Self-defeating hypotheses

[fare]

"If AIs become better and cheaper than humans at EVERYTHING, humans will stop interacting with each other. Pan-human catastrophe!"

"If foreigners become better and cheaper than nationals at EVERYTHING, nationals will stop buying from each other. National catastrophe!"

Yeah, right. And if people outside your immediate family do everything

Comments

[selfishgene]

'universal acceptance of the principles of property rights' will only happen when everyone is well armed enough that respecting their rights is easier than crushing them.

[fare]

Not everyone, only everyone that matters. If AIs are our superiors, it only matters that THEY be armed, for instance. And if they are superiorly intelligent enough, they will understand the principles of praxeology.

[randallsquared]

This doesn't really address the hypothetical situation where an AI would value the raw materials humans consume more than anything humans can produce from those materials, right down to humans themselves. Selling the matter you're made of is self-defeating.

A more fundamental problem I'm having lately with some of these arguments, though, is that they imply that we should never have lasting, involuntary unemployment, since there should always be *something* that those people could do. Depending on who you believe, there are between 15 and 50 million people just in the US who want work but can't find it. This is what David D. Friedman recently called, on his blog http://daviddfriedman.blogspot.com/, "the sort of situation that price theory [microeconomics] tells us can't exist but that sometimes does".

[fare]

It *does* address this situation: if the AIs respect property rights, they won't violates those of humans, even when humans make poor use of their resources. Just like I think other people make poor use of their time, food, computers, etc., but I don't go enslaving them, starving them, getting their cycles away, etc.

DDFR qualifies this sentence saying that he's "loosely speaking". He is. And price theory tells us exactly why people can't find work when there's price control and barriers to transactions (= taxes).

[randallsquared]

"if the AIs respect property rights, they won't violate those of humans

Unemployment

[fare]

Add to direct, known taxes the unknown taxes tied to the current future "health insurance" regime. Add disincentives to work from work loss "insurance", dole, and various aids to the poor. Add the certainty of a tax regime change for the worse, but a complete uncertainty as to the details. Add a trillion dollar of "stimulus" distortion in the economy that completely skews supply and demand so that productive ventures cannot find customers, whereas unproductive ventures are funded.

There are lots of reasons today to not be either hired or hiring.

A vision of far-future relations between AIs and humans...

[fare]

Simon Funk's "After Life"

Shrödinger's Robot

[fare]

Like Shrödinger's immigrant who is simultaneously both working and stealing jobs, and not working and stealing welfare, Shrödinger's Robot is both so cheap that everyone uses his services and not each other's, and yet so expensive that no one can afford his services.