All posts by Peter

Book review: Capital in the Twenty-First Century, by Thomas Piketty.

Capital in the Twenty-First Century is decent at history, mediocre at economics, unimpressive at forecasting, and gives policy advice that is thoughtfully adapted to his somewhat controversial goals. His goals involve a rather different set of priorities than I endorse, but the book mostly doesn’t try to persuade us to adopt his goals, so I won’t say much here about why I have different priorities.

That qualifies as a good deal less dumbed-down-for-popularity than I expected from a bestseller.

Even when he makes mistakes, he is often sufficiently thoughtful and clear to be somewhat entertaining.

Piketty provides a comprehensive view of changes in financial inequality since the start of the industrial revolution.

Piketty’s main story is that we’ve experienced moderately steady increases in inequality, as long as conditions remained somewhat normal. There was a big break in that trend associated with WW1, WW2, and to lesser extents the Great Depression and baby boom. Those equalizing forces (mainly decreases in wealth) seem unlikely to repeat. We’re back on a trend of increasing inequality, with no end in sight.

Continue Reading

Most Universal Basic Income (UBI) proposals look a bit implausible, because they want to solve poverty overnight, and rely on questionable hopes for how much taxpayers can be persuaded to support[1].

They also fall short of inspiring my idealistic motives, because they want to solve poverty only within the countries that implement the UBI (i.e. they should be called national basic income proposals). That means even those of us living in relatively successful countries would be gambling on the continued success of the country they happen to live in. I imagine some large upheavals in the next century or so that will create a good deal of uncertainty as to which countries prosper.

Political movements to create national basic income run the risk of being hijacked by political forces that are more short-sighted and less altruistic.

Whereas I’m more interested in preparing for the more distant risks of a large-scale technological unemployment that might accompany a large increase in economic growth.

UBI without taxation?

Manna is a somewhat better attempt. It’s a cryptocurrency with a one account per human rule, and regular distributions of additional (newly created) currency to each account.

It provides incentives to sign up (speaking of which, I get rewards if you sign up via this link). It’s less clear what incentive people have to hold onto their manna[2].

It’s designed so that, given optimistic assumptions, the price of manna will be stable, or maybe increase somewhat. Note that those optimistic assumptions include a significant amount of altruism on the part of many people.

Cryptocurrencies gained popularity in part because they offered a means of trust that was independent of their creator’s trustworthiness.

Manna doesn’t attempt to fully replicate that feature, because they’re not about to fully automate the one-human-one-account rule. They’ve outsourced a good deal of the verification to cell phone companies, but the system will still be vulnerable to fraud unless a good deal of human labor goes into limiting people to one account each.

The obvious outcome is that people stop buying manna, so it becomes worth too little for people to bother signing up.

I suspect most buying so far has been from people who think any cryptocurrency will go up. That’s typical of a bubble.

That may have helped to jumpstart the system, but I’m concerned that it may distract the founders from looking for a long-term solution.

Why use a cryptocurrency?

Some of what’s happening is that crypto enthusiasts expect crypto to solve all problems, and apply crypto to everything without looking for evidence that crypto is helpful to the problem at hand. The cryptocurrency bubble misled some people into thinking that cryptocurrencies created free lunches[3] (manna comes from heaven, right?), and a UBI is a good use for a free lunch.

I recommend instead that you think of manna as primarily a charity, which happens to get some advantage from using a cryptocurrency.

Cryptocurrencies provide fairly cheap ways of transmitting value.

The open source nature of the mechanism makes it relatively easy to verify most aspects of the system.

These may not sound like terribly strong reasons, but it looks to me like much of the difficulty in getting widespread adoption of valuable new charities is that donors won’t devote much effort to evaluating charities. So only the most easily verified charities succeed on their merits, and the rest succeed or fail mainly on their marketing ability.


It seems almost possible that the price of manna could be stable or rise reliably enough to act as a good store of value.

But it won’t happen via the thoughtless greed that drove last year’s cryptocurrency buying frenzy. It requires something along the lines of altruism and/or signaling.

It seems to require the “central bank” to use charitable donations to buy manna when the price of manna declines.

It also requires something unusual about the average person’s attitude toward manna. Would it be enough for people and businesses to accept manna as payment, for reasons that involve status signaling? That doesn’t seem quite enough.

It’s also important to persuade some people to hold the manna for a significant time.


There’s little chance that can be accomplished by making manna look as safe as dollars or yuan. The only possibility that I can imagine working is if holdings of manna provide a good signal of wealth and wealth-related status. Manna seems to be positioned so that it could become a substitute for a fancy car or house as a signal of wealth. With that level of acceptance, it might provide a substitute for bank accounts as a store of value.

Signaling motives might also lead some upper-class people/businesses to use it as medium of exchange.

To work well, manna would probably need to be recognized as a charity, with a reputation that is almost as widely respected as the Red Cross. I.e. it would need to be a fairly standard form of altruism.

The main UBI movement wants to imagine they can solve poverty with one legislative act. Manna uses a more incremental approach, which provides less hope of solving poverty this decade, but maybe a bit more hope of mitigating larger problems from technological unemployment several decades from now.


Manna seems to be run by the first group of people who decided the idea was worth doing. Typically with a new technology, the people who will manage it most responsibly wait a few years before getting involved, so my priors are that I should hesitate before deciding this particular group is good enough.

Manna currently isn’t fair to people who can’t afford a cell phone, but if other aspects of manna succeed, it’s likely that cell phone companies will find a way to get cell phones to essentially everyone, since the manna will pay for the phones. Also, alternatives to cell phones will probably be implemented for manna access.

The high-level rhetoric says any human being is eligible for manna, but a closer look shows that anyone under 18 is treated as only partly qualified – manna accumulates in their name, and they get access to the manna when they come of age. The arbitrariness of this threshold is unsettling. We’ll get situations where people become parents, yet don’t have access to manna. Or maybe that’s not much of a problem because someone else will enable children to borrow, using their manna as collateral?

The problems will become harder if someone needs to figure out what qualifies a human being in an Age of Em, where uploaded minds (human, and maybe bonobo) can be quickly duplicated.

I’m not too clear on how the governing board will be chosen – they say something about voting, which sort of suggests a global democracy. That runs some risk of short-sighted people voting themselves more money now at the cost of a less stable system later. But the alternative governing mechanisms aren’t obviously great either.

I’d have more confidence if manna were focused exclusively on a UBI. But they want to also enable targeted donations, by providing verified age, gender, location, and occupation data, and “verified needy” status indications generated by other charities. Maybe a one or two of those would work out well, but I see some important tension between them and the “NO DISCRIMINATION” slogan on the home page.

The people in charge also want to solve “instability … resulting from too much money being held in too few hands and used for reckless financial speculation” without convincing me they understand what causes instability.

I’d be concerned about macroeconomic risks in the unlikely event that manna’s use became widespread enough that wages were denominated in it. Manna’s creators express Keynesian concerns about aggregate demand, suggesting that the best we could hope for from a manna monetary policy is that it would repeat the Fed’s occasional large mistakes. I’d prefer to aim for something better than that.

Current central banks have enough problems with promoting monetary stability. If they’re replaced by an organization which has a goal that’s more distinct from monetary stability, I expect monetary stability to suffer. I don’t consider it likely that manna will replace existing currencies enough for that to be a big concern, but I find this scenario hard to analyze.

Like most charities, it depends more on support from the wealthy than from the average person. Yet the rhetoric behind Manna seems designed to alienate the wealthy.

Is current People’s Currency Foundation sufficiently trustworthy? Or should someone create a better version?

I don’t know, and I don’t expect to do enough research to figure it out. Maybe OpenPhil can investigate enough?

Is this Effective Altruism?

The near-term benefits of Manna or something similar appear unimpressive compared to GiveDirectly, which targets beneficiaries in a more sophisticated (but less transparent?) way.

But Manna’s simpler criteria make it a bit more scaleable, and make it somewhat easier to gain widespread trust.

The main costs that I foresee involve the attention that is needed to shift people’s from charities such as the Red Cross or their alma mater as the default charity, toward manna. Plus, of course, whatever is lost from the charities who get fewer donations. There’s no shortage of charities that produce less value than a well-run UBI would, but the social pressure that I’m imagining is too blunt an instrument to carefully target the least valuable charities as the things that manna should replace.


I don’t recommend significant purchases of manna or donations to the People’s Currency Foundation now. Current efforts in this area should focus more on evaluating these ideas further, figuring out whether a good enough implementation exists, and if it should be scaled up, then we should focus more on generating widespread agreement that this is a good charity, and not focus much on near-term funding.

I give Manna a 0.5% chance of success, and I see an additional 1% chance that something similar will succeed. By success, I mean reliably providing enough income within 30 years so that at least 10 million of the world’s poorest people can use it to buy 2000 calories per day of food. That probability seems a bit higher than the chance that political action will similarly help the world’s poorest.


[1] – e.g. pointing to tax rates that were tolerated for a while after a world war, without noticing the hints that war played an important role in getting that toleration, and without noting how tax rates affect tax avoidance. See Piketty’s Capital in the Twenty-First Century, figures 13.1 and 14.1, for evidence that tax rates which are higher than current rates haven’t generated more revenues.

[2]Wikipedia says of the original manna: ‘Stored manna “bred worms and stank”‘.

[3] – or maybe the best cryptocurrencies do create free lunches, but people see more free lunches than are actually created. The majority of cryptocurrencies have been just transfers of money from suckers to savvy traders.

Book review: The Book of Why, by Judea Pearl and Dana MacKenzie.

This book aims to turn the ideas from Pearl’s seminal Causality into something that’s readable by a fairly wide audience.

It is somewhat successful. Most of the book is pretty readable, but parts of it still read like they were written for mathematicians.

History of science

A fair amount of the book covers the era (most of the 20th century) when statisticians and scientists mostly rejected causality as an appropriate subject for science. They mostly observed correlations, and carefully repeated the mantra “correlation does not imply causation”.

Scientists kept wanting to at least hint at causal implications of their research, but statisticians rejected most attempts to make rigorous claims about causes.

Continue Reading

Book review: The Plant Paradox, by Steven R. Gundry.

This book describes a good diet, which Gundry seems to have put together by combining ideas from other competent nutritionists, testing many variations on himself and his patients, and keeping the components that showed the best results.

Unfortunately, his rhetoric is designed to convince us that he’s got brilliant, revolutionary theories. That rhetoric bears little resemblance to careful reasoning. It seems more likely that he tried a bunch of arguments on his patients, and kept the ones that more effectively scared them into following his diet.

His puns are mightier than his scalpel (“No More Mr. Knife Guy”).

I first noticed Gundry via the ApoE4 website, at about the time I gave up hoping that I could eat coconut fat without raising my cholesterol. I saw that Gundry recommended no coconut fat for my genotype, but liked coconut fat for others. That led me to think: here’s a paleo-friendly nutritionist who knows more than average.

Then I switched for 5 or 6 weeks to a diet more in line with his Apoe4 advice [1], and was surprised at how much more my cholesterol levels dropped than I expected. I only have modest evidence suggesting that that dietary change was the main reason for the cholesterol drop, but it’s still a bit of evidence that Gundry knows something valuable that I would have otherwise have overlooked.
Continue Reading

Book review: The Life You Can Save, by Peter Singer.

This book presents some unimpressive moral claims, and some more pragmatic social advocacy that is rather impressive.

The Problem

It is all too common to talk as if all human lives had equal value, yet act as if the value of distant strangers’ lives was a few hundred dollars.

Singer is effective at arguing against standard rationalizations for this discrepancy.

He provides an adequate summary of reasons to think most of us can easily save many lives.
Continue Reading

No, this isn’t about cutlery.

I’m proposing to fork science in the sense that Bitcoin was forked, into an adversarial science and a crowdsourced science.

As with Bitcoin, I have no expectation that the two branches will be equal.

These ideas could apply to most fields of science, but some fields need change more than others. P-values and p-hacking controversy are signs that a field needs change. Fields that don’t care much about p-values don’t need as much change, e.g. physics and computer science. I’ll focus mainly on medicine and psychology, and leave aside the harder-to-improve social sciences.

What do we mean by the word Science?

The term “science” has a range of meanings.

One extreme focuses on “perform experiments in order to test hypotheses”, as in The Scientist In The Crib. I’ll call this the personal knowledge version of science.

A different extreme includes formal institutions such as peer review, RCTs, etc. I’ll call this the authoritative knowledge version of science.

Both of these meanings of the word science are floating around, with little effort to distinguish them [1]. I suspect that promotes confusion about what standards to apply to scientific claims. And I’m concerned that people will use the high status of authoritative science to encourage us to ignore knowledge that doesn’t fit within its paradigm.

Continue Reading

Experts have been debating the causes of the industrial revolution for a long time, with few signs of agreement. I suggest that’s due to a human-centric bias which assumes that no other species could have caused human progress.

I became curious after reading that, while dogs were domesticated, cats intermixed with humans while showing few signs of domestication. Cats have cooperated with humans for long enough that I’d expect at least a little bit of co-evolution. If it wasn’t the humans selecting for the most docile or friendly cats, then maybe it was the cats selecting for the most docile or friendly humans.

One interesting hint is that various human cultures tell different stories about how many lives a cat has. As far as I can tell, the cultures that were most advanced in the 18th century (Britain, China) say cats have 9 lives, the more southern parts of Europe say 7, and Arab cultures say 6. That’s a fairly striking correlation between how highly cultures think of cats and how prospurrous they were near the start of the industrial revolution.[1]

People feed me, shelter me and love me ... I must be God

The hardest aspect of explaining the industrial revolution is explaining why the China’s level of development around 1800 didn’t enable it to lead the world. Why did Europe do better than China? Britain seems to have stopped eating cats sometime in the 18th century, while cat-eating has only started to become controversial in southern China this century. I’m less sure what the story is for northern China, where cat-eating apparently doesn’t happen today, but I haven’t found evidence about when cat-eating became unpopular there. Southern China is the region that’s generally said to have been advanced enough that it could have experienced an industrial revolution around 1800. Maybe northern China was backwards then for reasons unrelated to cats, maybe they were held back by their cat-eating southern compatriots, or maybe they adopted a cat-safe culture quite recently – could that have been the change that triggered China’s impressive growth over the past 40 years? Can anyone point me to better evidence on this topic?

Another possible reason for China’s lag is that a completely different species of cat (the leopard cat, Prionailurus bengalensis) acted as “pets” in China about 5000 years ago. Felis catus may have pawsed their human domestication plans in China due to distractions from their struggles with the leopard cat.

I can imagine many strategies that the cats could have used to purr-suade humans to change:

  • Cats probably have some influence over who the spread diseases to.
  • They could influence human mating choices, by causing distractions when “bad” humans court each other, versus purring peacefully when “good” humans court.[2]
  • Cat “ownership” can show trustworthiness, once cats establish conditions under which humans recognize cats as high status and/or recognize that cat-friendliness implies a person is less prone to pointless conflict. Cats exert some influence on who they associate with, and they can use that to increase the status of “good” humans, and/or increase the mating opportunities of “good” humans.
  • They could influence which areas have more or less rodents, thereby causing “bad” villages to have more food eaten by rodents than was he case with “good” villages.
  • Eating rodents protected nearby humans from diseases spread by rodents. This was especially important during the black plague.

What benefits would the cats have been selecting for?

I presume an important part of their plan would have been selecting for a wider moral circle, since that would make humans safer for cats to live near.

A wider moral circle is correlated with higher trust, and lower violence. These are likely important for cooperation among groups that are much larger than the Dunbar number. See Fukuyama for more on why that mattered.

So, regardless of whether cats planned to advance human civilization, or were merely protecting themselves, their interests in domesticating humans helped generate conditions that were conducive to an industrial revolution.

[1] – A more exotic version of this story is that we’re living in a simulation, and cats are the avatars of the beings who run the simulation. They’re arrogant enough to taunt us by reusing the same avatar just enough for humans to suspect something, but they stop before leaving enough evidence for anything to be proven.

[2] – I don’t want to express any opinion here about the nature-nurture debate, since there are many ways in which cats could have changed human behavior, and I have little hope of tracking down enough evidence to show which strategies the cats actually used. Feline influence on mating could achieve its results via genetic selection – that would require unusual patience, but produce the most stable results. Or the cats could have focused on making the good humans higher status, and the bad humans lower status – that could potentially produce faster results, but is more likely to depend on continuing feline manipulations of human culture.

See also the book A Farewell to Alms for a more detailed argument that British society has long been subjected to selection pressure which made it ripe for the industrial revolution. Alas, it neglects cats.

Scott Sumner asks whether those of us[1] who talked about a housing bubble are predicting another one now.

Sumner asks “Is it possible that the housing boom was not a bubble?”.

It’s certainly possible to define the word bubble so that it wasn’t. But I take the standard meaning of bubble in this context to mean something like a prediction that prices will be lower a few years after the time of the prediction.

Of course, most such claims aren’t worth the electrons they’re written on, for any market that’s moderately efficient. And we shouldn’t expect the news media to select for competent predictions.

Sumner’s use of the word “bubble” isn’t of much use to me as an investor. If prices look like a bubble for a decade after their peak, that’s a good reason to have sold at the peak, regardless of what happens a decade later.

If I understand Sumner’s definition correctly, he’d say that the 1929 stock market peak looked for 25 years like it might have been a bubble, then in the mid 1950s he would decide that it had been shown not to be a bubble. That seems a bit strange.

Even if I intended to hold an investment for decades, I’d care a fair amount about the option value of selling sooner.


The U.S. is not currently experiencing a housing bubble. I can imagine a small housing bubble developing in a year or two, but I’m reasonably confident that housing prices will be higher 18 months from now than they are today.

Several signs from 2005/2006 that I haven’t seen recently:

I mostly used to attribute the great recession to the foolish leverage of the banking system and homebuyers, who underestimated the risks of a significant decline in housing prices.

I’ve somewhat changed my mind after reading Sumner’s writings, and I now think the Fed had the power to prevent most of the decline in gdp, unless it was constrained by some unannounced limit on the size of its balance sheet. But I still think it’s worth asking why we needed unusual Fed actions. The fluctuations in leverage caused unusual changes in demand for money, and the Fed would have needed to cause unusual changes in the money supply to handle that well. So I think the housing bubble provides a good explanation for the timing of the recession, although that explanation is incomplete without some reference to the limits to either the Fed’s power or the Fed’s competence.

[1] – he’s mainly talking about pundits who blamed the great recession on the housing bubble. I don’t think I ever claimed there was a direct connection between them, but I did imply an indirect connection via banking system problems.

Book review: Surfing Uncertainty: Prediction, Action, and the Embodied Mind, by Andy Clark.

Surfing Uncertainty describes minds as hierarchies of prediction engines. Most behavior involves interactions between a stream of information that uses low-level sensory data to adjust higher level predictive models of the world, and another stream of data coming from high-level models that guides low-level sensory processes to better guess the most likely interpretations of ambiguous sensory evidence.

Clark calls this a predictive processing (PP) model; others refer to is as predictive coding.

The book is full of good ideas, presented in a style that sapped my curiosity.

Jeff Hawkins has a more eloquent book about PP (On Intelligence), which focuses on how PP might be used to create artificial intelligence. The underwhelming progress of the company Hawkins started to capitalize on these ideas suggests it wasn’t the breakthrough that AI researchers were groping for. In contrast, Clark focuses on how PP helps us understand existing minds.

The PP model clearly has some value. The book was a bit more thorough than I wanted at demonstrating that. Since I didn’t find that particularly new or surprising, I’ll focus most of this review on a few loose threads that the book left dangling. So don’t treat this as a summary of the book (see Slate Star Codex if you want that, or if my review is too cryptic to understand), but rather as an exploration of the questions that the book provoked me to think about.

Continue Reading