Here's what I was able to gather from watching the video. (It's short — you should watch it before reading the below!)
The tool seems to have a 3-pane workspace. On the right is a live-updating preview of the artwork you create. On the top left is a workspace for drawing vector graphics, with lots of extra widgets and annotations and such for manipulating the graphics you draw. On the bottom left is a node-and-wire editor, where (as far as I can tell) each vector graphic element you create the top left pane gets a corresponding node, exposing some programmatic attributes for that element. There's also a left sidebar with various drawing tools and a list of common functions (measurements, data type conversions, etc). An interesting idea seems to be that the vector graphic you construct in the top left is actually a "function", and you can effectively call other functions inside the definition of a new function by nesting them. This gives you some "step and repeat"-like behaviour as a natural consequence.
Putting on my "visual programming critic" hat for a sec, this looks superficially like Toby Schachman's Cuttle, but with a GUI that seems heavily inspired by 3D / compositing apps like Blender, Houdini, or something from Autodesk. The video has no link to a website, no way to follow the progress of the project. I tracked this guy down on LinkedIn a few years back, and at the time he said he had no plans to keep working on DeepUI stuff, thinking instead that something like Enso was a more promising direction for visual programming. In any event, I'm really glad he's back and made this video, even if nothing else ever comes of it. I went frame-by-frame through parts of this video. There's some interesting thinking on display here.
https://www.youtube.com/watch?v=7SOv9zntnMY
Here's what I was able to gather from watching the video. (It's short — you should watch it before reading the below!)
The tool seems to have a 3-pane workspace. On the right is a live-updating preview of the artwork you create. On the top left is a workspace for drawing vector graphics, with lots of extra widgets and annotations and such for manipulating the graphics you draw. On the bottom left is a node-and-wire editor, where (as far as I can tell) each vector graphic element you create the top left pane gets a corresponding node, exposing some programmatic attributes for that element. There's also a left sidebar with various drawing tools and a list of common functions (measurements, data type conversions, etc). An interesting idea seems to be that the vector graphic you construct in the top left is actually a "function", and you can effectively call other functions inside the definition of a new function by nesting them. This gives you some "step and repeat"-like behaviour as a natural consequence.
Putting on my "visual programming critic" hat for a sec, this looks superficially like Toby Schachman's Cuttle, but with a GUI that seems heavily inspired by 3D / compositing apps like Blender, Houdini, or something from Autodesk. The video has no link to a website, no way to follow the progress of the project. I tracked this guy down on LinkedIn a few years back, and at the time he said he had no plans to keep working on DeepUI stuff, thinking instead that something like Enso was a more promising direction for visual programming. In any event, I'm really glad he's back and made this video, even if nothing else ever comes of it. I went frame-by-frame through parts of this video. There's some interesting thinking on display here.