Comment by natural219 | Hacker News Reader

Comment by natural219 | Hacker News Reader

Parent Full thread

Eliezer, as always, vastly overstates the value of his own intelligence and intelligence in general. I believe this essay is one of the worst he has ever written.

The reasoning on this is correct. The problem is, he is arguing from hilariously inaccurate axioms. For instance, he seems axiomatically convinced that "creating trillions of dollars of value" is an objective good. That is totally fine for people ideologically "all in" to ideologically cooky notions of macroeconomics, which is one of the most intellectually shaky disciplines still in circulation.

Since he seems familiar with the "two economists walking down the street" parable, I will offer him this one:

===

Two economists were walking down the street one day when they passed two large piles of dog turds.

The first economist said to the other, "I'll pay you $20,000 to eat one of those piles of turds." The second one agrees and chooses one of the piles and eats it. The first economist pays him his $20,000.

Then the second economist says, "I'll pay you $20,000 to eat the other pile of turds." The first one says okay, and eats the turds. The second economist pays him the $20,000.

They resume walking down the street.

After a while, the second economist says, "You know, I don't feel very good. We both have the same amount of money as when we started. The only difference is we've both eaten turds."

The first economist says: "Ah, but you're ignoring the fact that we've engaged in $40,000 worth of trade!"

Disagree with your view of the essay, but your anecdote is interesting. I learned something from analyzing the anecdote with inappropriate sincerity.

Why- of their own free will- would each economist offer $20,000 to eat the turd? There can essentially only be one answer. They value the schadenfreude or sense of power derived from watching their fellow economist eat the turn at greater than $20,000.

Why would each economist eat the turd? They value an extra $20,000 more than they value avoiding the disgust associated with eating the turd.

Long story short, that $40,000 of seemingly destructive trade appears to have improved both their lives according to their own value functions.

This story is a remarkable reductio-ad-absurdum of the Subjective Theory of Value.

No, it isn't. It introduces a non-standard situation where the actors' motivations are hard to model, and results in a non-standard outcome. That doesn't help build the firmer X-free worldmodel that I would count as a refutation of X.

It requires us to posit someone who's money-hungry enough to do something disgusting for a large amount of money and yet who still would spend all of that gain to get some (for-most-people) trivial kick out of seeing someone else do the same.

Well, yes, if two people both had that bizarre preference set, and you fully intuited the implications of that situation, it wouldn't be strange to view it as creation of that much value, because both of them satisfied each other's values.

I don't see how this parable carries more kick than if you cut off the parable halfway through and said, "hah! Someone ate feces for money! That's silly, isn't it?" Yes, bizarre values lead to bizarre outcomes. But then, if that's all you want to prove, then just "go Rule 34" and link some fetish sites.

To the extent that the parable has a valuable lesson, it's "back-and-forth exchange of money for services, if officially canceled out, is a kind of gain in utility that doesn't show up in formal accounting". (E.g. Alice provides romantic satisfaction to Bob, who simultaneously provides romantic satisfaction to Alice. They do not transfer money on net. "We're right where we were before, and didn't create any value." Um, yeah you did.)

Or, if the point is that decisions can be regretted (which you could tweak the Alice/Bob example to include) , sure, but that doesn't shed a lot of light on the Subjective Theory of Value.

It seems abundantly clear to me that humans routinely take actions that are far more irrational than eating a turd for $20,000. There are literally TV shows dedicated to that type of thing.

The whole point is to draw absurdity to the Subjective Theory of Value -- the claim that, if a trade occurs, the trade must be Good, because both participants (with full information!) made a rational (!) calculation that this trade was in their best (!) long-term (!) interests. I find this proposition -- one that rests at the heart of economics -- to be incredibly dangerous and outrageously wrong, but since our entire economy / political system sort of depends on this being true, I don't know what to do about it.

The weird step in the example isn't eating a turd for $20000. It's paying 20000 to see someone eat a turd.

That the thought experiment is badly constructed is clear once you take the money and indirection out of the picture. Does two - clearly somewhat weird - people agreeing to eat a turd if the other person eats a turd, somehow disprove economics?

Clearly for each of the two people it's the case that they get more from watching the other person eat a turd than it costs them to eat one. Maybe they're perverts. Whatever. Value was actually clearly produced.

I was replying to the idea that the parable demonstrates those points for someone who doesn't already agree. (And that's not the SVT you're describing, but the SVT plus some ancillary claims that I agree are flawed -- see the last paragraphs of my other reply.)

I'm curious to pick your brain about this. What is a better concept to describe the idea that "if a trade occurs, it is Good?" I agree STV does contain real insights about the nature of trading, but there's some ideological extrapolation that transforms obviously correct things like marginal diminishing utility to "justifying every trade as it occurs as inexorably good"

See also: http://www.paulgraham.com/wealth.html

You're actually right, I think -- I don't know of a specific (standard) term to describe that fallacy, and every time I've criticized its appearance, I've had to make up a neologism or spell it out expliclity [1]. ("Vulgar Keynesianism" is one term for it, but it's not the only school that does it so it's a fair term for the precice concept.)

In any case, I was just trying to make the point that

a) The turd-trade example does not help the debate about the underlying concepts -- regardless of label -- because it's uses an intuition breaking situation to build intuitions, which is no more insightful than pointing and laughing at people with bizarre preferences.

b) STV is not the same as the thing the parable is criticizing.

[1] Example comment I remember: https://news.ycombinator.com/item?id=12242525#12243088

Man. Thank you so much. "Vulgar Keynesianism" is indeed a very good summary of my central complaint. although it arrives at it from a totally opposite direction. Sidenote: I miss when Paul Krugman was actually good :(.

http://web.mit.edu/krugman/www/vulgar.html

Ah. Thank you for teaching me that the Subjective Theory of Value is the specific cooky idea I find disastrously wrong in modern economics.

Why do I suspect that this reasoning is circular?

You would do well to check if your beliefs are falsifiable. If you believe that GDP objectively, universally, and precisely maps to real-world value, no thought experiment will disabuse you of that notion.

Are you claiming that "GDP objectively, universally, and precisely maps to real-world value" is non-falsifiable? Unless you reason is that nobody can even tell you what real-world value is, that proposition can easily be falsified.

If you ever observe a trade that changes GDP in a way that does not correspond to the change in real-world value, the belief is shown to be false. Of course that isn't just a thought experiment, you have to actually do the observations.

I said something slightly different; I claimed the commentor's beliefs were potentially unfalsifiable, and then I characterized those beliefs as "GDP mapping objectively etc to real world value"

I bet "GDP mapping objectively etc to real world value" is in fact falsifiable (and false, at least to some extent).

You would have to compare the GDP to some measure that better maps to real world value. And the choice of measure to use would be controversial. HDI? GNH? HPI? GNW?

I agree that GNP is a terrible measure of real world value (a lot of things increase GNP that decrease human well-being on a global or long-term scale). But if you want to falsify it, you need some other measure. Is the other measure falsifiable?

If this is one of his worst, I must read the rest. You are unsatisfied with an essay that is merely of correct reasoning, can you recommend some even better reading material than this?

Collection of Eliezer essays: https://www.lesserwrong.com/rationality

Yeah. Other LessWrong posts are usually great. I recommend almost everything at http://slatestarcodex.com, http://ribbonfarm.com, http://meaningness.com and other insight sinks.

"Eyeballing this question, my off-the-cuff answer—based mostly on the impressions related to me by every friend of mine who has ever dealt with medicine on a research level—is that I wouldn’t necessarily expect any medical researcher ever to have done a formal experiment on the first thought that popped into my mind for treating this extremely common depressive syndrome. Nor would I strongly expect the intervention, if initial tests found it to be effective, to have received enough attention that I could Google it.

"But this is just my personal take on the adequacy of 21st-century medical research."

I am guessing that the subtitle of his new book is, 300,000 words on why I have more smarts than everyone else?

> For instance, he seems axiomatically convinced that "creating trillions of dollars of value" is an objective good.

Is he?

The fact is, many people do behave as it is an objective good, if only to justify their own wealth. I think the arguments leans on that more that it leans on the belief that GDP growth is actually a good thing.

And as much as GDP growth sucks as a utility function, it does seem to be correlated with much goodnes. How strongly correlated exactly is the $3^^^3 question.

I think it's implicit that the 'trillions of dollars of value' are dollars of actual value, like purchases of goods and services that people actually want. Your parable is amusing, but not analogous at all. Of course if each economist extracted >$20,000 of sadistic pleasure out of the other eating poop, then, in fact, there was a net benefit to both of them. (But sadistic or capricious behavior throws a big wrench into economics, anyway...)

Why is that implicit? The economy seems full of huge transactions that have nothing to do with the quality of people's lives. I feel like this is... a pretty obvious position to arrive at if you start critically examining the types of transactions that take place, especially at the tops of finance. Some finance whiz is going to claim that derivatives, credit default swaps, and billions made from LIBOR rigging somehow benefits the common man. Many people actually believe this!

Don't forget to add "and created 2 jobs" to the punchline.

Reply on news.ycombinator.com