Is this really a Mercator projection? It doesn't appear to maintain the invariant that lines of constant bearing are straight lines.
If I pick a point somewhere in the middle of Manhattan, the top point of Manhattan is somewhere near the top of the light colored area and the bottom point of Manhattan nearish the bottom of the light colored area. This means that if I draw straight lines on the the map from San Francisco to these two points, the angle between them is something like 30 degrees. They pass through very roughly the top and bottom of Nevada. But there's no line of constant bearing that passes from SF through the top of Nevada to the top of Manhattan while at the same time one that passes through the bottom of Nevada to the bottom of Manhattan.
This is all very wishy-washy, but it doesn't look right to me.
I think it's just a play on the fact that mercator distorts distances significantly, rather than actually being accurate. It's a 3-second website you open, exhale in hilarity and close.
"Lines of constant bearing" (or "rhumb lines") depend on the choice of poles.
A rhumb line relative to true north looks straight on a standard Mercator projection, but can look like a spiral on another Mercator-style projection where the pole and center-point have been swapped.