Log in

No account? Create an account
eyes black and white

Self-defeating hypotheses

"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 better and cheaper than your family as far as you're concerned, you'll never talk to your spouse and kids again? Family-zastrophe!

Let's say I'm one of the persons with whom you admittedly share some close bond; now what if people other than the two of us satisfy our needs better and cheaper than the two of us can — we will stop interacting with each other, and will it be a catastrophe for the two of us as we'll both lose all the precious things we bring to each other? Pair-zastrophe?

The fallacy is that on the one hand, the denounced Third Party is supposed to better than anything, and at the same time, the lost relationship is supposed to be something so valuable that it's vastly better than anything else. But you can't have it both ways. As in the story of universal solvent that can dissolve anything and the universal container that can resist any solvent, one has to give way to the other. And as in the story, both claims are dubious at best.

If indeed the Third Parties are so good and cheap that we'll both refer to them and stop interacting with each other, then by definition, we're having more fun, more satisfaction, by interacting with the Third Parties rather than with each other. And for whom is it a catastrophe? For neither of us. By assumption, we're both better satisfied this way. Maybe I'll marry one of those beautiful, sexy, intelligent and agreeable strangers instead of marrying you; and you'll similarly marry one of them instead of marrying me. And by assumption, we'll both be happier than if we had been together. Yay, life!

Or is the bond between us so strong and so valuable that nothing that the Third Parties may offer is worth dissolving that bond? Then why are you afraid that we'll destroy this bond? Who are you calling stupid? Yourself, me or everyone else? Are you going to leave your family behind to starve to death because the Matrix is offering you a more pleasurable though virtual family? Why do you think anyone would and make that the premise of your catastrophic prediction?

Are you claiming that most people are stupid or evil, except yourself and the vis-à-vis you're trying to convince of your views with such arguments? Besides the remarkable conceit you display, you should realize the vanity of the attempt to save such stupid or evil people, especially when the strategy you choose would be to either convince them all through an appeal to their intelligence and morality that you deny exist, or to somehow make yourself their dictator.

And while you meditate at the impending doom of humanity or your nation, etc., you may consider the miracle by which this flock of stupid or evil people reached that state of affairs that you're so afraid to lose; here is a miracle you obviously can't fathom, and which should give you faith that some force is at work that you have failed to identify so far.

For this whole scare of machines, of foreigners or of any other third party is but an ingrained reflex of defiance against alterity and a defense of existing known or supposed relationships taken to the point of neurosis. Sadly, there is no shortage of crooks who will excite other people's neuroses to profit from them. Happily, however, the scare is wholly unfounded.

So what happens when machines or foreigners get better and cheaper at something? The general answer is: the law of comparative advantage applies. We get more of what the third party provides, in exchange for which we do more of what the third party requires; by hypothesis, this costs us less to produce than previous and/or other available alternatives, and we enjoy the fact that the third party offers us a better deal than we used to have before (or else we wouldn't have switched). Getting more for less, that's the only reason why we ever voluntarily switch from one arrangement to the next.

How did that work in the past? When washing machines, vacuum cleaners, gas stoves, etc., freed hours of housework for women (mostly), does that mean that women (and men) lost the ability to wash clothes, clean the house, cook a dinner, etc., and are now running the risk of going naked, dirty and famished? No, it means that in addition to having clean clothes, clean houses, and quick dinners, we have all that we can do with the free time gained: more gratifying jobs, healthy activities, cozy time with our families, etc. Or sure, getting stoned in front of the TV if that's what you're into. All that for the investment of a measly priced machine and for a little bit of cheap water, electricity, or gas. If you don't want to invest yourself, you can still enjoy the benefits of technology based on other people's investment: go to the laundromat, hire a cleaner or buy takeout.

Similarly, when we start getting cheaper clothes from here, cheaper robots from there, cheaper food from another place, etc., it doesn't mean we are losing anything to foreigners; we are gaining all they offer to us, in exchange from what we offer to them, which by very definition costs us less than it would cost us to get the same things ourselves. So we get cheap fruits from some southern country? It means we offer them some manufactured goods or something in exchange, that is cheaper for us to create than those fruits would be, and is more expensive for them to create than those fruits are. Both parties benefit and enjoy the additional free time from the trade.

And what if at some point, we become unable to pay them with something they desire? Then in this worst case scenario, we'd be back to the point where you'd like to force us to be now: we wouldn't trade, and would have to get the desired thing by ourselves. Really, what you propose to achieve through some liberticidal and onerous use of force — the prohibition of trade with some maligned third parties — is actually but the worst possible outcome of what would happen for free if everyone's freedom to trade was respected.

Are you now claiming that in a temporary deal like that, we'll have lost our ability to create and will eventually find ourselves poorer? Well, now you're calling others stupid or evil for their lack of foresight, but it's the same unwarranted conceit as above. You're claiming you can see further than anyone else and are demanding that others be submitted to your brilliant schemes, by appealing to superior foresight of the very people you're calling incapable of such. Yet, expectation of change in future scarcity is already counted in the price of goods and the price of investing in the future production of such goods. When it appears that the third party source of some good will dry up, people stockpile, the price starts going up before the source is over, and investment in alternative sources starts. Once again, at the very worst we'll be back to where we were before, after having enjoyed a free ride for the price difference while it lasted.

Is your claim that you can actually see further than everyone else? Well, put your money where your mouth is, then: invest at the time that you think other people are failing to, and reap the profits. You can then both rightfully boast of your superior foresight, be proud of having accomplished a good action (saving "your people" from a dirth of the misanticipatedly lacking good) and use the proceeds to advance more of your ideas. The more people you claim agree with you, the easiest such an investment should be.

Is your claim that the third party, once strong enough, will crush you? Well, said third party, by hypothesis, is interested in what you're offering for its services, and finds the trade useful. As long as the trade is useful, it has no interest in fighting you. And if for whatever reason it ever becomes stronger indeed, advocating as you're doing the forceful prohibition of trade with the third party is setting a very bad precedent for what will happen to you: since you accept the principle that might makes right, you'll soon enough be victim of that principle. Instead of calling for forceful prohibition by the central authority of whoever happens to be the strongest now, you should call for the respect for property rights. Where these rights are universally recognized, the third party won't attack you even when vastly stronger than you, for fear of having to respond to other similar third parties wary of the respect of rights. Where property rights are sacrificed to the altar of some collective welfare, you'll soon be the victim of a change in who controls enough force to claim unopposed to represent the collective.

Are you claiming that humans, once obsoleted by AIs, will disappear like draft horses did? But there are infinitely more horses in America now than before slave horses were introduced on that continent by white men. And those horses that exist now are probably happier than ever were their ancestors that toiled at the height of the horse enslavement racket. Moreover, horses are miserable precisely because they are animals without rights, without rights because they are unable to petition for the rights they are being denied, unable to make and respect covenants that delimit their and another party's respective rights, etc. Humans can negotiate, respect and enforce rights, and are therefore susceptible to be recognized mutual property rights.

Do you mean that ultimately AIs will acquire most resources in the universe, and leave humans with but what they have, that will only sustain them so long before they starve out and die? But if property rights are respected, then by very hypothesis, humans through voluntary trade will only ever but get more than they would have without machines. They may eventually starve and die out, but without the peaceful interaction with machines, by very hypothesis, they would have starved and died out even sooner: whatever extra resources they would have had, they would not have been able to make as good use of them, and would have extracted a shorter and less agreeable life from them — which is the very reason why they accepted to trade those resources with the machines. Of course, if interactions are not peaceful but warlike, then things could go wrong for humans, but the same is true when the war doesn't involve machines, and the damage is to be ascribed to war, i.e. the denial of mutual property rights, rather than to the advancement of machines as such. And once again, the proposed "solution" is to start now and for certain a war that you fear might happen in the future, making the worst imaginable outcome an effective certainty. Instead, the actual solution is in the universal acceptance of the principles of property rights as being most sacred.

In the end, what if machines actually become so good and ultra cheap that they replace us in about all the jobs we can do? Then by definition, at the cost of almost nothing (the cheap price in question, that we pay to them), we get all the free time in the world, to do whatever it is we really want: whether it's reading books, having sex with cyborgs, or raising actual human kids. And we'll keep exchanging services with each other, so that those few who are most able to get what the machines want will be those who produce it, while other people offer all kind of human services, from psychological support to massages to entertainment to whatever the hell we'll desire when we're free from all the hassles overcome through machines. More people doing agreeable human jobs, fewer doing horrible mechanical work, isn't that the essence of progress?


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

But what's the probability of AIs respecting (human) property rights? If we examine how humans treat the territories of entities far less capable than we are, the chances don't look good. It isn't really that I'm arguing that there's anything to be done about this -- I've never heard a plausible argument for something that could fix it -- but I do expect humanity's run to be effectively over as soon as superintelligent AI is built.

As for price control, other regulation, and taxes being the answer to why we have 15-50 million people who want to work and can't find jobs, it's not so clear to me. Even after the huge hit to the economy that the federal, state, and local governments represent, the average USian is far wealthier in real terms than the average USian in, say, the 1950s, which was a boom time, mostly. Further, the highest tax rates then were far higher than now, too. Government interference in the usual way is certainly one piece of the puzzle, but I don't see how it can be all of it, since technology has provided more growth of real wealth than the US government has soaked up. This isn't to say I have any other answer, though. I'm pretty confused about it.


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.

Libertarian AIs

AIs, with their even greater and more obvious inequalities, and faster interactions, will need rules to avoid conflict even more so than humans. They may quickly come to adopt property rights as the natural way to minimize and resolve conflict with each other. The mechanism trivially extends to humans, who can petition for rights and respect the rights of others. AIs who trample on the rights of humans may be seen as dangerous by other AIs, and may forfeit their own rights wrt their better and faster AIs -- a very bad signal to send.

Re: Libertarian AIs

I'm preparing another essay on Evolution and Law, Immunity and Identity on the Rapacious Hardscrapple Frontier.

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

Simon Funk's "After Life"

Edited at 2015-04-25 03:54 pm (UTC)

Shrödinger's Robot

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.

eyes black and white

January 2018



Powered by LiveJournal.com