Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Twenty basic rules for intelligent money management

32 James_Miller 19 March 2015 05:57PM

1.  Start investing early in life.

The power of compound interest means you will have much more money at retirement if you start investing early in your career.  For example, imagine that at age eighteen you invest $1,000 and earn an 8% return per year.  At age seventy you will have $54,706.  In contrast, if you make the same investment at age fifty you will have a paltry $4,661 when you turn seventy.
Many people who haven't saved for retirement panic upon reaching middle age.  So if you are young don't think that saving today will help you only when you retire, but know that such savings will give you greater peace of mind when you turn forty.
When evaluating  potential marriage partners give bonus points to those who have a history of saving.  Do this not because you want to marry into wealth, but because you should want to marry someone who has discipline, intelligence and foresight.

continue reading »

How much does consumption affect production?

5 erratim 05 January 2015 03:51PM

A ewe for a ewe

In a discussion with Benquo over his recent suffering-per-calorie estimates I learned that there have been a few different proponents of incorporating short term elasticities into such estimates. But do empirical short term elasticities really improve our estimates of consumption's long term effect on production? For example, if I decide to reduce my lifetime consumption of chicken by one, should I expect the long term production of chicken to drop by ~1, ~0, or something in between?

I believe we should have a relatively strong prior that long term production has a  roughly 1:1 relationship with consumption, including for small individual decisions. Below are a couple arguments I find compelling, and a major exception that is not a short term elasticity.

Black box economies in general

If I go to a large alien civilization of uncertain economic structure and surprise them by buying(?) one widget, how should I expect that to affect their long term production of widgets? Seems like I should expect it to increase by one, because now they have one less than they used to. If it was originally decided that that widget should be produced; why wouldn't they decide to replace it when lost?

Neoclassical capitalism in the long term

In a simplified market, I expect there to be a lowest price at which chickens can be reliably produced at scale ("the Cost"). If producers expect the market price to be less than the Cost in the future, they will shut down production to avoid losses. If they expect it to be more than the Cost in the future, they might expand operations to make more profit. In the long term (when we can ignore temporary shocks to the system and producers have time to make adjustments), I expect the equilibrium price of chicken to approach the Cost of chicken (because other prices lead to conditions that push the price back toward the Cost). In other words, my prior is that the "price elasticity of supply" in the arbitrarily long term becomes arbitrarily high (imagine a virtually horizontal supply curve).

How many chickens will be produced at that long term price? However many are worth the Cost to consumers. If 50% of chicken consumers permanently become vegetarians, I expect that eventually the chicken industry will end up producing about 50% fewer chickens at a price similar to today's.

Similarly if consumption is reduced by just one chicken. My prior is that producers have an unbiased estimate of consumption, and that doesn't change when I eat one less chicken (so my best guess about their long term estimate of consumption drops by one when I forgo one chicken).

Time breaks the elastic limit

Compare my prior that every chicken forgone causes (in the long term) one less chicken to be produced, to the estimates that it only causes 6% or 76% of a chicken to not be produced (as Peter Hurford points out in the second case, the enormous range in these estimates alone is enough to raise flags).

Those numbers sound plausible in the short term when there's a backup in the chicken pipeline and a drop in price because producers were caught off guard by the drop in consumption. But if the vegetarians hold their new diets, won't the producers eventually react to the changed market? When they do I bet the equilibrium price will be somewhere close to the original Cost, and the quantity produced will be about 50% less (not 3% less or even 38% less). I think the thing these elasticity estimates are forgetting is that the producers aren't satisfied (in the long term) with the lower price that results from a chicken glut caused by vegetarianism. If they were, they'd be producing more chickens now.

Said another way, it all comes down to the difference between producers' reaction in the short term vs. the long term. In the short term, when someone decides not to eat a chicken, it goes to the next highest bidder (so price drops and production doesn't change much). But in the long term, producers produce all the chickens that will be demanded at the Cost (they want to produce as many as they can at that price, but if they produce any more, the chickens will be sold at a loss). When one person permanently becomes vegetarian, we should expect that long term size of the industry decreases accordingly.

When the long term Cost changes with industry size

To be clear, if we could actually measure consumption's effect on long term production in specific cases, it would rarely be exactly 1:1, though my prior is that it will average out to that over time. The exception is if consumption consistently affects the long term price in a particular direction. For example, here are some reasons that I might expect the Cost of chicken to grow or shrink as the size of the chicken industry increases:


  • Finite inputs such as limited agricultural land (Cost grows with size)
  • The production process also creates another product like eggs (Cost grows with size if marginal production is used for both)
  • Gains to scale such as factory farming (Cost shrinks with size)
  • R&D or innovation (Cost shrinks with size)
  • Favorable government policies (Cost shrinks with size)


If we have sufficiently certain estimates on any of these effects, we can certainly try to model them, although it would be a very different exercise than using empirical estimates of short-term elasticities. As it is, I have no idea which of the above effects win out (ie, whether the "consumption elasticity of the Cost" is positive or negative in the long term).

I think we would make our estimates more simple and accurate by sticking with the prior that eating one less chicken causes about one less chicken to be produced in the long term.

The Superstar Effect

10 adamzerner 03 January 2015 06:11AM

Modern microconomist Alfred Marshall explains that technology has greatly extended the power and reach of the planet's most gifted performers....He referenced a classical of the British opera singer Elizabeth Billington. She was a well-acclaimed soprano with a strong voice, that, naturally did not have access to a microphone or amplifier in 1798, let alone to MTV, CDs, iTunes, and Pandora. She could only reach a small audience. This limited her ability to dominate the market in the way that artists to do today. Marshall wrote, “so long as the number of persons who can be reached by a human voice is strictly limited, it is not very likely that any singer will make an advance on the £10,000 said to have been earned in a season by Mrs. Billington at the beginning of the last century, nearly as great [an increase] as that which the business leaders of the present generation have made on those of the last.” 

- Wikipedia

Technology has made it easy for us to reach large audiences. And to do so at no marginal cost. If a musician writes a song and puts it on iTunes, it doesn't cost him any money for one more person to download it.

The fact that technology has made it easy for us to reach large audiences has implications on the consumer side of things as well. As a consumer, I can go on iTunes and choose the best music to buy. To understand my point, consider a different world. In this world iTunes doesn't exist. In this world the best music is 200 miles away, but mediocre music is only 5 miles away. Because traveling 200 miles is inconvenient, I choose the mediocre music.

In today's world of iTunes, this doesn't happen. Technology exists that allows us to reach large audiences and to do so at little/no marginal cost. And so, the consumer can (and will) choose the best the market has to offer.

Now for the implications on the supply side. We've already seen that consumers can and will choose the best the market has to offer. "The best the market has to offer" is usually provided by a small number of talented people. Think about it: the best artists, performers, writers, athletes etc. These talented people end up serving a large proportion of the market, and are paid accordingly. This... is The Superstar Effect.

Because of these joint consumption economies, there is a unique opportunity to create and capture value. If you are the best, you capture insane amounts of value. Thus, there is a huge incentive to be the best.

So, should you invest in an attempt to outdo The Superstar and capture this value? Well, investment decisions are all about expected value. Balancing risk with reward. In this case, the potential reward is huge. Astronomical. These joint consumption economies allow you to reach tremendous markets. However, the question is "how big is the risk?".

Outdoing The Superstar is a large and complex task, and I won't pretend to have all the answers. However, I've had this nagging suspicion in the back of my mind for years. My suspicion is that people drastically overestimate this risk, and that with a good plan and enough resources, you could have an excellent chance at "outdoing The Superstar".

Before moving on, let me go through the logic one more time:

  • Today's joint consumption economies allow for firms to reach large amounts of people with little/no marginal costs.
  • Thus, consumers have tons of options to choose from.
  • Consumers often have similar enough tastes such that a large percentage of them end up choosing The Superstar.
  • In serving all of these people, The Superstar has created and captured a ton of value.
  • If another firm came along and outdid The Superstar, this new firm would replace The Superstar. It would now be the one to serve the large market, and would be compensated accordingly. There is a large reward for outdoing The Superstar.
  • Investment is all about balancing risk and reward. Investing in an attempt to outdo The Superstar has a very large potential reward. The question is, "what's the risk?".

Outdoing The Superstar

People seem to view large ventures like starting startups as a roll of the dice. They say things like, "9 out of 10 startups fail". I don't see things that way. I don't see it as "a roll of the dice". I see it as a deterministic puzzle that can be solved.

I should qualify that previous statement. I'm not trying to make a philosophical point, just a practical one. People seem to be afraid of what I'll call, Large Puzzles. Because of their size and complexity, people seem to be put off by them, and they fall back on outside view arguments like "9 in 10 startups fail".

I'll admit that Large Puzzles are complex, but I maintain that with enough resources and with a good plan, a lot of them are very solvable. I sense that a lot of these large joint consumption winner-take-all industries are ripe for the taking, and that with enough resources and a good plan, they can be taken.

My confidence isn't that high though. I don't understand these Large Puzzles well enough to really say. What I'm referring to are "relatively strong suspicions", not "beliefs" (my thoughts are cloudy enough such that I'm having trouble being more precise than this, sorry).


This is a bit of an aside and a rant, but here we go. Investors currently seem to be heavily biased towards investing in businesses that can be built incrementally. They want to...

  • See some sort of promise/traction before investing at all (usually).
  • Invest 10s/100s of thousands of dollars in a seed round.
  • See some more traction before they invest a couple/10s of millions in a series A.
  • See some more traction before they invest 10s/100s in the next round.
  • etc. etc.
What about firms that are trying to replace The Superstar? Such a task usually requires very large amounts of upfront investment. Because of the winner-take-all nature of these industries, you usually need to exceed a certain threshold of "firepower" before you have a shot at showing some traction, let alone at replacing The Superstar.

However, the fact remains that investment decisions are all about expected value. Risk vs. reward. Risk isn't inherently bad, it just needs to be balanced by the reward. An in the case of superstar industries, the potential reward is huge.

In fact, the idea that the distribution of returns in an investment fund follows a power law seems to be well accepted. This means that it makes sense for an investor to seek huge exits. Replacing a Superstar seems like a great way to do that to me.

But in reality, it seems that investors don't actually understand the power law. It seems that they try desperately to "minimize risk", and look desperately for signs of traction, and end up investing mostly in companies that can be built incrementally. Unfortunate.


The Large Puzzle that I understand best is Education (which causes my System I to care disproportionately about it). I'll indulge myself and say it: the education system today is shit.

I think that Elon Musk said it well. He said (paraphrasing):

Consider The Dark Knight. It's awesome! It has all the best actors, directors, special effects etc. Now imagine if you took the same script and asked the local middle school to reproduce it. It'd suck. That's education.

I think that this division of resources is really the core of the problem. Things you could do once you pool resources:

  • Put a lot of effort towards making each lesson great (in dath ilan, "One hour of instruction on a widely-used subject got the same kind of attention that an hour of prime-time TV gets on Earth"). Figure out how to word things properly. What examples to use. What analogies to give. Make lessons visual, animated, interactive. Gamify them and make them fun (when appropriate). Make them beautiful. Apply design thinking. Make them skimmable so students can refer back to them when they're studying. Include convenient references to things the student might have a question on.
  • Break lessons into chunks and organize them according to their dependencies (this is an important and difficult task). I'm a big believer that knowledge is hierarchical. That concepts have dependencies (to know A, you have to know B). I think it makes a lot of sense to have students learn things that they have the proper foundation for. I think this makes more sense in the negative: you shouldn't have students learn things that they don't have the proper foundation for. (This is a bit of an aside, but I think that mastery should be fixed, and time should be variable. Currently it's the opposite.)
  • Open up time for teachers to spend personal attention on their students. In today's system, they're usually too busy to do this. (Note: even with these great lessons, I still think that teachers will be useful. The lessons could be pretty good, so I'm not sure if they'd be necessary, but I suspect that they'd still be useful. I think using a human will still be the best way to diagnose and address the holes in a student's understanding.)
  • Come up with great practice problems, exercises, projects etc. 
  • Make tests way more accurate and effective. Make them smaller. And for gods sake, have them created by a separate financial entity than the entity that does the teaching!
  • This applies to a lot of what I said above, but iterate, iterate, iterate!! See what works and what doesn't work and change. Given the amount of "experimental subjects (students)" and "technicians (teachers)", there's a tremendous opportunity to do this. Effective collaboration and coordination might be tough, but I sense that it's doable.
Sorry, I may have mixed in a few opinions that aren't directly related to the idea of pooling resources and that should really be asides.

Anyway, I think that the Large Puzzle of Education is very solvable. I think that with enough resources, you could do a good enough job such that it becomes an industry where The Superstar Effect takes over. Where one Superstar addresses a large proportion of the market. And I think that this would have a huge and beneficial impact on the world.

Unemployment explanations

28 Stuart_Armstrong 07 November 2014 05:12PM

When I knew nothing of economics, unemployment wasn't mysterious. People wanted a job, and couldn't get one - well, people often want stuff they can't get. Nothing strange there, just one of those things.

Then I learnt some simple economics, and it became more mysterious. The employment market is a market, with the salary being the price. Why doesn't this market clear? Why doesn't the price (salary) simply adjust, and then everyone gets a job? It seemed profoundly mysterious that this didn't happen.

I've been gradually introducing myself to more economics (mostly indirectly) and I've encountered a lot of explanations for this perpetual market failure. Thus the mystery of unemployment is, if not resolved, at least somewhat explained. Since I would really have enjoyed reading a collection of unemployment explanations when I was initially puzzled (almost any explanation of unemployment you read in the press is worthless) I thought I'd do this for others. So here is my (entirely personal and idiosyncratic) summary of the main explanations I've encountered.


continue reading »

Three Parables of Microeconomics

25 jimrandomh 09 May 2014 06:18PM

(Epistemic status: Satire.)


First Parable: Equilibrium Pricing

Highway Offramp 72 leads to the isolated town of Townton. Visitors are greeted by two fuel stations, Carbonaceous Fossils (CF) and Hydrogenated Chains (HC), on opposite sides of the main road. There are no other gas stations for many miles. Together, these two stations sell 1000 gallons per day. Since their products are indistinguishable, and they have prominently posted prices, every driver will choose the cheaper one; or if the prices are the same, they will split half and half.  Both pay $1.50/gal for their stock and charge $2/gal to drivers, so half the drivers stop at each.

The owner of CF reasons as follows: If I keep my current price of $2, I will make 500*(2-1.5)=$250 of profit. But if I lower my price to $1.99, I will get twice as much business and make 1000*(1.99-1.5)=$490 of profit. The next morning, he updates his price.

Across the street, the owner of HC (who is having a bad day, due to the complete lack of customers), reasons the same way. The next morning HC has updated its price to $1.98; the morning after that CF lowers its price to $1.97; and so on.

Because CF and HC's owners are law-abiding model citizens, they never talk to each other about prices. That would be collusion, which is illegal. Later that month, with CF's price down to $1.52 and HC's price at $1.51, the local community center holds Game Theory night, where both owners attend a local economist's presentation on the Iterated Prisoner's Dilemma.

The next morning, both stations charge $1.52. The morning after that, $1.53. The morning after that, $1.54, and so on. Later that year, CF reasons as follows: If I keep my current price of $20...

(Moral: Gas station attendants should study game theory.)


Second Parable: Comparative Advantage

Two farmers, Alex and Bertha, grow potatoes and carrots. In one year, Alex can either grow 4 barrels of potatoes or 10 barrels of carrots, or some linear combination of the two, such as 2 barrels of potatoes and 5 barrels of carrots. Bertha is better at farming, and can produce 15 barrels of potatoes or 20 barrels of carrots, or some combination of the two. Doctors agree that everyone should eat exactly equal numbers of potatoes and carrots - an excess of one over the other would be unacceptable. So in the first year, having just settled a new frontier and not having met their neighbors, Alex plants 2.9 barrels' worth of each, and Bertha plants 8.6 barrels of each.

continue reading »

The Robots, AI, and Unemployment Anti-FAQ

47 Eliezer_Yudkowsky 25 July 2013 06:46PM

Q.  Are the current high levels of unemployment being caused by advances in Artificial Intelligence automating away human jobs?

A.  Conventional economic theory says this shouldn't happen.  Suppose it costs 2 units of labor to produce a hot dog and 1 unit of labor to produce a bun, and that 30 units of labor are producing 10 hot dogs in 10 buns.  If automation makes it possible to produce a hot dog using 1 unit of labor instead, conventional economics says that some people should shift from making hot dogs to buns, and the new equilibrium should be 15 hot dogs in 15 buns.  On standard economic theory, improved productivity - including from automating away some jobs - should produce increased standards of living, not long-term unemployment.

Q.  Sounds like a lovely theory.  As the proverb goes, the tragedy of science is a beautiful theory slain by an ugly fact.  Experiment trumps theory and in reality, unemployment is rising.

A.  Sure.  Except that the happy equilibrium with 15 hot dogs in buns, is exactly what happened over the last four centuries where we went from 95% of the population being farmers to 2% of the population being farmers (in agriculturally self-sufficient developed countries).  We don't live in a world where 93% of the people are unemployed because 93% of the jobs went away.  The first thought of automation removing a job, and thus the economy having one fewer job, has not been the way the world has worked since the Industrial Revolution.  The parable of the hot dog in the bun is how economies really, actually worked in real life for centuries.  Automation followed by re-employment went on for literally centuries in exactly the way that the standard lovely economic model said it should.  The idea that there's a limited amount of work which is destroyed by automation is known in economics as the "lump of labour fallacy".

Q.  But now people aren't being reemployed.  The jobs that went away in the Great Recession aren't coming back, even as the stock market and corporate profits rise again.

A.  Yes.  And that's a new problem.  We didn't get that when the Model T automobile mechanized the entire horse-and-buggy industry out of existence.  The difficulty with supposing that automation is producing unemployment is that automation isn't new, so how can you use it to explain this new phenomenon of increasing long-term unemployment?

Baxter robot

continue reading »

A New Interpretation of the Marshmallow Test

73 elharo 05 July 2013 12:22PM

I've begun to notice a pattern with experiments in behavioral economics. An experiment produces a result that's counter-intuitive and surprising, and demonstrates that people don't behave as rationally as expected. Then, as time passes, other researchers contrive different versions of the experiment that show the experiment may not have been about what we thought it was about in the first place. For example, in the dictator game, Jeffrey Winking and Nicholas Mizer changed the experiment so that the participants didn't know each other and the subjects didn't know they were in an experiment. With this simple adjustment that made the conditions of the game more realistic, the "dictators" switched from giving away a large portion of their unearned gains to giving away nothing. Now it's happened to the marshmallow test.

In the original Stanford marshmallow experiment, children were given one marshmallow. They could eat the marshmallow right away; or, if they waited fifteen minutes for the experimenter to return without eating the marshmallow, they'd get a second marshmallow. Even more interestingly, in follow-up studies two decades later, the children who waited longer for the second marshmallow, i.e. showed delayed gratification, had higher SAT scores, school performance, and even improved Body Mass Index. This is normally interpreted as indicating the importance of self-control and delayed gratification for life success.

Not so fast.

In a new variant of the experiment entitled (I kid you not) "Rational snacking", Celeste Kidd, Holly Palmeri, and Richard N. Aslin from the University of Rochester gave the children a similar test with an interesting twist.

They assigned 28 children to two groups asked to perform art projects. Children in the first group each received half a container of used crayons, and were told that if they could wait, the researcher would bring them more and better art supplies. However, after two and a half minutes, the adult returned and told the child they had made a mistake, and there were no more art supplies so they'd have to use the original crayons.

In part 2, the adult gave the child a single sticker and told the child that if they waited, the adult would bring them more stickers to use. Again the adult reneged.

Children in the second group went through the same routine except this time the adult fulfilled their promises, bringing the children more and better art supplies and several large stickers.

After these two events, the experimenters repeated the classic marshmallow test with both groups. The results demonstrated children were a lot more rational than we might have thought. Of the 14 children in group 1, who had been shown that the experimenters were unreliable adults, 13 of them ate the first marshmallow. 8 of the 14 children in the reliable adult group, waited out the fifteen minutes. On average children in unreliable group 1 waited only 3 minutes, and those in reliable group 2 waited 12 minutes.

So maybe what the longitudinal studies show is that children who come from an environment where they have learned to be more trusting have better life outcomes. I make absolutely no claims as to which direction the arrow of causality may run, or whether it's pure correlation with other factors. For instance, maybe breastfeeding increases both trust and academic performance. But any way you interpret these results, the case for the importance and even the existence of innate self-control is looking a lot weaker.

New report: Intelligence Explosion Microeconomics

45 Eliezer_Yudkowsky 29 April 2013 11:14PM

SummaryIntelligence Explosion Microeconomics (pdf) is 40,000 words taking some initial steps toward tackling the key quantitative issue in the intelligence explosion, "reinvestable returns on cognitive investments": what kind of returns can you get from an investment in cognition, can you reinvest it to make yourself even smarter, and does this process die out or blow up? This can be thought of as the compact and hopefully more coherent successor to the AI Foom Debate of a few years back.

(Sample idea you haven't heard before:  The increase in hominid brain size over evolutionary time should be interpreted as evidence about increasing marginal fitness returns on brain size, presumably due to improved brain wiring algorithms; not as direct evidence about an intelligence scaling factor from brain size.)

I hope that the open problems posed therein inspire further work by economists or economically literate modelers, interested specifically in the intelligence explosion qua cognitive intelligence rather than non-cognitive 'technological acceleration'.  MIRI has an intended-to-be-small-and-technical mailing list for such discussion.  In case it's not clear from context, I (Yudkowsky) am the author of the paper.


I. J. Good's thesis of the 'intelligence explosion' is that a sufficiently advanced machine intelligence could build a smarter version of itself, which could in turn build an even smarter version of itself, and that this process could continue enough to vastly exceed human intelligence.  As Sandberg (2010) correctly notes, there are several attempts to lay down return-on-investment formulas intended to represent sharp speedups in economic or technological growth, but very little attempt has been made to deal formally with I. J. Good's intelligence explosion thesis as such.

I identify the key issue as returns on cognitive reinvestment - the ability to invest more computing power, faster computers, or improved cognitive algorithms to yield cognitive labor which produces larger brains, faster brains, or better mind designs.  There are many phenomena in the world which have been argued as evidentially relevant to this question, from the observed course of hominid evolution, to Moore's Law, to the competence over time of machine chess-playing systems, and many more.  I go into some depth on the sort of debates which then arise on how to interpret such evidence.  I propose that the next step forward in analyzing positions on the intelligence explosion would be to formalize return-on-investment curves, so that each stance can say formally which possible microfoundations they hold to be falsified by historical observations already made.  More generally, I pose multiple open questions of 'returns on cognitive reinvestment' or 'intelligence explosion microeconomics'.  Although such questions have received little attention thus far, they seem highly relevant to policy choices affecting the outcomes for Earth-originating intelligent life.

The dedicated mailing list will be small and restricted to technical discussants.

continue reading »

Is Sunk Cost Fallacy a Fallacy?

20 gwern 04 February 2012 04:33AM

I just finished the first draft of my essay, "Are Sunk Costs Fallacies?"; there is still material I need to go through, but the bulk of the material is now there. The formatting is too gnarly to post here, so I ask everyone's forgiveness in clicking through.

To summarize:

  1. sunk costs are probably issues in big organizations
    • but maybe not ones that can be helped
  2. sunk costs are not issues in animals
  3. they appear to be in children & adults
    • but many apparent problems can be explained as part of a learning strategy
  4. there are few clear indications sunk costs are genuine problems
  5. much of what we call 'sunk cost' looks like simple carelessness & thoughtlessness

(If any of that seems unlikely or absurd to you, click through. I've worked very hard to provide multiple citations where possible, and fulltext for practically everything.)

I started this a while ago; but Luke/SIAI paid for much of the work, and that motivation plus academic library access made this essay more comprehensive than it would have been and finished months in advance.


Prediction is hard, especially of medicine

47 gwern 23 December 2011 08:34PM

Summary: medical progress has been much slower than even recently predicted.

In the February and March 1988 issues of Cryonics, Mike Darwin (Wikipedia/LessWrong) and Steve Harris published a two-part article “The Future of Medicine” attempting to forecast the medical state of the art for 2008. Darwin has republished it on the New_Cryonet email list.

Darwin is a pretty savvy forecaster (who you will remember correctly predicting in 1981 in “The High Cost of Cryonics”/part 2 ALCOR’s recent troubles with grandfathering), so given my standing interests in tracking predictions, I read it with great interest; but they still blew most of them, and not the ones we would prefer them to’ve.

The full essay is ~10k words, so I will excerpt roughly half of it below; feel free to skip to the reactions section and other links.

continue reading »

View more: Next