Hands-on: Astra drives Blender to make 3D motion graphics, but still needs heavy hand-holding
builderjaydub · x · 2026-09-19
The author asked Astra to study an award-winning reference, reproduce several scenes, and render a 20s video entirely in Blender. Verdict: not professional-grade yet, but on the right track.
- The experience feels like vibe coding in 2025: functional results, but without constant hand-holding the output is complete slop
- Prompts evolved from "build this scene" to detail-level nudges like "the middle piece needs more movement", "surfaces are too dark", "light should feel like it's spilling from inside"
- The final video still looks like a school project
- Astra clearly leads on Blender and opens a new growth curve — 3D modeling, spatial reasoning, computer use — that matters for future model generations; capabilities are far from saturation
More from Multimodal
- Free course covers the 4 hard problems of production generative media agents — Al_Grigor · 2026-11-04
- PixelBench: testing multimodal models pixel-by-pixel instead of with QA — bdsqlsz · 2026-11-04
- Kitten-TTS 2 offers free voice cloning on laptops, no GPU needed — solyarisoftware · 2026-11-04
- "We'll Take Your RAM": a fully local-hardware cyberpunk meme from 2030 San Francisco — Wonderful-Yak-3348 · 2026-11-04
- Openfilm: open-source AI video agent turns coding agents into video makers — solyarisoftware · 2026-11-04
- Microsoft reopens VibeVoice: 10-second voice cloning, 90-minute local TTS — solyarisoftware · 2026-11-04