Reddit hot take: World Labs' deterministic camera control beats rolling the dice with AI video
Spiritual_Skirt_9312 · reddit · 2026-11-10
A Reddit post (with a screenshot of Fei-Fei Li praising World Labs' Atlas camera control) argues World Labs is doing what OpenAI should have done with Sora from day one.
- Current AI video is a slot machine: prompts like 'slow pan left' on Runway/Kling yield 360 flips and mushy faces — fine for 5-second flexes, useless for specific camera angles.
- Atlas is different: instead of a 2D diffusion model guessing how a room looks when you move, you move a virtual camera inside an actual 3D representation, with pixel-perfect control and export to 3D Gaussian Splats rather than flat MP4.
- Caveats: dynamic physics and character interactions are likely still janky, but deterministic camera motion beats 'roll the dice 30 times and pray.' The author asks VFX/Blender folks for takes.
More from Multimodal
- Is ControlNet Obsolete? New DiT Image Models Ship Without It — Manicarus · 2026-11-10
- Full prompt revealed: how to recreate the Van Gogh convenience store in MiniMax H3 Image — mhdfaran · 2026-11-10
- 10 open-source GitHub repos for voice AI: from local ASR to zero-shot voice cloning — Shruti_0810 · 2026-11-10
- Fable 5.2 generates animated nature scene with ~70k trees from one prompt — Merzmensch · 2026-11-10
- MOSS-Image roundup: Control LoRAs, 461 new characters, and 20-minute LoRA training on a 5090 — optimisticalish · 2026-11-10
- StreamEval from Tsinghua, NVIDIA, CMU and Berkeley exposes gaps in streaming video generation across 38 models — jiqizhixin · 2026-11-10