We don't typically find gold in 1kg lumps. IIUC it is quite sparse even in the ore processed in gold mining. Surely running a bunch of connectors through an appropriate purification process would be cheaper than processing a similar volume of ore? If it's not, that tells us something too.
I predict this will be the real killer app of self-assembling and -maintaining nanomachines: processing landfills into ingots of useful material. That's a pretty long range prediction however.
If markets are working more or less efficiently, digging natural gold must still be cheaper. There's a lot of toxic stuff in landfill and it's usually located close to population centres.
Gold-harvesting nano robots would be awesome. Just don't let those suckers loose downtown...
Yeah, that's kind of what I was alluding to with my "that tells us something too" comment. We can't assume that markets are "efficient" in the general case, but we can say that there must be some point at which the ratio of the concentration of gold (or platinum, etc.) in landfills to that in mines will prove irresistible to someone talented enough to harvest it. Long before that point, recyclers will figure out which connectors they want to buy at a premium. They already are quite keen to purchase old catalytic converters.
WRT letting nanomachines loose outside the landfill, I'm not actually as concerned about the "gray goo" scenario as some are. It has seldom been a problem in other engineering endeavors that mechanisms have required too little maintenance, and so in a sense all of our machines face an uphill battle. One could probably make a thermodynamic argument that a given volume of soil contains only so much potential energy, and then calculate how much gold can be moved with that. Then "if the gold concentration is below this value, do nothing" could be a good safety rule to design in.
The gray goo scenario requires self-replicating nanomachines, so not applicable to harvesters that can't replicate.
The problem of where the energy comes from for micro- and nanomachines is very tough. But if you can make the landfill harvesting work, no doubt they'd harvest the hell out of the electronics in an office tower.
Robot foraging and energetics is my area of expertise. Pleasantly surprised to see a cogent comment come up on HN, thanks and nice to meet you!
I've been to the highest yielding part of one of the biggest gold mining operations in the world. I went 2.3km under ground, and down 6km of tunnels horizontally, to a part where they were getting 40 grams of gold per cubic ton of what is essentially hard rock.
I predict this will be the real killer app of self-assembling and -maintaining nanomachines: processing landfills into ingots of useful material. That's a pretty long range prediction however.