In response to a question about autonomous agents creating scenes in Blender and how long until AI can complete a series without human interaction: “Claude 3.7 can do that already but the scenes aren’t great. Feeding them into vtv + first frame as style reference can lead to promising results. Today this isn’t quite production ready yet, but it’s an impotent step to getting there. I see a ton of immediate utility once Luma , ltx, wan and others get full v2v working - then we can pass render jobs frame output to this process for post / comp work in 2.5d with full artistic control and all the benefits of path tracing as well” @JulesUrbach 13.03.25 $RENDER