The Devs recently added support for templates which can handle dynamic layouts (i.e drag and drop). https://panel.holoviz.org/reference/templates/React.html#tem...
Here's an example that uses almost entirely higher level components: https://dash-bootstrap-components.opensource.faculty.ai/exam...
We've also been working on `dash.templates`, which provide opinionated, prebuilt UIs - no layout code required: https://community.plotly.com/t/introducing-dash-labs-dash-2-...
For more complex 3D objects, Dash users can use dash-vtk. This includes things like point clouds, CFD simulations, 3D mesh, or 3D images.
So I'd be curious to see if encouraging certain conversational behaviors might actually improve the reasoning and maybe even drive towards consensus.