Silver is a fantastic conductor, better than copper, but tarnishes far too quickly for contacts that are expected to actuate consistently over a prolonged period of time. This is a problem because silver oxides are poor conductors (as are copper oxides). Gold is an inferior conductor to silver and copper, but not by that much and it doesn't tarnish. But is it relevant? That depends on the technology you build the keyboard out of, but I think the answer is probably a firm "maybe".
As far as technology: there are basically three major varieties. One has a plastic membrane that separates two conductive surfaces; physical pressure forces the two surfaces in contact and actuates the key. These are usually fairly large contact surfaces and also are relatively sealed, but you can get situations where you need to press harder on the key to get it to actuate, and if little bits of crap get between the surfaces then the key is pretty well dead. There are a variety of ways of forcing that actuation; simple physical pressure (your microwave pad), a plunger (less prone to troublesome false actuations), the scissor switches you see in laptops or the modern Apple keyboards, but they're all basically similar as far as electrical conductivity and tend to fail in relatively similar ways.
The Model M is actually a membrane like this under the surface, but the membrane is almost totally sealed off and is operated by extremely reliable and powerful (well, for a keyboard) actuator--the buckling spring operates a small lever in the guts of the keyboard, pressing the devices in contact. As a result they get excellent lifetime and a very consistent and desirable action. The size of the lever and the distance between the key and the actuation (and also the sheer size of a Model M) mean that the contact surfaces can be relatively large. Earlier IBM keyboards used a capacitive action which is very rare today; the only one I know of is the Happy Hacking Keyboard.
The Cherry MX switches the Das uses have a mechanical switch that makes and breaks contact on every keystroke. They are mechanically more compact than the buckling spring type with similarly desirable key action, but the switch means they need to be made in a very durable way. Cherry says the MXs are good for 50 million operations; I don't know how many operations IBM guaranteed theirs for but I'm sitting next to one made in 1989 so I presume it is "a lot".
Which is a lot of words to say that "gold might make a difference and anyway Cherry doesn't sell good switches without the gold plating and they're the only people of any repute making those type of keys any more".
Edit: one other further irritation for people who type fast; N-key rollover is a term you may have heard of. On a PS/2 keyboard, you can literally activate every key on the keyboard at once and the computer can track them all. This is because the PS/2 tracks key up and key down events but doesn't try to keep state so much. USB does and because of this true N-key rollover is impossible without custom software on both ends; the best we can do is I think six, so in this case the Das is state of the art, such as it is. Supposedly the v4 Sidewinder got around this, probably through a custom driver mode.