHunyuan · 2026-09-22