Oh, well done - that certainly wins my award for most awful tesselation of the year 🙂 I'm sure Rhino can do better than that if it tries !
Even the almighty Z-ReMesher is going to struggle with that.
I haven't looked at HP's plan for rescuing it, but mine begins with deleting the thickness polys leaving a flat layer, doing UQ to select the outline of that, and extracting a spline (edge to spline) from it, which is then going under an Extrude, where, TBH I am probably going to stick with the ngon option - the surface is entirely flat, so as long as it produces a renderable surface - the triangulation going on behind the scenes is still likely to be moderately better than what Rhino did (let's face it - it couldn't be worse!) and if we don't have to look at ngon lines, then they need not worry us !
However if we do care about nice(r) topology, then the delaunay options in the extrude caps can work for us if we set the source spline to be uniform and a low point value such as 1, and choose Quad-dominant and suitable density...
CBR