Conversely, because VNM utility is out here, axiomized for the sovereign preferences of a single agent, we don’t much expect it to show up in there, in a discussion if utilitarian preference aggregation. In fact, if we do encounter it in there, it’s probably a sign of a failed abstraction barrier.
Money is like utility—it is society’s main represenation of utility.
Be careful, money is an economics and game theory thing that can give you evidence about people’s preferences, but I would not directly call it a representation of utility. It is likewise not directly relevant to utilitarianism.
I’d like to take a crack at discussing how money (and fungiblish consumablish resources in general) relate to utility, but that’s a big topic on it’s own, and I think it’s beyond the scope of this article.
Money is a form of value. It has an equivalent to wireheading—namely hyperinflation. And we have maximisers of share value—namely companies. So: money is a kind of utility—though obviously not the utility of people.
Are there any expected-money maximizers in the world? (risk aversion is not allowed; utility is linear in utility, so if money is utility, it must have linear utility. Does it?)
Does anyone value money for its own sake? Or do they value what it can buy?
Is money a quantity associated with an entire world-history?
It seems accurate to say that it’s treated in a utility-like way within certain incentive systems, but actually calling it a form of utility seems to imply a kind of agency that all the money-optimizers I can think of don’t have. Except perhaps for automated trading systems, and even those can have whatever utility curves over money that their designers feel like setting.
Not really, no. They have goals in the sense that aggregating their subunits’ goals gives us something of nonzero magnitude, but their ability to make plans and act intentionally usually seems very limited compared to individual humans’, never mind well-programmed software. Where we find exceptions, it’s usually because of an exceptional human at the helm, which of course implies more humanlike and less money-optimizerlike behavior.
Where we find exceptions, it’s usually because of an exceptional human at the helm, which of course implies more humanlike and less money-optimizerlike behavior.
Right. So, to a first approximation, humans make reasonable money-optimizers. Thus the “Homo economicus” model.
I think it is perfectly reasonable to say that companies have “agency”. Companies are powerfully agent-like entities, complete with mission statements, contractual obligations and reputations. Their ability to make plans and act intentionally is often superhuman. Also, in many constitutuencies they are actually classified as legal persons.
So, money is a representation of utility. Representations of utilities don’t have to be “linear in utility”. I already said “obviously not the utility of people”, so whether people value money for its own sake doesn’t seem very relevant.
Perhaps a better point of comparison for money would be with utility-related signals in the brain—such as dopamine.
I don’t like having to say “representation of utility”. Representations are all we have. There is no utility apart from representations.
It has an equivalent to wireheading—namely hyperinflation.
This is a difference to utility. Not a similarity. Wireheading gives low utility (for most plausible utility functions) but huge measures for other things that are not utility, like ‘happiness’. It is the reason it would be utterly absurd to say “The federal government can print arbitrarily large amounts of utility”.
And we have maximisers of share value—namely companies.
You can approximate (or legislate) companies that way and. It wouldn’t be quite as inaccurate as saying “we have homo economicus” but it’d be a similar error.
So:
The statements following “So” do not follow from the statements preceding it. The preceding statements are respectively negatively relevant and irrelevant. So “So” does not fit between them.
money is a kind of utility—though obviously not the utility of people.
There is a relationship between money and utility. It is not an “is a kind of” relationship. (If Nyan takes a crack at explaining what the actual relationship is between fungible resources and utility it will most likely be worth reading.)
So: I was using “utility” there to mean “representation of utility”. In fact I did previously say that money was a “representation of utility”.
This is a case where there are only really representations. Utility is defined by its representations (in its “that which is maximised” sense). Without a representation, utility doesn’t really exist.
To be more sepcific about about hyperinflation, that is a form of utility counterfeitting. It’s on the “wireheading” side, rather than the “pornography” side. This isn’t really an analogy, but an instance of another phenomenon in the same class. Hyperinflation produces poor outcomes for countries, just as wireheading produces poor outcomes for those that choose it. This is a similarity—not a difference. I am not sure why you don’t recognise the relationship here. Are you sure you that have thought the issue through?
Money is like utility—it is society’s main represenation of utility.
Be careful, money is an economics and game theory thing that can give you evidence about people’s preferences, but I would not directly call it a representation of utility. It is likewise not directly relevant to utilitarianism.
I’d like to take a crack at discussing how money (and fungiblish consumablish resources in general) relate to utility, but that’s a big topic on it’s own, and I think it’s beyond the scope of this article.
Money is a form of value. It has an equivalent to wireheading—namely hyperinflation. And we have maximisers of share value—namely companies. So: money is a kind of utility—though obviously not the utility of people.
Is it?
Are there any expected-money maximizers in the world? (risk aversion is not allowed; utility is linear in utility, so if money is utility, it must have linear utility. Does it?)
Does anyone value money for its own sake? Or do they value what it can buy?
Is money a quantity associated with an entire world-history?
It seems accurate to say that it’s treated in a utility-like way within certain incentive systems, but actually calling it a form of utility seems to imply a kind of agency that all the money-optimizers I can think of don’t have. Except perhaps for automated trading systems, and even those can have whatever utility curves over money that their designers feel like setting.
You don’t think economic systems have “agency”? Despite being made up of large numbers of humans and optimising computer systems?
Not really, no. They have goals in the sense that aggregating their subunits’ goals gives us something of nonzero magnitude, but their ability to make plans and act intentionally usually seems very limited compared to individual humans’, never mind well-programmed software. Where we find exceptions, it’s usually because of an exceptional human at the helm, which of course implies more humanlike and less money-optimizerlike behavior.
Right. So, to a first approximation, humans make reasonable money-optimizers. Thus the “Homo economicus” model.
I think it is perfectly reasonable to say that companies have “agency”. Companies are powerfully agent-like entities, complete with mission statements, contractual obligations and reputations. Their ability to make plans and act intentionally is often superhuman. Also, in many constitutuencies they are actually classified as legal persons.
So, money is a representation of utility. Representations of utilities don’t have to be “linear in utility”. I already said “obviously not the utility of people”, so whether people value money for its own sake doesn’t seem very relevant.
Perhaps a better point of comparison for money would be with utility-related signals in the brain—such as dopamine.
I don’t like having to say “representation of utility”. Representations are all we have. There is no utility apart from representations.
This is a difference to utility. Not a similarity. Wireheading gives low utility (for most plausible utility functions) but huge measures for other things that are not utility, like ‘happiness’. It is the reason it would be utterly absurd to say “The federal government can print arbitrarily large amounts of utility”.
You can approximate (or legislate) companies that way and. It wouldn’t be quite as inaccurate as saying “we have homo economicus” but it’d be a similar error.
The statements following “So” do not follow from the statements preceding it. The preceding statements are respectively negatively relevant and irrelevant. So “So” does not fit between them.
There is a relationship between money and utility. It is not an “is a kind of” relationship. (If Nyan takes a crack at explaining what the actual relationship is between fungible resources and utility it will most likely be worth reading.)
Thanks for the encouragement! I do plan to do that soon, but I am hardly an agent that can be described as following through on “plans”.
So: I was using “utility” there to mean “representation of utility”. In fact I did previously say that money was a “representation of utility”.
This is a case where there are only really representations. Utility is defined by its representations (in its “that which is maximised” sense). Without a representation, utility doesn’t really exist.
To be more sepcific about about hyperinflation, that is a form of utility counterfeitting. It’s on the “wireheading” side, rather than the “pornography” side. This isn’t really an analogy, but an instance of another phenomenon in the same class. Hyperinflation produces poor outcomes for countries, just as wireheading produces poor outcomes for those that choose it. This is a similarity—not a difference. I am not sure why you don’t recognise the relationship here. Are you sure you that have thought the issue through?
Something like “Money: The Unit of Caring” by EY?
Money is like utility but different.