> In order to put the emission peak in the visible spectrum, the filament would need to run at ~5700 °C (...) No metal can survive these conditions: Tungsten melts at "only" 3422 °C.

Hmm, can we possibly find something that does survive at 5700 °C to use as a light source? Perhaps molten metal confined in a magnetic field...