The Happiness Machine

[daghain]

We’re all familiar with the concepts of the Love Machine and the Sex Machine, but I am trying to picture what a Happiness Machine would look like. In my head it is a huge 50’s-sci-fi-style mass of wires and gear with a few vacuum tubes stuck on for nostalgia. The interface of this machine would be a comfy armchair for the subject to sit in, a

Comments

[leech]

The same amount of money counts for different amounts of happiness in different people's hands.

[daghain]

Sure, that's the principle that allows charities to function. But is the sum of many small losses comparable to one large gain?

[leech]

That's where I think the differences are important. Nobody's going to be too excited about Donald Trump sitting in the Happiness Machine. I claim that in those cases, pulling the lever decreases net happiness.

(Assuming that all these distributions and happiness functions really exist, which I don't; that requires rather a lot of faith in human rationality.)

[daghain]

Nobody's going to be too excited about Donald Trump sitting in the Happiness Machine.

Sure, but that's envy. People wake up in the morning and read in the paper that Donald Trump made another million for reasons that have nothing to do with them at all, and they still feel unhappy. Take envy and altruism out of it and suppose nobody knows about the Happiness Machine. They just wake up are shy half a penny.

Now if the recipient is Trump, then sure, he'll be less happy than a destitute guy would be. But I don't see why it follows that the sum total of unhappiness from using the Machine is somewhere between Trump's happiness and Homeless Guy's. That would make sense if you viewed the Happiness Machine transfer as the sum of 299,999,999 seperate half-cent transfers (all to Trump, all happiness-destroying). But precisely because people aren't perfectly rational, we *can't* do that. I know I'm happier to get an $800 tax refund than I would be if I just got $30 more every payday, and so it would be with the Machine.

[blazepoet]

The way I see it, the process starts to break down when it becomes a noticeable drain on some of the population.

I don't know what price that would be probably 20-50$ because then the net change in happiness would be measurable and the hundreds of millions of small sadness' would overwhelm the few happiness.

Of course the lottery works similar to this.... of course the track record of lottery winners does not support how long the happiness from the happiness machine would last.

[daghain]

Alright, let me geek this up for you.

Let f(x) be the Happiness Machine function that maps an allocation x into the same allocation plus a Machine activation.

Let u(x) be a measure of the amount of happiness generated by a resource/money allocation x. If X is the allocation we have now, it seems pretty plausible to say that u(f(X)) > u(X).

But we know that f^3,000,000(X) = X, so u(f^3,000,000(X)) = u(X). If u(f(X)) > u(X), then u(f(X)) > u(f^3,000,000(X)). Thus there must exist an x such that u(f(x)) < u(x). But I have a hard time thinking of an x for which that would be true. It's much easier to think, as you are, of how u(f^5,000(x)) < u(x). But there has to be one particular activation of the machine that *by itself* decreases happiness in the world, and I can only imagine such a thing happening in the context of many other activations

[anonymous]

As mentioned in the first comment, the assumption that $1 is an identical value to all people is a fallacy. This is something I wish weren't true, but seems undeniable. Thus, much like taking a glass of water from me and giving it to someone just out of the desert would likely increase net happiness, taking $1 from me and giving it to someone who can't afford food would likely be a net gain

[daghain]

Sorry for the late response, but it felt kind of odd to see an anonymous comment only hours after my first posting in months. Do I know you?

Econ deals with the problem in two ways (and justifies the two in a plethora of ways). One is just to say that $1 is worth the same to everyone simply because we don't have another convenient yardstick to compare one person's loss to another's gain. The other is to declare peoples' utilities incomparable and only focus on Pareto Improvements (allocations that improve one person's lot without hurting anyone else).

I do feel silly for forgetting about altruism, but my point was basically to take a not-necessarily-obvious economic assumption (that 100,000 people made poorer by half a cent exactly balances one guy made richer by $500) and show a strange thing that would happen if it didn't hold. I purposely did this without telling anyone the assumption I was indirectly defending, and it's interesting to read the responses.

pushing the limits of an-cap, if I may

[zaxec]

I'm not sure this contribution is warranted or on topic, but you are ignoring something which I feel is perhaps important

Re: pushing the limits of an-cap, if I may

[daghain]

I agree that some people (virtually *all* people, if we don't include altruism) would opt out of having their half-cent taken if asked. And yet I'm claiming that each pull is a Pareto Improvement. So what gives