Are diode models like totally broken or should I ignore this warning or ... what should I do if I need to use a diode?

See https://open-source-silicon.slack.com/archives/C016HUV935L/p1664403409640839?thread_ts=1663972071.204679&cid=C016HUV935L

Hmm what should I do while the maintainers deliberate their patch?

I am pretty confident that the actual units are very weird and I have a complicated hypothesis about why that is involving the weirdness of the level 3 diode and the fact that depending on what parameters are given to it, the parameters can be interpreted with wildly different scale factors; that is further complicated by somebody's unfortunate decision a long time ago to put this 1e-6 global scalefactor into the device models. According to my hypothesis,

area=1e+12

(units = picometers squared) actually is correct for a 1um x 1um area diode. I currently have the tech file for magic using this interpretation of the diode scaling (area in square microns is multiplied through by 1e+12 while perimeter in microns is multiplied through by 1e+6).