@RaymondWongPhD The tools that claim to replace will go down, as ultimately, creative artists need assistants to help them with repetitive tasks rather than a one-shot crap
@metropolisworld@iamlukethedev Well, a Gaussian Splat reconstruction is not a world, even though it looks impressive. You can see the live analysis of that in my posts
🚨 Meet 01C's 3D agent 'Amara'
Describe a world. Bring your own assets or let Amara generate them. You steer the vision; Amara builds the scene and thousands of articulated objects, fully editable, getting sharper every time you iterate.
More details below 👇
@ltx_model@nvidia Depends on what you call a video model. I believe a video based scene prediction is not a world model while it is very impressive. Some real interactable world creator landed today
https://t.co/xxloeFAHGJ
🚨 Meet 01C's 3D agent 'Amara'
Describe a world. Bring your own assets or let Amara generate them. You steer the vision; Amara builds the scene and thousands of articulated objects, fully editable, getting sharper every time you iterate.
More details below 👇
@ashkan01C Can't wait to test and rate it. When is it available? It seems to be on @UnrealEngine, is it gonna be only there, or extended to other platforms? It reads to me that it is not Gaussian Splatting-based; is it correct?
Final verdict:
Tripo Splat passes the demo test.
It mostly passes the workflow test too, but with caveats.
Amazing for fast, high-fidelity splat-based visuals.
Less ideal when you need editable materials, collisions, game-engine readiness, or universal mesh-based production.
So the takeaway:
Tripo Splat is fast, lightweight, and visually impressive.
It feels genuinely useful, especially when visual fidelity and speed matter.
But it is not a full replacement for mesh/PBR workflows yet.