Obviously everyone can do their own thing, regardless of if it's based in reality or in Boomer facebook memes. Save a ton of teaspoons by filling up every morning from a Jerry can if you think air gap matters that much. It's the only logical thing to do. It's amazing stored vehicles don't just have empty tanks by the time owners come back for them. Like all those cars sitting for a year or so while their owners are on deployment? Tanks dry as a bone when they get back... I was gone for two years and still had a full tank upon return. Or does it have to 'slosh' to evaporate?
I can count on zero finger the number of gas tank explosions I've seen in vehicle crashes, and I've been on plenty.
*If you don't care about demolitions, feel free to skip the rest of this* Most everyone is familiar with the fire triangle, fuel/oxygen/ignition, but an explosion also requires the correct dispersion of the fuel and oxygen at the time the ignition source is present. If the fuel/air mix is in the correct range, defined by a lower and upper explosive limit (LEL / UEL), it explodes. If it's outside you either get combustion or you get nothing, depending. You aren't achieving that *and* having an ignition source inside a gas tank in a crash.
Gasoline fumes have a fairly tight LEL and UEL (say 2%-7% for easy figuring). It's simply not a realistic concern to have both an ignition source and correct dispersion inside a tank at any fuel level. The spill is a different matter. You can get a weak explosion (and a strong fire) from the fumes immediately over the spilled gas.
Now an "empty" one somebody is welding on... the fumes of an empty tank can get that 2%-7% comparatively easy. The tank is acting as containment, the welding torch is a steady source of ignition so there's more time for the dispersion to be correct, etc.