I agree that EA is worth rescuing from some of its high-profile advocates, but I'm still worried about being turned into a paperclip.
(Steven Pinker not being worried makes me worry more, because he's reliably wrong about everything outside his area of expertise. I think he and Richard Dawkins must have made some sort of pact.)
re: japan, i still can never really understand economic reports and analysis of japan’s economy. i’ve been visiting japan for the last 40 years and all i see is advancements everywhere and is still the most futuristic place til this day in addition to social stability, social welfare, and overall widespread wealth like ive never seen. you will never ever see distressed people walking around like zombies fiening for their vice of choice on the streets, it’s just not there. what am i missing?
I think the experience of being a tourist is radically different from the experience of citizens.
Roger Cohen once wrote an irritating column about how Italy is actually really nice despite all the economic problems people complain about, which didn't impress my Italian friends much. And my general experience is that living in developing countries is very pleasant as long as you're not on a local income.
As someone that lives in Tokyo, I feel the experience is quite nice?
On a good, but far from great, local salary, my partner and I are able to eat out all the time and live in an apartment that is impossibly nice by SF standards. While I was able to eat out all the time in SF, being mostly single and in tech, that just isn't possible in most/all of Western Europe.
Public infrastructure just works. Trains are generally on time and go everywhere I care about going, utilities are reliable, and street surfaces are well maintained. In SF, public transit was a joke, electricity and internet cut out many times a year, and on the off chance of rain, the uneven road surface would pool up water for cars to splash into anyone walking.
Safety and cleanliness is extremely high. It's kinda hard to relax in public in the US or even Europe. Maybe used needles aren't a problem with functional shoes. Maybe drug zombies are not terribly likely to actually attack you. Maybe the salaries pay enough to buy a new bike every few months. However, just not having to deal with all of that is incredibly nice.
Tourists see public spaces but rarely see private spaces. Japan has fantastic public spaces but pretty poor private spaces (the average dwelling is 93 square metres). The US is way on the other end of the spectrum as far as developed countries go; many public spaces (city streets, public transport etc) rather suck, but go inside people's houses where they actually live and they're the nicest in the world.
Drug zombies are a separate issue. The other weird thing about the US is that the population of complete bottom-crawlers incapable of life in a civilised society is about ten times larger than it is in most other Western countries. A complete failure to enforce drug laws doesn't help either.
> Though, being able to afford drugs in a way is a sign of wealth.
A pill of fentanyl on the streets of Seattle is $3-5 as per Seattle PD, and bulk purchases can go below $0.50 a pill.
That isn't really a sign of wealth amount of money for developed countries. Mass overdoses from cocaine might be a sign of wealth, but that isn't really what is happening.
I was mostly thinking of the time you'd need for your "hobby". Some people have a lot of time because they have money, some have time because they can't make any money.
I think a really big part of it is that Japan is optimized for selling cheap things. Right by the station in Azabujuban, which is a pricy business district, you can get a hot bento lunch with chicken and veggies and rice for $3, or fried chicken and a rice ball for $2. You could live half an hour from that station by train for $150/mo in 150 square feet. The $15 jeans from Uniqlo are nice enough that I wear them over stuff I brought from the US. And I'm in a Facebook group that's drowning in "$8/hr installing solar panels, no language skills required, housing and food provided" style job listings. You can buy lightweight cars that go 30mph for $5000.
This seems like a different thing than PPP - expensive stuff is still expensive. Cheap stuff is just really nice and easy to find.
Is Japan futuristic? I feel like the most recent technology improvement I saw was that a lot more restaurants started taking credit cards. It has great well-run infrastructure sure, but not brand new ideas.
I've also heard Korea is even newer.
Their wages are also kind of low, especially with the exchange rate at $1=150jpy. A lot of people are leaving for Australia which pays like twice as much.
It's less that Japan is futuristic, and that the West is incredibly behind on the adoption of many great technologies. Though to be fair, a lot of what makes Japan so futuristic also applies to Korea, often to a greater extent.
For example, Japan hit 90% adoption rate of heat pumps in the 2000s, and the west has lagged behind so far that it's common for expats to make fun of Japanese people running the air conditioning in reverse like cavemen who never discovered they could just burn dinosaur juice instead.
Or that self checkouts and fast food ordering kiosks are mature, reliable devices that are so common and easy to use, that even plenty of retired tech-phobic types use them without issue. It's just incredible how bad and unpopular self checkouts and fast food ordering kiosks in the US still are in comparison.
Or everything about toilets. Or faucets that allow you to set a desired temperature. Or hot water that comes on in seconds and never runs out. Or how common video chat doorbell systems are in apartments. Or motion sensor based lights that turn on when you walk in the door. Or how you never have to climb stairs into the train except on a handful of legacy historic tramways. Or ordering tablets and call buttons at restaurants. Or elevators where you can cancel mistake floor requests. The list really goes on and on and on.
The West is so far behind that many of these technologies have become so mature, people like you have trouble recognizing how far the West is behind on them. It feels like a conscious cultural choice that the self checkout at Walmart is a buggy piece of shit while the self checkout at Seiyu just works, not the US being so far behind on basic technology.
If anything, that's one of the big problems Japan has. Japan never managed to capitalize on its position of futuristic thought leader back when it had it, and has since lost it.
You'd think that the vast majority of American homes having hot water that takes ages to come on, delivered through faucets that required constant fidgeting to maintain a constant temperature, would be an opportunity to sell some futuristic Japanese technology that solves all of that. However that just doesn't happen.
How so? Tokyo is an amazing place, but so is Seoul. I find the latter's subway to be cleaner, newer, and easier to navigate. I'd give Seoul the edge in airports, too. Both cities are connected by HSR. I've enjoyed my time in both cities, and if you forced me to choose one to live in, I'd probably go with Tokyo mainly because its winters are quite a bit milder and feels less likely to be attacked (all those artillery tubes aimed at the South Korean capital!).
But I personally don't see an infrastructure advantage for Tokyo.
I always look forward to these "five things" posts, because it's rare that i don't find a few informative and thought provoking. This one was for me five for five, so I wanted to thank you for doing these posts.
One comment. I think wealth, rather than income, is a much more important indicator of inequality for the very rich. That's the measure that the 0.1% and above care about. And that's the measure used for comparison purposes among the very rich.
But that is tremendously more complicated, it involves valuing illiquid assets, privately held companies. Integrating debt etc. So if there's a wide range of empirical outcomes on income inequality, wealth numbers over time are gonna be close to worthless
I think it’s worth distinguishing between two arguments.
The good argument: the real value of illiquid assets has changed over time (e.g. investors become more comfortable with high P/E ratios when interest rates were zero). Therefore it’s hard to know whether the rich are actually richer than they used to be.
The bad argument: illiquid assets aren’t the same as liquid assets, therefore we can’t really measure them, and therefore billionaires aren’t actually stupendously rich. (I see this one float around Reddit a lot, usually as a defense of low tax rates).
It sounds like you’re making the good argument though.
Yeah, it's the former. I have worked in wealth management and I am intimately familiar with all these devils in the details, and how much accounting in illiquids is much more art than science lol.
It doesn't make me in favor of low taxes by any means, But it does lead me to want much higher estate taxes with much lower exemption thresholds but less interest in trying to work out annual wealth taxes. But I feel that way for plenty of reasons.
I agree that wealth is even more difficult to measure. But it's far "stickier" than income and it really is how the very rich measure themselves. So true inequality, at this time, may be impossible to measure.
Well stickier in some sense. But if 70% of your wealth is tied to your public company, and it makes more profit than last year, but because interest rates quadruple, the forward p/e decreases 60%, mark to market you are significantly less wealthy. However, on a discounted cash flow basis, maybe it's ultimately inconsequential long term, but the liquidation value has fallen dramatically. Given how much wealth in that tier is tied to equity ownership public or private, and how correlated that can be, it seems like a quagmire. I get that in some sense the 'economic and political power' from owning a firm doesn't really shift with those considerations, but that feels separate.
Yes, perhaps harder to measure - BUT that doesn't mean that the effect isn't real. Just that there isn't a sufficient incentive (i.e. someone willing to pay for the research) to do it. Hmm, wonder why that is.....
My sense is Deaton fell in love with a conceptualization of his own that he initially found clever, and wouldn't back off out of pride or stubbornness. It's an understandable and common human foible. But one that is nonetheless damaging.
Americans do tend to die younger than their counterparts in other rich countries. And as far as I know the gap has widened over the years (the USA's relative position would have looked a lot more favorable in, say, 1970). It's not unreasonable to posit that America's greater tolerance of economic inequality and less effective safety net are among the drivers of this phenomenon.
But as I understand it, Deaton tried to come up with a neat and tidy theory that connected the above with the rise of the MAGA movement. And this necessarily focuses on the white working class. And that's where the idea falls apart, because there's overwhelming evidence the particular spike among this cohort is associated with opioids. Needless to say there have been major differences in how America and the rest of the rich world have handled the emergence of the new generation of these drugs. In short, US practices were basically *designed* to get people addicted to drive pharm products. It's not despair that has caused so many working class white Americans to die prematurely; it's been the intersection of brain physiology and very poorly designed painkiller regulations that have done that.
It’s an American issue for sure but it’s basically guns, cars, obesity and opioids, not the lack of a safety net. For one thing, America’s safety net for people 65+ is as good as anywhere else.
You're conflating what Deaton was referring to—a sharp rise post 2000 (especially post 2010) in mortality among non college whites—with the overall phenomenon of lower US life expectancy (vis a vis other rich countries). I agree the factors you cite all play a role in the latter. As to the former—which Deaton erroneously pins on Dickensian steroids—it's opioids.
Obesity is a problem, but the "deaths of despair" narrative was specifically trying to explain a sharp rise in mortality among the white working class in the new millennium, and obesity rates have been rising for many decades. Also, Blacks suffer from obesity at higher rates than whites.
There are number of contributing reasons as to why Americans don't live as long on average as the denizens of other high income countries, for sure. But general, lower life expectancy in the US is *not* what Deaton's "deaths of despair" claims were referring to.
I'm having trouble in my head determining whether it's a fair comparison. Obviously, magicking away "cardio-metabolic deaths" (a huge cause of death, possibly the majority) is going to have a much larger effect than magicking away minor causes of death.
What I do find interesting is that the very recent decrease in US life expectancy does seem to be attributable entirely to drugs. Without drugs, life expectancy has been flat since 2015 as it has in most other countries.
Re the cardio metabolic line: these plots show differences, so I guess the differences will be due to eating and other behavioral choices. But I do wish I could see the article to know more details.
It’s on a spectrum right? Makes sense to mitigate it to the benefit of people who were disadvantaged in one or more of the myriad ways one can be. Plus the whole point of a robust economy is to benefit the humans operating within it. It’s supposed to be in service of us (all of us) rather than vice versa.
The American Left is really bad at touting its accomplishments. So the narrative becomes "we've made no progress" instead of "we've made progress but will have to keep working as conditions evolve."
Yes exactly. Income inequality is better than at basically any time in civilized human history, the US had a very large and growing upper middle class, and poor people in the US are more likely to suffer from obesity than hunger. Let’s focus on real problems please
Yes, Income inequality has been improving continuously throughout human history. That is, up until the last 50 years where it has been getting worse. This is a problem and we absolutely will focus on it.
Well, that was true well into the 80s, but fairly clear that it has turned since then - all the studies that Noah cited (well, except for the most extreme) show a growing income inequality (and I'd argue much more growing wealth inequality). And I sure know a lot of poor and homeless folks who don't suffer from obesity (especially the children).
For a sizable time after WWII, we were moving towards a diamond shaped economy (and Europe as well), but at least in the US, that has been shifting ("hollowing out of the middle class") - hidden for a while as women (and hence 2 earner families) joined the workforce.
The ideal amount of inequality in society is probably not zero, but it probably is lower than any existing national economy so it seems like a somewhat misplaced concern to worry about an economy becoming too equal.
The ideal amount of income inequality is probably equal to the actual amount of productivity inequality. If some people are one hundred times more productive than others then they should be one hundred times richer.
The ideal amount of _productivity_ inequality is probably lower than it is, and this decrease in inequality should come entirely from improved productivity at the low end.
Yes, maybe, but I can tell you just from software engineering, that income is rarely related to productivity in the real world. And I doubt that teachers, etc. are not productive compared to the gross low income
An issue is that some people are valued on their past work and some people are valued on their expected future work. Generally you become switch by switching from the first to the second.
eg becoming a hit artist and selling off shares in your future income.
I believe that a society where pay reflected productivity would actually be more unequal than the status quo. People just aren't equally productive! Even within the same job, some of my colleagues are better at the job than me, some are worse, but everyone working under the same job title earns the same.
First one should consider the political bent of the authors. One set appears to be in the pro-socialism/anti-capitalistic realm that colors their assessments and decisions. The other is more grounded in the capitalistic system that has overwhelmingly benefited humanity.
Another problem is the value-liquidity issue. At one time not so long ago there were more millionaires per capita in Idaho than any other state, yet the standard of living was one of the lowest. The reason was land ownership. Large acreages equaled low liquidity, and many of those same millionaires were barely economically keeping their heads above water. So how does this fit the inequity model?
Income inequality is indeed necessary for a strong, innovative economy. The question is *how much* is optimal, especially after accounting for taxes and transfers.
Deaton did point to one feature of as-existing "Capitalism" that contributes to stress/despair: employer "provided" health insurance. Although paid for by employees' lower wages, for any specific employee, it is a cost to the employer, a fixed cost per person. This makes it much higher percentage of wage for the minimum wage employee that the CEO. This inclines employers to substitute capital and higher wage labor for low wage labor and to strategies of part-time, contractor status employment to avoid that cost. It was a mistake of ACA not to have structured the program so that employers could "dump" low-income employees on to a more generous ACA.
The ACA was about as ambitious as it could have been given the political constraints. A public option does seem like a reasonable next step (and a mechanism to uncouple insurance from employment) but it's not happening without large Democratic majorities or Republicans supporting it (neither is happening anytime soon).
There is certainly merit in the government making the lives of citizens easier and more predictable like simplifying tax prep and making enrollment in government programs automatic and seamless. It is stressful to worry that a paperwork error can drastically affect your life! It's odd, but a lot of Americans don't seem to like the government to function TOO well.
That's not a feature of capitalism, it's a consequence of the US tax code (i.e. government, not capitalism). Most countries don't do this, employees pay their own health insurance.
Isn't one issue that health _care_ is much more expensive in the US than in most European countries, before we even get to the question of health _insurance_?
Actually, the US pays nearly 20% of GDP for Healthcare, while every modern free market democracy pays 10% of GDP. That is an annual GDP waste of $2.4 Trillion every year.
Health Insurance is the biggest scam ever perpetrated on the American public.
a. Health is NOT a product. Sons, daughters, wives and husbands are fungible like Ford vs Chevy vs Dodge trucks.
b. Health Insurance NEVER made one person healthier, but has made health outcomes worse as we know that Americans are living far shorter lives that people under Ntional Healthcare.
c. All we need is uncapped mandoratory Medicare payments on all income by every person.
From the Economist dated Oct 26th 2023. Headline: "How health-care costs stopped rising"
[Note -- the Economist uses 'America/American' to refer to the United States.]
Text: "For a long time, health care was eating the world. From 1950 to 2009 American spending on hospitals, medics and the like rose from 5% of gdp to 17%. Between the late 1970s and the mid-2010s British public spending on health rose by 4% a year in real terms, much faster than the economy’s growth of 2% a year. From 1980 to 2010 overall French prices rose by 150%; the price of caring for a sick or old person rose by 250%. Among economists, the proposition “health care’s share of gdp rises” was almost as close to an iron law as “free trade is good” or “rent controls do not work”.
"The iron has now melted. Even as populations age, and as the world continues to deal with the fallout from the covid-19 pandemic, health care is no longer taking over the economy. Across the rich world health care’s share of gdp jumped in 2020 and 2021, because of a combination of pandemic-related spending and lower gdp. However, it has since fallen back to close to its level in 2008 (see chart 1). Because of this “flattening of the curve”, health spending today is somewhere in the region of $2trn below its pre-2009 trend.
"In some countries the changes are still more dramatic. The ratio of health-care spending to gdp has fallen from its pre-covid peak in Australia and Sweden. In Norway it has tumbled by a remarkable 2.5 percentage points of gdp from its level in 2016. Even in America—the land of costly health care—something has changed. A new measure published by the Bureau of Economic Analysis suggests that the share of spending going on health care has been falling since before the pandemic. A widespread slowdown has never happened before. It has not even come close to happening.
"Inflation in the global health-care industry, which once looked Argentine, now looks rather more normal. Consider a broad measure of American health-care prices, which includes not only things purchased directly by consumers but also those paid for on their behalf, such as by insurers. From the 1970s to the 2000s annual inflation almost always exceeded the average (see chart 2). But in around 2010 that relationship flipped—and much the same is true elsewhere. Relative to the “gdp deflator”, an economywide measure of inflation, the deflator in health and social care across the rich world has pretty much stopped rising. In the 1990s Japanese health inflation soared relative to average prices, but has fallen since 2015. In the French health-and-social-care sector, once dreadful at cost control, prices now grow in line with the economywide average....few people saw this curve-flattening coming. What prompted it?"
"Begin with supply-side factors. Falling health-care inflation is consistent with rising health-care productivity. Economists typically think productivity gains in health care are difficult to come by because the service is labour-intensive. It is, for example, unrealistic to expect a phlebotomist to draw a blood sample 3% faster, year after year. Typically this results in “cost disease”, where spending must rise over time in order for the service merely to stand still. After all, even if phlebotomists do not get more efficient, they still expect a pay rise every now and then—without them they may be enticed to other sectors. William Baumol, who identified this phenomenon, worried about health-care spending swallowing up an ever-larger share of gdp.
"At the same time, it never seemed plausible that health care was entirely immune to productivity gains. Even for an occasional patient it is blindingly obvious that health systems are ravaged by inefficiencies: paper-based forms instead of digital ones; hours spent filing insurance claims; different parts of the system not talking to one another. Meanwhile, some systems do seem to have improved. According to America’s Bureau of Labour Statistics, labour productivity in health care and social assistance fell by 13% between 1990 and 2000, but then made up all the lost ground from 2000 to 2019. In Britain, a study found that staff in the National Health Service (nhs) provided 17% more care pound for pound in 2016 than they did in 2004, compared with productivity growth of 7% in the economy as a whole. All this may have helped keep cost growth under control.
"Another supply-side factor—technological change—may also play a role. Over the long sweep of history, innovations have tended to raise health-care spending. This is in part because they often make therapies available for conditions that were previously impossible to treat. In the 1960s, for instance, the advent of dialysis machines was quite literally a life-saver for people suffering from kidney failure. And yet in the rich world the current cost of a year of dialysis for a patient is somewhere between $40,000 and $60,000, close to these countries’ per-person gdp.
"The nature of technological innovation in health care may now be changing. One possibility is that there has been a generalised slowdown in treatments that represent medical breakthroughs and are costly, such as dialysis. But this is difficult to square with a fairly healthy pipeline of drugs coming to market. Another possibility, which is perhaps more plausible, is that the type of advancements has changed, involving a shift from whizzy curative treatments to less glamorous preventive ones. There is decent evidence that the increased use of aspirin, a very low-cost preventative treatment, in the 1990s has cut American spending on the treatment of cardiovascular diseases today.
"Demand-side factors may also be keeping health-care spending in check. In America the Affordable Care Act (aca)—which was introduced in 2010, at about the time costs tailed off—tightened up the ways in which the government reimburses companies that provide treatment. The aca also made it more difficult for doctors to prescribe unnecessary treatments (seven expensive scans, perhaps, instead of one cheap one) in order to make more money.
"There are similar trends elsewhere. Following the global financial crisis of 2007-09, many cash-strapped European governments decided to reduce spending. This included limiting staff pay rises. The average basic earnings of British nurses are at least 10% lower in real terms than in 2010. Other governments have reduced spending by cutting services, sometimes to the bone. Take Greece, where the ratio of health spending to gdp is the same as it was in 2005. A paper published by the imf noted that, even before covid, the country’s health care was struggling, with “widening inequalities and large unmet needs, especially among the poor”.
"Other governments have saved money by replacing brand-name pharmaceutical offerings with generic equivalents. In the median European country for which there are data, generics take up 50% of the market by volume, up from 33% in 2010. After a patent on adalimumab, which is often known as “Humira” and is used to treat rheumatoid arthritis and other conditions, expired in 2018, the nhs saved around £150m ($200m) a year by bulk-buying non-brand versions of the drug.
"Another demand-side factor relates to overall economic growth. Health care is a “superior good”. When people get a dollar richer, they want more than a dollar more in health care—maybe demanding, say, mental-health care in addition to more traditional life-saving treatments. Across countries there is a strikingly strong relationship between prosperity and spending on health, even in places where the government provides the bulk of the health care (see chart 3). Americans spend so much on health because they are so much richer than almost everyone else.
"These days growth in income per person across the rich world is far slower than it was before 2008. According to our estimates, this explains 40-60% of the curve flattening. This part of the story is therefore a pyrrhic victory: health spending is not growing in part because the world has stopped getting much richer.
"How long will the curve stay flat? Spending is being pulled in different directions. An ageing population will continue to push up demand. In some countries the pandemic appears to have dealt a blow to health-care productivity, which may not yet have shown up in the data. On the other hand, America’s Inflation Reduction Act allows Medicare to use its purchasing power to bludgeon pharma companies into lowering prices. And economic growth remains weak. What is clear for now, though, is that the received wisdom is wrong. Health care need not eat the world."
Specifically on the 50% of untaxed ..short and long term capital gains, high estate wealth transfers. Mandatory means evennon companies hiding income thru financial engerineerd losses
"50% of untaxed ..short and long term capital gains"
What capital gains are not taxed? Selling one's primary residence can give a nice exclusion (I think is up to 500K for married), but so much home appreciate is just inflation, so you are otherwise taxed on inflation.
I'm no fan of EA, nor an AI-doomer, but your comment on OpenAI seems mistaken to me.
"And “shut it all down” is what the OpenAI board seems to have had in mind when it pushed the panic button and kicked Altman out. But the effort collapsed when OpenAI’s workers and financial backers all insisted on Altman’s return. Becuase they all realized that “shut it all down” has no exit strategy."
It seems plain that "shut it all down" is not what the board "had in mind". If so, then OpenAI would not exist as it currently does.
I would suggest, instead, that the board believed in the principles of the founding of the corporation; that is: "The Company exists to advance OpenAI, Inc.'s mission of ensuring that safe artifical general intelligence is developed an benefits all of humanity." That is, not necessarily to grow or to make a profit (explicitly, regarding the latter). And that they realized (too late, apparently) that running the company as an SV startup ("grow!", "move fast!", "break things!", "get rich!") was not conducive to the goals of the foundation. And likely (also too late, apparently) realized that coupling themselves too tightly to Microsoft was a similar mistake.
Now Microsoft is in the driver's seat, and I don't see the new board as having much focus on the founding principles.
Are people really doubtful about increased disaffection among the US working class? And the supposed European counterexamples aren't too convincing as they're less economically cutthroat than the US--if anything wouldn't that strengthen the argument that there's some causal nexus between US capitalistic culture/socioeconomics and the mounting disaffection? Even if we set aside particular health outcomes or dispute vocabulary like despair and stress, it still seems like there's some pernicious social phenomenon that merits scrutiny.
Combining 2 thoughts from this post: income inequality and deaths of despair. Sociologists who study these topics have data that clearly show the relationship between impoverishment (the bottom end of inequality) and mortality/health (shortened life expectancy, including "deaths of despair" and reduced adult height). Please see Dr. Peter Turchin's data at peterturchin.com. In terms of cause and effect, evidence-based sociologists clearly show a chain of cause/effect in many societies worldwide: inequality > impoverishment > rising fascism, because the impoverished look to a strong leader to destroy the system for them. Noah, please see Dr. Turchin's data and provide your reaction.
I have followed Turchin's writing for years. As for your comment, inequality and impoverishment are not the same things. It is one thing to be against impoverishment, but shifting this to an argument against inequality requires some huge assumptions, many of which don’t work. Entrepreneurs and surgeons getting paid more to enrich or save more lives does nothing to affect my height, nor cause me to drink more or become a junky. Indeed, they are likely to improve my life.
I don’t doubt that there could be a link between between "Unfair Inequality" (note the first word does most of the work here) and totalitarian support (fascism, socialism, etc). But note too that dysfunctional incentives and welfare traps can do the same types of things:
— encouraging uneducated non English speakers to come to the country
— encouraging clannish-based cultures to immigrate to liberal societies
— paying mothers not to get married or work
— subsidizing young men to get useless degrees
— paying welfare to people to stay in places with no jobs
"Inequality" is one of the most disingenuous terms in usage today. Those doing the statistics wildly distort it and include or exclude things for the sake of their narrative. They pretend that inequality is synonymous with unfairness or guaranteed to lead to fascism.
The last part of this sentence is an oversimplification.
Fascism scholar Roger Griffin describes fascism as palingenetic (rebirth) ultranationalism. Fascism is a mass politics, and there isn't a single class that acts as a kernel of political potential for fascism to emerge.
So poors aren't the kernel to mobilize fascism, even though they can support fascism for those ends. They are also supporting the same party as the beleaguered merchant and financial class fearing a challenge from below, military brass who yearn for a martialized culture -- the Nazis didn't invent the image of the German as an officious, disciplined brute; the militarization of culture was formed and perfected by von Bismarck -- religious leaders who fear and detest a modernity they can't control; even the bohemians like poets, artists and musicians who see the world in idealized, utopian terms, etc.
Fascism was never about coherence, but rather an eclecticism which makes it a formless ideology that allows anyone to project their perceptions onto it and see what they want.
Totally agree with everything you wrote. I phrased it as I did for simplicity and impact. The actual mechanics of a rise of a counter-governmental movement (call it what you will; I called it fascism) must include disaffected elites (or those who aspire to elite status) who have the education, social network, and skills in writing & speaking to lead the masses in their direction. The masses must be disaffected and ready to be mobilized. That's where the US is today, with a high "mass mobilization potential," the source of which is 2 generations of impoverishment due to income inequality. When a counter-governmental elite harnesses their discontent, the US will be in real trouble. (Trump may or may not be it.) See the work of Dr. Peter Turchin (peterturchin.com) and his fellow researchers for the evidence (data and models) that supports this common phenomenon.
Income inequality is not the driving force for U.S. fascism. Racial ressentiment and status anxiety among upper-income strata Whites are the salient factors.
Trump's base is the sore winner.
Ta-Nehisi Coates foretold what Trumpism was going to be in 2017 in his Atlantic essay "The First White President." He ran the numbers on who Trump voters really were by income, education level and geographic area. Trump won Whites as a race, period, and his voters were no different than previous Republican candidates.
As a chaser to the shot of "The First White President," Robert Pape did a demographic profile of the Jan. 6 insurrectionists. The modal insurrectionist is a status-anxious asshole boss. Again, sore winner.
I'm familiar with the work of Peter Turchin and his hypothesis of elite overproduction. The trouble is, if you advance his idea of disaffected elites (in this case, more postgraduate degrees than positions of political power or economic stations that could absorb them), they are pulling the country leftward and are a countervailing power against fascism. Turchin is also the one who says the French Revolution was a backlash by Jan Bradys (the middle and younger children of the aristocracy glommed onto liberté, égalité, fraternité as a way to topple primogeniture, or power accruing to their firstborn brothers).
My belief in anyone's theory as to how anything operates is their evidence related to cause and effect. Turchin has the most credible data on social upheaval. I agree that disaffected elites are pulling the country leftward, but those elites are not speaking to the disaffected masses. The disaffected will respond to someone from the right who speaks in terms that matter to them: political violence against the government.
Piketty: “inequality denial is as dangerous as climate denial.”
Prices in a properly functioning market act as signals and incentives. That means that in a properly functioning market it is entirely possible that inequality in wages or differences between capital and labor are the signals telling us to invest more in skills and investment and entrepreneurial ideas and less in rote labor. The incentive is to get a better education, move to better jobs, or start your own lawn-care business.
If Piketty or anyone else thinks that rising inequality is de facto a negative, then I would argue they are missing a fundamental grasp of economics. That said, I understand that inequality could be a sign of rent seeking or market distortions, and that rising inequality can create various externalities.
Adding insult to injury, when I see economists decry inequality, top on their list of remedies is usually more powerful unions (which are by definition market distorting, rent-seeking labor cartels).
Prices do two things. They give incentives and they transfer income. You may approve of the incentive giving without approving of the income distributed as result.
What do you base your disapproval on? I assume you don’t disapprove of brain surgeons making more than janitors? When is unequal unfair in your opinion, and when should it override price signals?
I saw that in the aftermath of the OpenAI coup, they added Larry Summers to the board and that only added to my existential dread.
Will we all feel better about AI running amok because Larry Summers is there to write a reassuring op-ed about how the AI’s decision to liquidate 70% of us into Soylent is going to be tremendous for the stock market and economy?
No.
And that’s basically all he’s ever been good for. Larry Summers cares so much more about money as a line going up, than he does about people or humanity as a whole.
re: " narratives; Piketty thundered on Twitter that “inequality denial is as dangerous as climate denial.” (Note: No, it is not.)"
Just an ignorant dilatant, but IIRC Piketty argues that inequality distorts economic/public policy decisions to favor the wealthy. The Fossilistas are a prime example.
Noah: I think one of the strongest arguments in favor of renewables at this point is the very long term problem of heat waste. I'd imagine you've read about this phenomenon: many decades from now, global warming will re-emerge as a huge problem even *after* we've completely decarbonized simply because, when humans employ energy to do things, some of it dissipates into the atmosphere and increases temperatures.
Anyway, as I understand it, the only energy technologies that can completely avoid this (very serious) issue, long term, are solar and wind, provided the systems are properly designed. Nuclear doesn't cut it. It's not a problem that will become huge in scale for a very long time (maybe several centuries) but it's a very real issue.
The article says that waste heat will pose a problem in about 200 years, but not really today. We replace our electricity generating infrastructure about once every 50 years. So we have 4 technology generations to figure this problem out. Also, waste heat goes away the same day you shut down the power plant, while CO2 keeps on warming indefinitely.
It’s an interesting problem to think about. And I’m in favor of any argument to get people to build more solar. But I wouldn’t consider this a serious reason to not build nuclear power today (if building nuclear was seriously on the table). Just like how natural gas power plants are complementary to renewables and cleaner than coal, we shouldn’t let perfect be the enemy of good.
It says it will be a problem in 200 years _if_ we keep doubling our energy usage every 30-50 years.... or in other words, if we increase our energy consumption by a factor of 100 or so.
But of course it's very unlikely that we could get 100x our current energy consumption from (terrestrial) solar anyway, even if we could think of a good way to use 100x more energy. It would require tiling too much of the Earth's surface with solar panels.
The problem with heat is it can’t radiate in to space due to the heat absorbing chemicals being released in to the atmosphere. This is where nuclear IS the answer. It’s energy dense, efficient with minimal waste and doesn’t release heat trapping chemicals into the atmosphere. Also consider that solar panels may not emit heat but the things it powers, like generators, lite bulbs, electric heaters transmission loss etc, does. Also basic math says that solar and wind can’t produce the amount of energy needed to satisfy todays needs much less increasing future needs as Germany is finding out the hard way. Climate change is a problem but it won’t be solved by ridge green ideology and wishful thinking. Energy use is directly tied to prosperity and abundance. Energy security is directly tied to national security. I seriously doubt the average citizen and their leaders would be willing to accept less of either. Doomberg has a great series of articles in Substack on this which I highly recommend.
We want less net CO2, not less energy. True we should be developing nuclear and geothermal, too. The more different kinds of zero-CO2 technologies we have the more lower the average cost as some technologies are lese expensive for different uses.
> consider that solar panels may not emit heat but the things it powers… does
This is the insight of the article - renewables capture energy that would have turned into waste heat if we didn’t capture it. They actually cool the earth a little bit during the capture process and heat it back up when the energy is used, coming out close to neutral.
> Basic math says that solar and wind can’t produce the amount of energy needed to satisfy todays needs
The basic math works out quite fine, we would only need to cover a fraction of the world’s deserts in solar panels to power all of humanity. It’s the details that are hard - transmission, storage, minerals, permitting, etc.
As you so eloquently pointed out with your details, the math doesn’t work out. The amount of engineering, current battery technology, transmission loss trying to get the energy from deserts to places where most people live, the low energy density which is further impacted by darkness and clouds not to mention the ecological impact makes this impractical.
It would make for an interesting paper. How much DAC would you need to offset waste heat and what would that entail in terms of energy and materials? Surely the DAC machines would also create waste heat…
So define "cheap energy". It might be cheap for generating entity but not for the user, as we are now seeing in CA where the costs of renewables are ever increasing and severely affecting especially lower income people.
Heat is wasted also when electricity (regardless of how it is generated) is converted to mechanical energy (in EVs, heat pumps, motors) or computation (AI, data centers, crypto) or heating. Nuclear electricity generation is engineered to not waste any heat and convert as much as possible to electricity. Wind turbines also waste heat converting rotational energy to electricity and solar panels reflect back IR as waste heat. Waste heat is not going to be a problem until long after we will have technology to capture it, and it seems like you may just be inventing new reasons to be anti-nuclear.
Effective altruism is like macroeconomics: a good intuition pump that some people mistake for all-inclusive blueprints.
I'm very glad the people working on AI are keeping an eye out for the wackier possibilities. The actual data doesn't so far seem to justify either hosannas or panic, but it's good the actual professionals take their own work seriously.
The even scarier thing is the plenitude of countervailing incentives--financial and otherwise--for people to ignore the wackier possibilities, or even to disparage those who warn about them.
EA seems to have found a good one; just tell rich people they're going to live forever and then scare them into worrying about far-off theoretical problems.
Actually, that's a big open question in alignment research: how much does "commercially reliable" overlap with "safe for humanity"?
Suppose you make a financial-trader AI. If that AI lies to managers about the expected profits, or even just flatters money managers instead of being honest, it's a bad product that investment banks and hedge funds won't feel safe purchasing.
What does that mean? It means that even if AI makers only care about profit, they still want AI that won't lie to or flatter users in critical applications. We don't need extra regulations for companies to fear deceptive or sycophantic AI. Deceptive and sycophantic AIs are, in many applications, a bad commercial product that will cost you sales and reputation.
On the other hand, lying about near-term consequences and the distant future are two different kind of lies. You can imagine an AI that's very honest about its financial advice, because it knows it'll be caught, but sneakily running its own programs in the background, because it knows those won't be found out until it's too late.
So we have the key question: is "don't lie, it's bad for our business" enough the same as "don't lie, it's bad to wipe out humanity"? When AI makers train the AI to be short-term honest with users, will that be enough to also get us long-term honesty and safety?
There's often a lot of uncalled-for anthropomorphization here.
Being able to make the decision to lie, having any self-reflection ability, having any long-term memory ability, and being able to trade stocks are all completely different skills. There's no reason to think any system would come with any particular combination of these skills just because it's been called an AI. And what we currently call AI (LLMs) doesn't have any of them.
1. the model actually learned wrong facts, or else knows something "imprecisely" (imagine the right answer gets "rounded off" because the model is too small).
2. Answer generation is done via a sampler algorithm that only goes forward and isn't itself "intelligent". So if you ask a question it doesn't know, and then the first word it generates is "yes", it sort of forces the model into making up an answer because it can't start over again.
There's a lot of research into improving this, but some of the issues are fundamental.
It seems like Piketty’s morally tinged outrage about “inequality denial” rests on the assumption that the outlandish wealth accumulation of the underserving rich somehow impairs the economic welfare of the lower classes and the poor. How true is that?
I mean, if you take consumption as a more direct measure of standard of living, in the US inflation-adjusted personal consumption expenditures per household continues to rise decade over decade.
Which is not to argue that there isn’t some low hanging fruit to improve people’s economic security, like universal health care coverage.
Rich people hoarding money is bad for poor people if the world is zero-sum. And unfortunately zero-sum thinking was a self-fulfilling prophecy in the 2010s. Low economic growth encourages zero sum thinking about redistribution, and economic growth was low because of a fixation on balancing the budget. And people had a fixation on balancing the budget because they were in a zero sum mindset.
"economic growth was low because of a fixation on balancing the budget"
Economic growth 2008-2020 was low because of bad monetary policy, specifically, under-target inflation. Balancing the budget (at least to extent of financing all non-high yield investment from taxes) would be pro growth.
Great stuff re: EAs, despair, and coal. But solar and land use:
Jesse Jenkins: ...[In] The most cost-effective of our net-zero scenarios, [wind] spans an area that is equal to Illinois, Indiana, Ohio, Kentucky, and Tennessee put together. And the solar farms are an area the size of Connecticut, Rhode Island, and Massachusetts.
Warning: On topic but also on the nose ... Can AI ever advance to the point where it can put Illinois, Indiana, Ohio, Kentucky, and Tennessee to their highest use and turn everyone in them into wind turbines? 🤔
I agree that EA is worth rescuing from some of its high-profile advocates, but I'm still worried about being turned into a paperclip.
(Steven Pinker not being worried makes me worry more, because he's reliably wrong about everything outside his area of expertise. I think he and Richard Dawkins must have made some sort of pact.)
Could you give some examples of Pinker's wrongness outside his area of expertise?
I found Pinker's view on AI somewhat naive, but on FOS and Better Angels, I though he was right.
re: japan, i still can never really understand economic reports and analysis of japan’s economy. i’ve been visiting japan for the last 40 years and all i see is advancements everywhere and is still the most futuristic place til this day in addition to social stability, social welfare, and overall widespread wealth like ive never seen. you will never ever see distressed people walking around like zombies fiening for their vice of choice on the streets, it’s just not there. what am i missing?
I think the experience of being a tourist is radically different from the experience of citizens.
Roger Cohen once wrote an irritating column about how Italy is actually really nice despite all the economic problems people complain about, which didn't impress my Italian friends much. And my general experience is that living in developing countries is very pleasant as long as you're not on a local income.
As someone that lives in Tokyo, I feel the experience is quite nice?
On a good, but far from great, local salary, my partner and I are able to eat out all the time and live in an apartment that is impossibly nice by SF standards. While I was able to eat out all the time in SF, being mostly single and in tech, that just isn't possible in most/all of Western Europe.
Public infrastructure just works. Trains are generally on time and go everywhere I care about going, utilities are reliable, and street surfaces are well maintained. In SF, public transit was a joke, electricity and internet cut out many times a year, and on the off chance of rain, the uneven road surface would pool up water for cars to splash into anyone walking.
Safety and cleanliness is extremely high. It's kinda hard to relax in public in the US or even Europe. Maybe used needles aren't a problem with functional shoes. Maybe drug zombies are not terribly likely to actually attack you. Maybe the salaries pay enough to buy a new bike every few months. However, just not having to deal with all of that is incredibly nice.
Tourists see public spaces but rarely see private spaces. Japan has fantastic public spaces but pretty poor private spaces (the average dwelling is 93 square metres). The US is way on the other end of the spectrum as far as developed countries go; many public spaces (city streets, public transport etc) rather suck, but go inside people's houses where they actually live and they're the nicest in the world.
Drug zombies are a separate issue. The other weird thing about the US is that the population of complete bottom-crawlers incapable of life in a civilised society is about ten times larger than it is in most other Western countries. A complete failure to enforce drug laws doesn't help either.
The US's violence and drug problems are better compared to other New World countries. Though, being able to afford drugs in a way is a sign of wealth.
> Though, being able to afford drugs in a way is a sign of wealth.
A pill of fentanyl on the streets of Seattle is $3-5 as per Seattle PD, and bulk purchases can go below $0.50 a pill.
That isn't really a sign of wealth amount of money for developed countries. Mass overdoses from cocaine might be a sign of wealth, but that isn't really what is happening.
I was mostly thinking of the time you'd need for your "hobby". Some people have a lot of time because they have money, some have time because they can't make any money.
i can’t say i’m a tourist, my life is deeply intertwined with japan and i’m in there like swimwear
I think a really big part of it is that Japan is optimized for selling cheap things. Right by the station in Azabujuban, which is a pricy business district, you can get a hot bento lunch with chicken and veggies and rice for $3, or fried chicken and a rice ball for $2. You could live half an hour from that station by train for $150/mo in 150 square feet. The $15 jeans from Uniqlo are nice enough that I wear them over stuff I brought from the US. And I'm in a Facebook group that's drowning in "$8/hr installing solar panels, no language skills required, housing and food provided" style job listings. You can buy lightweight cars that go 30mph for $5000.
This seems like a different thing than PPP - expensive stuff is still expensive. Cheap stuff is just really nice and easy to find.
Is Japan futuristic? I feel like the most recent technology improvement I saw was that a lot more restaurants started taking credit cards. It has great well-run infrastructure sure, but not brand new ideas.
I've also heard Korea is even newer.
Their wages are also kind of low, especially with the exchange rate at $1=150jpy. A lot of people are leaving for Australia which pays like twice as much.
It's less that Japan is futuristic, and that the West is incredibly behind on the adoption of many great technologies. Though to be fair, a lot of what makes Japan so futuristic also applies to Korea, often to a greater extent.
For example, Japan hit 90% adoption rate of heat pumps in the 2000s, and the west has lagged behind so far that it's common for expats to make fun of Japanese people running the air conditioning in reverse like cavemen who never discovered they could just burn dinosaur juice instead.
Or that self checkouts and fast food ordering kiosks are mature, reliable devices that are so common and easy to use, that even plenty of retired tech-phobic types use them without issue. It's just incredible how bad and unpopular self checkouts and fast food ordering kiosks in the US still are in comparison.
Or everything about toilets. Or faucets that allow you to set a desired temperature. Or hot water that comes on in seconds and never runs out. Or how common video chat doorbell systems are in apartments. Or motion sensor based lights that turn on when you walk in the door. Or how you never have to climb stairs into the train except on a handful of legacy historic tramways. Or ordering tablets and call buttons at restaurants. Or elevators where you can cancel mistake floor requests. The list really goes on and on and on.
The West is so far behind that many of these technologies have become so mature, people like you have trouble recognizing how far the West is behind on them. It feels like a conscious cultural choice that the self checkout at Walmart is a buggy piece of shit while the self checkout at Seiyu just works, not the US being so far behind on basic technology.
If anything, that's one of the big problems Japan has. Japan never managed to capitalize on its position of futuristic thought leader back when it had it, and has since lost it.
You'd think that the vast majority of American homes having hot water that takes ages to come on, delivered through faucets that required constant fidgeting to maintain a constant temperature, would be an opportunity to sell some futuristic Japanese technology that solves all of that. However that just doesn't happen.
How do they get instant hot water?
korea’s industry maybe so, but seoul is nowhere near tokyo in terms of infrastructure
How so? Tokyo is an amazing place, but so is Seoul. I find the latter's subway to be cleaner, newer, and easier to navigate. I'd give Seoul the edge in airports, too. Both cities are connected by HSR. I've enjoyed my time in both cities, and if you forced me to choose one to live in, I'd probably go with Tokyo mainly because its winters are quite a bit milder and feels less likely to be attacked (all those artillery tubes aimed at the South Korean capital!).
But I personally don't see an infrastructure advantage for Tokyo.
I always look forward to these "five things" posts, because it's rare that i don't find a few informative and thought provoking. This one was for me five for five, so I wanted to thank you for doing these posts.
One comment. I think wealth, rather than income, is a much more important indicator of inequality for the very rich. That's the measure that the 0.1% and above care about. And that's the measure used for comparison purposes among the very rich.
But that is tremendously more complicated, it involves valuing illiquid assets, privately held companies. Integrating debt etc. So if there's a wide range of empirical outcomes on income inequality, wealth numbers over time are gonna be close to worthless
I think it’s worth distinguishing between two arguments.
The good argument: the real value of illiquid assets has changed over time (e.g. investors become more comfortable with high P/E ratios when interest rates were zero). Therefore it’s hard to know whether the rich are actually richer than they used to be.
The bad argument: illiquid assets aren’t the same as liquid assets, therefore we can’t really measure them, and therefore billionaires aren’t actually stupendously rich. (I see this one float around Reddit a lot, usually as a defense of low tax rates).
It sounds like you’re making the good argument though.
Yeah, it's the former. I have worked in wealth management and I am intimately familiar with all these devils in the details, and how much accounting in illiquids is much more art than science lol.
It doesn't make me in favor of low taxes by any means, But it does lead me to want much higher estate taxes with much lower exemption thresholds but less interest in trying to work out annual wealth taxes. But I feel that way for plenty of reasons.
I'd start with eliminating step-up of capital gains on inheritance.
I agree that wealth is even more difficult to measure. But it's far "stickier" than income and it really is how the very rich measure themselves. So true inequality, at this time, may be impossible to measure.
Well stickier in some sense. But if 70% of your wealth is tied to your public company, and it makes more profit than last year, but because interest rates quadruple, the forward p/e decreases 60%, mark to market you are significantly less wealthy. However, on a discounted cash flow basis, maybe it's ultimately inconsequential long term, but the liquidation value has fallen dramatically. Given how much wealth in that tier is tied to equity ownership public or private, and how correlated that can be, it seems like a quagmire. I get that in some sense the 'economic and political power' from owning a firm doesn't really shift with those considerations, but that feels separate.
Yes, perhaps harder to measure - BUT that doesn't mean that the effect isn't real. Just that there isn't a sufficient incentive (i.e. someone willing to pay for the research) to do it. Hmm, wonder why that is.....
Are you saying that because the rich care about wealth rather than income it's what we should all care about as the important measure of inequality?
No, I'm saying at the uppermost reaches, wealth matters more than income as a measure of inequality.
Re: deaths of despair
My sense is Deaton fell in love with a conceptualization of his own that he initially found clever, and wouldn't back off out of pride or stubbornness. It's an understandable and common human foible. But one that is nonetheless damaging.
Americans do tend to die younger than their counterparts in other rich countries. And as far as I know the gap has widened over the years (the USA's relative position would have looked a lot more favorable in, say, 1970). It's not unreasonable to posit that America's greater tolerance of economic inequality and less effective safety net are among the drivers of this phenomenon.
But as I understand it, Deaton tried to come up with a neat and tidy theory that connected the above with the rise of the MAGA movement. And this necessarily focuses on the white working class. And that's where the idea falls apart, because there's overwhelming evidence the particular spike among this cohort is associated with opioids. Needless to say there have been major differences in how America and the rest of the rich world have handled the emergence of the new generation of these drugs. In short, US practices were basically *designed* to get people addicted to drive pharm products. It's not despair that has caused so many working class white Americans to die prematurely; it's been the intersection of brain physiology and very poorly designed painkiller regulations that have done that.
It’s an American issue for sure but it’s basically guns, cars, obesity and opioids, not the lack of a safety net. For one thing, America’s safety net for people 65+ is as good as anywhere else.
You're conflating what Deaton was referring to—a sharp rise post 2000 (especially post 2010) in mortality among non college whites—with the overall phenomenon of lower US life expectancy (vis a vis other rich countries). I agree the factors you cite all play a role in the latter. As to the former—which Deaton erroneously pins on Dickensian steroids—it's opioids.
Yeah. And Burn-Murdoch's graphs show it's mostly obesity.
Then too, the safety net could be helping more with opioids.
Obesity is a problem, but the "deaths of despair" narrative was specifically trying to explain a sharp rise in mortality among the white working class in the new millennium, and obesity rates have been rising for many decades. Also, Blacks suffer from obesity at higher rates than whites.
There are number of contributing reasons as to why Americans don't live as long on average as the denizens of other high income countries, for sure. But general, lower life expectancy in the US is *not* what Deaton's "deaths of despair" claims were referring to.
I got the impression the problem was almost entirely drugs, not obesity, can you link to these graphs?
On an individual level I don't have much sympathy for those who take drugs, but definitely more needs to be done to stop them.
The graphs are here: https://twitter.com/jburnmurdoch/status/1641799922583326720
The dominance of obesity (cardio-metabolic) is stunning if that's the full story. Of course, this is all vs. England.
Thanks for the graph.
I'm having trouble in my head determining whether it's a fair comparison. Obviously, magicking away "cardio-metabolic deaths" (a huge cause of death, possibly the majority) is going to have a much larger effect than magicking away minor causes of death.
What I do find interesting is that the very recent decrease in US life expectancy does seem to be attributable entirely to drugs. Without drugs, life expectancy has been flat since 2015 as it has in most other countries.
Great point about drugs after 2015.
Re the cardio metabolic line: these plots show differences, so I guess the differences will be due to eating and other behavioral choices. But I do wish I could see the article to know more details.
“Income inequality” is one of those expressions that sounds bad but is quite obviously a necessary feature in any robust, competitive economy.
It’s on a spectrum right? Makes sense to mitigate it to the benefit of people who were disadvantaged in one or more of the myriad ways one can be. Plus the whole point of a robust economy is to benefit the humans operating within it. It’s supposed to be in service of us (all of us) rather than vice versa.
The American Left is really bad at touting its accomplishments. So the narrative becomes "we've made no progress" instead of "we've made progress but will have to keep working as conditions evolve."
Yes exactly. Income inequality is better than at basically any time in civilized human history, the US had a very large and growing upper middle class, and poor people in the US are more likely to suffer from obesity than hunger. Let’s focus on real problems please
Yes, Income inequality has been improving continuously throughout human history. That is, up until the last 50 years where it has been getting worse. This is a problem and we absolutely will focus on it.
Well, that was true well into the 80s, but fairly clear that it has turned since then - all the studies that Noah cited (well, except for the most extreme) show a growing income inequality (and I'd argue much more growing wealth inequality). And I sure know a lot of poor and homeless folks who don't suffer from obesity (especially the children).
For a sizable time after WWII, we were moving towards a diamond shaped economy (and Europe as well), but at least in the US, that has been shifting ("hollowing out of the middle class") - hidden for a while as women (and hence 2 earner families) joined the workforce.
The ideal amount of inequality in society is probably not zero, but it probably is lower than any existing national economy so it seems like a somewhat misplaced concern to worry about an economy becoming too equal.
The ideal amount of income inequality is probably equal to the actual amount of productivity inequality. If some people are one hundred times more productive than others then they should be one hundred times richer.
The ideal amount of _productivity_ inequality is probably lower than it is, and this decrease in inequality should come entirely from improved productivity at the low end.
Yes, maybe, but I can tell you just from software engineering, that income is rarely related to productivity in the real world. And I doubt that teachers, etc. are not productive compared to the gross low income
An issue is that some people are valued on their past work and some people are valued on their expected future work. Generally you become switch by switching from the first to the second.
eg becoming a hit artist and selling off shares in your future income.
I believe that a society where pay reflected productivity would actually be more unequal than the status quo. People just aren't equally productive! Even within the same job, some of my colleagues are better at the job than me, some are worse, but everyone working under the same job title earns the same.
First one should consider the political bent of the authors. One set appears to be in the pro-socialism/anti-capitalistic realm that colors their assessments and decisions. The other is more grounded in the capitalistic system that has overwhelmingly benefited humanity.
Another problem is the value-liquidity issue. At one time not so long ago there were more millionaires per capita in Idaho than any other state, yet the standard of living was one of the lowest. The reason was land ownership. Large acreages equaled low liquidity, and many of those same millionaires were barely economically keeping their heads above water. So how does this fit the inequity model?
Income inequality is indeed necessary for a strong, innovative economy. The question is *how much* is optimal, especially after accounting for taxes and transfers.
Deaton did point to one feature of as-existing "Capitalism" that contributes to stress/despair: employer "provided" health insurance. Although paid for by employees' lower wages, for any specific employee, it is a cost to the employer, a fixed cost per person. This makes it much higher percentage of wage for the minimum wage employee that the CEO. This inclines employers to substitute capital and higher wage labor for low wage labor and to strategies of part-time, contractor status employment to avoid that cost. It was a mistake of ACA not to have structured the program so that employers could "dump" low-income employees on to a more generous ACA.
The ACA was about as ambitious as it could have been given the political constraints. A public option does seem like a reasonable next step (and a mechanism to uncouple insurance from employment) but it's not happening without large Democratic majorities or Republicans supporting it (neither is happening anytime soon).
There is certainly merit in the government making the lives of citizens easier and more predictable like simplifying tax prep and making enrollment in government programs automatic and seamless. It is stressful to worry that a paperwork error can drastically affect your life! It's odd, but a lot of Americans don't seem to like the government to function TOO well.
Your last sentence could have ended with "government[.]"
Not far from the truth lol
That's not a feature of capitalism, it's a consequence of the US tax code (i.e. government, not capitalism). Most countries don't do this, employees pay their own health insurance.
Single payer, one national Healthcare system. Destroy all Healthcare Insurance as negative value add.
Obama big mistake
Isn't one issue that health _care_ is much more expensive in the US than in most European countries, before we even get to the question of health _insurance_?
Actually, the US pays nearly 20% of GDP for Healthcare, while every modern free market democracy pays 10% of GDP. That is an annual GDP waste of $2.4 Trillion every year.
Health Insurance is the biggest scam ever perpetrated on the American public.
a. Health is NOT a product. Sons, daughters, wives and husbands are fungible like Ford vs Chevy vs Dodge trucks.
b. Health Insurance NEVER made one person healthier, but has made health outcomes worse as we know that Americans are living far shorter lives that people under Ntional Healthcare.
c. All we need is uncapped mandoratory Medicare payments on all income by every person.
From the Economist dated Oct 26th 2023. Headline: "How health-care costs stopped rising"
[Note -- the Economist uses 'America/American' to refer to the United States.]
Text: "For a long time, health care was eating the world. From 1950 to 2009 American spending on hospitals, medics and the like rose from 5% of gdp to 17%. Between the late 1970s and the mid-2010s British public spending on health rose by 4% a year in real terms, much faster than the economy’s growth of 2% a year. From 1980 to 2010 overall French prices rose by 150%; the price of caring for a sick or old person rose by 250%. Among economists, the proposition “health care’s share of gdp rises” was almost as close to an iron law as “free trade is good” or “rent controls do not work”.
"The iron has now melted. Even as populations age, and as the world continues to deal with the fallout from the covid-19 pandemic, health care is no longer taking over the economy. Across the rich world health care’s share of gdp jumped in 2020 and 2021, because of a combination of pandemic-related spending and lower gdp. However, it has since fallen back to close to its level in 2008 (see chart 1). Because of this “flattening of the curve”, health spending today is somewhere in the region of $2trn below its pre-2009 trend.
"In some countries the changes are still more dramatic. The ratio of health-care spending to gdp has fallen from its pre-covid peak in Australia and Sweden. In Norway it has tumbled by a remarkable 2.5 percentage points of gdp from its level in 2016. Even in America—the land of costly health care—something has changed. A new measure published by the Bureau of Economic Analysis suggests that the share of spending going on health care has been falling since before the pandemic. A widespread slowdown has never happened before. It has not even come close to happening.
"Inflation in the global health-care industry, which once looked Argentine, now looks rather more normal. Consider a broad measure of American health-care prices, which includes not only things purchased directly by consumers but also those paid for on their behalf, such as by insurers. From the 1970s to the 2000s annual inflation almost always exceeded the average (see chart 2). But in around 2010 that relationship flipped—and much the same is true elsewhere. Relative to the “gdp deflator”, an economywide measure of inflation, the deflator in health and social care across the rich world has pretty much stopped rising. In the 1990s Japanese health inflation soared relative to average prices, but has fallen since 2015. In the French health-and-social-care sector, once dreadful at cost control, prices now grow in line with the economywide average....few people saw this curve-flattening coming. What prompted it?"
/1
"Begin with supply-side factors. Falling health-care inflation is consistent with rising health-care productivity. Economists typically think productivity gains in health care are difficult to come by because the service is labour-intensive. It is, for example, unrealistic to expect a phlebotomist to draw a blood sample 3% faster, year after year. Typically this results in “cost disease”, where spending must rise over time in order for the service merely to stand still. After all, even if phlebotomists do not get more efficient, they still expect a pay rise every now and then—without them they may be enticed to other sectors. William Baumol, who identified this phenomenon, worried about health-care spending swallowing up an ever-larger share of gdp.
"At the same time, it never seemed plausible that health care was entirely immune to productivity gains. Even for an occasional patient it is blindingly obvious that health systems are ravaged by inefficiencies: paper-based forms instead of digital ones; hours spent filing insurance claims; different parts of the system not talking to one another. Meanwhile, some systems do seem to have improved. According to America’s Bureau of Labour Statistics, labour productivity in health care and social assistance fell by 13% between 1990 and 2000, but then made up all the lost ground from 2000 to 2019. In Britain, a study found that staff in the National Health Service (nhs) provided 17% more care pound for pound in 2016 than they did in 2004, compared with productivity growth of 7% in the economy as a whole. All this may have helped keep cost growth under control.
"Another supply-side factor—technological change—may also play a role. Over the long sweep of history, innovations have tended to raise health-care spending. This is in part because they often make therapies available for conditions that were previously impossible to treat. In the 1960s, for instance, the advent of dialysis machines was quite literally a life-saver for people suffering from kidney failure. And yet in the rich world the current cost of a year of dialysis for a patient is somewhere between $40,000 and $60,000, close to these countries’ per-person gdp.
"The nature of technological innovation in health care may now be changing. One possibility is that there has been a generalised slowdown in treatments that represent medical breakthroughs and are costly, such as dialysis. But this is difficult to square with a fairly healthy pipeline of drugs coming to market. Another possibility, which is perhaps more plausible, is that the type of advancements has changed, involving a shift from whizzy curative treatments to less glamorous preventive ones. There is decent evidence that the increased use of aspirin, a very low-cost preventative treatment, in the 1990s has cut American spending on the treatment of cardiovascular diseases today.
/2
"Demand-side factors may also be keeping health-care spending in check. In America the Affordable Care Act (aca)—which was introduced in 2010, at about the time costs tailed off—tightened up the ways in which the government reimburses companies that provide treatment. The aca also made it more difficult for doctors to prescribe unnecessary treatments (seven expensive scans, perhaps, instead of one cheap one) in order to make more money.
"There are similar trends elsewhere. Following the global financial crisis of 2007-09, many cash-strapped European governments decided to reduce spending. This included limiting staff pay rises. The average basic earnings of British nurses are at least 10% lower in real terms than in 2010. Other governments have reduced spending by cutting services, sometimes to the bone. Take Greece, where the ratio of health spending to gdp is the same as it was in 2005. A paper published by the imf noted that, even before covid, the country’s health care was struggling, with “widening inequalities and large unmet needs, especially among the poor”.
"Other governments have saved money by replacing brand-name pharmaceutical offerings with generic equivalents. In the median European country for which there are data, generics take up 50% of the market by volume, up from 33% in 2010. After a patent on adalimumab, which is often known as “Humira” and is used to treat rheumatoid arthritis and other conditions, expired in 2018, the nhs saved around £150m ($200m) a year by bulk-buying non-brand versions of the drug.
"Another demand-side factor relates to overall economic growth. Health care is a “superior good”. When people get a dollar richer, they want more than a dollar more in health care—maybe demanding, say, mental-health care in addition to more traditional life-saving treatments. Across countries there is a strikingly strong relationship between prosperity and spending on health, even in places where the government provides the bulk of the health care (see chart 3). Americans spend so much on health because they are so much richer than almost everyone else.
"These days growth in income per person across the rich world is far slower than it was before 2008. According to our estimates, this explains 40-60% of the curve flattening. This part of the story is therefore a pyrrhic victory: health spending is not growing in part because the world has stopped getting much richer.
"How long will the curve stay flat? Spending is being pulled in different directions. An ageing population will continue to push up demand. In some countries the pandemic appears to have dealt a blow to health-care productivity, which may not yet have shown up in the data. On the other hand, America’s Inflation Reduction Act allows Medicare to use its purchasing power to bludgeon pharma companies into lowering prices. And economic growth remains weak. What is clear for now, though, is that the received wisdom is wrong. Health care need not eat the world."
/end
Medicare taxes, including the two added Obamacare taxes, are already uncapped. Maybe you mean something different?
Specifically on the 50% of untaxed ..short and long term capital gains, high estate wealth transfers. Mandatory means evennon companies hiding income thru financial engerineerd losses
"50% of untaxed ..short and long term capital gains"
What capital gains are not taxed? Selling one's primary residence can give a nice exclusion (I think is up to 500K for married), but so much home appreciate is just inflation, so you are otherwise taxed on inflation.
I'm no fan of EA, nor an AI-doomer, but your comment on OpenAI seems mistaken to me.
"And “shut it all down” is what the OpenAI board seems to have had in mind when it pushed the panic button and kicked Altman out. But the effort collapsed when OpenAI’s workers and financial backers all insisted on Altman’s return. Becuase they all realized that “shut it all down” has no exit strategy."
It seems plain that "shut it all down" is not what the board "had in mind". If so, then OpenAI would not exist as it currently does.
I would suggest, instead, that the board believed in the principles of the founding of the corporation; that is: "The Company exists to advance OpenAI, Inc.'s mission of ensuring that safe artifical general intelligence is developed an benefits all of humanity." That is, not necessarily to grow or to make a profit (explicitly, regarding the latter). And that they realized (too late, apparently) that running the company as an SV startup ("grow!", "move fast!", "break things!", "get rich!") was not conducive to the goals of the foundation. And likely (also too late, apparently) realized that coupling themselves too tightly to Microsoft was a similar mistake.
Now Microsoft is in the driver's seat, and I don't see the new board as having much focus on the founding principles.
Are people really doubtful about increased disaffection among the US working class? And the supposed European counterexamples aren't too convincing as they're less economically cutthroat than the US--if anything wouldn't that strengthen the argument that there's some causal nexus between US capitalistic culture/socioeconomics and the mounting disaffection? Even if we set aside particular health outcomes or dispute vocabulary like despair and stress, it still seems like there's some pernicious social phenomenon that merits scrutiny.
I feel like there's increased disaffection among all classes.
If I had to blame something I'd blame the internet, for constantly shoving faraway problems into people's faces.
Combining 2 thoughts from this post: income inequality and deaths of despair. Sociologists who study these topics have data that clearly show the relationship between impoverishment (the bottom end of inequality) and mortality/health (shortened life expectancy, including "deaths of despair" and reduced adult height). Please see Dr. Peter Turchin's data at peterturchin.com. In terms of cause and effect, evidence-based sociologists clearly show a chain of cause/effect in many societies worldwide: inequality > impoverishment > rising fascism, because the impoverished look to a strong leader to destroy the system for them. Noah, please see Dr. Turchin's data and provide your reaction.
I have followed Turchin's writing for years. As for your comment, inequality and impoverishment are not the same things. It is one thing to be against impoverishment, but shifting this to an argument against inequality requires some huge assumptions, many of which don’t work. Entrepreneurs and surgeons getting paid more to enrich or save more lives does nothing to affect my height, nor cause me to drink more or become a junky. Indeed, they are likely to improve my life.
I don’t doubt that there could be a link between between "Unfair Inequality" (note the first word does most of the work here) and totalitarian support (fascism, socialism, etc). But note too that dysfunctional incentives and welfare traps can do the same types of things:
— encouraging uneducated non English speakers to come to the country
— encouraging clannish-based cultures to immigrate to liberal societies
— paying mothers not to get married or work
— subsidizing young men to get useless degrees
— paying welfare to people to stay in places with no jobs
"Inequality" is one of the most disingenuous terms in usage today. Those doing the statistics wildly distort it and include or exclude things for the sake of their narrative. They pretend that inequality is synonymous with unfairness or guaranteed to lead to fascism.
The last part of this sentence is an oversimplification.
Fascism scholar Roger Griffin describes fascism as palingenetic (rebirth) ultranationalism. Fascism is a mass politics, and there isn't a single class that acts as a kernel of political potential for fascism to emerge.
So poors aren't the kernel to mobilize fascism, even though they can support fascism for those ends. They are also supporting the same party as the beleaguered merchant and financial class fearing a challenge from below, military brass who yearn for a martialized culture -- the Nazis didn't invent the image of the German as an officious, disciplined brute; the militarization of culture was formed and perfected by von Bismarck -- religious leaders who fear and detest a modernity they can't control; even the bohemians like poets, artists and musicians who see the world in idealized, utopian terms, etc.
Fascism was never about coherence, but rather an eclecticism which makes it a formless ideology that allows anyone to project their perceptions onto it and see what they want.
Totally agree with everything you wrote. I phrased it as I did for simplicity and impact. The actual mechanics of a rise of a counter-governmental movement (call it what you will; I called it fascism) must include disaffected elites (or those who aspire to elite status) who have the education, social network, and skills in writing & speaking to lead the masses in their direction. The masses must be disaffected and ready to be mobilized. That's where the US is today, with a high "mass mobilization potential," the source of which is 2 generations of impoverishment due to income inequality. When a counter-governmental elite harnesses their discontent, the US will be in real trouble. (Trump may or may not be it.) See the work of Dr. Peter Turchin (peterturchin.com) and his fellow researchers for the evidence (data and models) that supports this common phenomenon.
Income inequality is not the driving force for U.S. fascism. Racial ressentiment and status anxiety among upper-income strata Whites are the salient factors.
Trump's base is the sore winner.
Ta-Nehisi Coates foretold what Trumpism was going to be in 2017 in his Atlantic essay "The First White President." He ran the numbers on who Trump voters really were by income, education level and geographic area. Trump won Whites as a race, period, and his voters were no different than previous Republican candidates.
As a chaser to the shot of "The First White President," Robert Pape did a demographic profile of the Jan. 6 insurrectionists. The modal insurrectionist is a status-anxious asshole boss. Again, sore winner.
I'm familiar with the work of Peter Turchin and his hypothesis of elite overproduction. The trouble is, if you advance his idea of disaffected elites (in this case, more postgraduate degrees than positions of political power or economic stations that could absorb them), they are pulling the country leftward and are a countervailing power against fascism. Turchin is also the one who says the French Revolution was a backlash by Jan Bradys (the middle and younger children of the aristocracy glommed onto liberté, égalité, fraternité as a way to topple primogeniture, or power accruing to their firstborn brothers).
My belief in anyone's theory as to how anything operates is their evidence related to cause and effect. Turchin has the most credible data on social upheaval. I agree that disaffected elites are pulling the country leftward, but those elites are not speaking to the disaffected masses. The disaffected will respond to someone from the right who speaks in terms that matter to them: political violence against the government.
Turchin has also boasted that he has 10,000 years of data. That's a tell as bright as a Vegas Strip marquee.
Piketty: “inequality denial is as dangerous as climate denial.”
Prices in a properly functioning market act as signals and incentives. That means that in a properly functioning market it is entirely possible that inequality in wages or differences between capital and labor are the signals telling us to invest more in skills and investment and entrepreneurial ideas and less in rote labor. The incentive is to get a better education, move to better jobs, or start your own lawn-care business.
If Piketty or anyone else thinks that rising inequality is de facto a negative, then I would argue they are missing a fundamental grasp of economics. That said, I understand that inequality could be a sign of rent seeking or market distortions, and that rising inequality can create various externalities.
Adding insult to injury, when I see economists decry inequality, top on their list of remedies is usually more powerful unions (which are by definition market distorting, rent-seeking labor cartels).
Prices do two things. They give incentives and they transfer income. You may approve of the incentive giving without approving of the income distributed as result.
What do you base your disapproval on? I assume you don’t disapprove of brain surgeons making more than janitors? When is unequal unfair in your opinion, and when should it override price signals?
The whole issue is driven by social politics not economic realism.
I saw that in the aftermath of the OpenAI coup, they added Larry Summers to the board and that only added to my existential dread.
Will we all feel better about AI running amok because Larry Summers is there to write a reassuring op-ed about how the AI’s decision to liquidate 70% of us into Soylent is going to be tremendous for the stock market and economy?
No.
And that’s basically all he’s ever been good for. Larry Summers cares so much more about money as a line going up, than he does about people or humanity as a whole.
re: " narratives; Piketty thundered on Twitter that “inequality denial is as dangerous as climate denial.” (Note: No, it is not.)"
Just an ignorant dilatant, but IIRC Piketty argues that inequality distorts economic/public policy decisions to favor the wealthy. The Fossilistas are a prime example.
Noah: I think one of the strongest arguments in favor of renewables at this point is the very long term problem of heat waste. I'd imagine you've read about this phenomenon: many decades from now, global warming will re-emerge as a huge problem even *after* we've completely decarbonized simply because, when humans employ energy to do things, some of it dissipates into the atmosphere and increases temperatures.
https://aeon.co/essays/theres-a-deeper-problem-hiding-beneath-global-warming
Anyway, as I understand it, the only energy technologies that can completely avoid this (very serious) issue, long term, are solar and wind, provided the systems are properly designed. Nuclear doesn't cut it. It's not a problem that will become huge in scale for a very long time (maybe several centuries) but it's a very real issue.
The article says that waste heat will pose a problem in about 200 years, but not really today. We replace our electricity generating infrastructure about once every 50 years. So we have 4 technology generations to figure this problem out. Also, waste heat goes away the same day you shut down the power plant, while CO2 keeps on warming indefinitely.
It’s an interesting problem to think about. And I’m in favor of any argument to get people to build more solar. But I wouldn’t consider this a serious reason to not build nuclear power today (if building nuclear was seriously on the table). Just like how natural gas power plants are complementary to renewables and cleaner than coal, we shouldn’t let perfect be the enemy of good.
It says it will be a problem in 200 years _if_ we keep doubling our energy usage every 30-50 years.... or in other words, if we increase our energy consumption by a factor of 100 or so.
But of course it's very unlikely that we could get 100x our current energy consumption from (terrestrial) solar anyway, even if we could think of a good way to use 100x more energy. It would require tiling too much of the Earth's surface with solar panels.
The problem with heat is it can’t radiate in to space due to the heat absorbing chemicals being released in to the atmosphere. This is where nuclear IS the answer. It’s energy dense, efficient with minimal waste and doesn’t release heat trapping chemicals into the atmosphere. Also consider that solar panels may not emit heat but the things it powers, like generators, lite bulbs, electric heaters transmission loss etc, does. Also basic math says that solar and wind can’t produce the amount of energy needed to satisfy todays needs much less increasing future needs as Germany is finding out the hard way. Climate change is a problem but it won’t be solved by ridge green ideology and wishful thinking. Energy use is directly tied to prosperity and abundance. Energy security is directly tied to national security. I seriously doubt the average citizen and their leaders would be willing to accept less of either. Doomberg has a great series of articles in Substack on this which I highly recommend.
We want less net CO2, not less energy. True we should be developing nuclear and geothermal, too. The more different kinds of zero-CO2 technologies we have the more lower the average cost as some technologies are lese expensive for different uses.
> consider that solar panels may not emit heat but the things it powers… does
This is the insight of the article - renewables capture energy that would have turned into waste heat if we didn’t capture it. They actually cool the earth a little bit during the capture process and heat it back up when the energy is used, coming out close to neutral.
> Basic math says that solar and wind can’t produce the amount of energy needed to satisfy todays needs
The basic math works out quite fine, we would only need to cover a fraction of the world’s deserts in solar panels to power all of humanity. It’s the details that are hard - transmission, storage, minerals, permitting, etc.
As you so eloquently pointed out with your details, the math doesn’t work out. The amount of engineering, current battery technology, transmission loss trying to get the energy from deserts to places where most people live, the low energy density which is further impacted by darkness and clouds not to mention the ecological impact makes this impractical.
But we can just use cheap energy to remove more CO2 from the air.
It would make for an interesting paper. How much DAC would you need to offset waste heat and what would that entail in terms of energy and materials? Surely the DAC machines would also create waste heat…
So define "cheap energy". It might be cheap for generating entity but not for the user, as we are now seeing in CA where the costs of renewables are ever increasing and severely affecting especially lower income people.
Heat is wasted also when electricity (regardless of how it is generated) is converted to mechanical energy (in EVs, heat pumps, motors) or computation (AI, data centers, crypto) or heating. Nuclear electricity generation is engineered to not waste any heat and convert as much as possible to electricity. Wind turbines also waste heat converting rotational energy to electricity and solar panels reflect back IR as waste heat. Waste heat is not going to be a problem until long after we will have technology to capture it, and it seems like you may just be inventing new reasons to be anti-nuclear.
Effective altruism is like macroeconomics: a good intuition pump that some people mistake for all-inclusive blueprints.
I'm very glad the people working on AI are keeping an eye out for the wackier possibilities. The actual data doesn't so far seem to justify either hosannas or panic, but it's good the actual professionals take their own work seriously.
The scary thing is that there is no financial incentive to look out for the wackier possibilities.
The even scarier thing is the plenitude of countervailing incentives--financial and otherwise--for people to ignore the wackier possibilities, or even to disparage those who warn about them.
EA seems to have found a good one; just tell rich people they're going to live forever and then scare them into worrying about far-off theoretical problems.
Actually, that's a big open question in alignment research: how much does "commercially reliable" overlap with "safe for humanity"?
Suppose you make a financial-trader AI. If that AI lies to managers about the expected profits, or even just flatters money managers instead of being honest, it's a bad product that investment banks and hedge funds won't feel safe purchasing.
What does that mean? It means that even if AI makers only care about profit, they still want AI that won't lie to or flatter users in critical applications. We don't need extra regulations for companies to fear deceptive or sycophantic AI. Deceptive and sycophantic AIs are, in many applications, a bad commercial product that will cost you sales and reputation.
On the other hand, lying about near-term consequences and the distant future are two different kind of lies. You can imagine an AI that's very honest about its financial advice, because it knows it'll be caught, but sneakily running its own programs in the background, because it knows those won't be found out until it's too late.
So we have the key question: is "don't lie, it's bad for our business" enough the same as "don't lie, it's bad to wipe out humanity"? When AI makers train the AI to be short-term honest with users, will that be enough to also get us long-term honesty and safety?
There's often a lot of uncalled-for anthropomorphization here.
Being able to make the decision to lie, having any self-reflection ability, having any long-term memory ability, and being able to trade stocks are all completely different skills. There's no reason to think any system would come with any particular combination of these skills just because it's been called an AI. And what we currently call AI (LLMs) doesn't have any of them.
How do you explain fantasy answers with present AI even when there is actual information available that AI could offer the customer?
Basically there's two kinds of issues:
1. the model actually learned wrong facts, or else knows something "imprecisely" (imagine the right answer gets "rounded off" because the model is too small).
2. Answer generation is done via a sampler algorithm that only goes forward and isn't itself "intelligent". So if you ask a question it doesn't know, and then the first word it generates is "yes", it sort of forces the model into making up an answer because it can't start over again.
There's a lot of research into improving this, but some of the issues are fundamental.
Making up an answer sounds a lot like human CYA (cover your ass) behavior, so it tempts us into an anthropomorphic assessment.
It seems like Piketty’s morally tinged outrage about “inequality denial” rests on the assumption that the outlandish wealth accumulation of the underserving rich somehow impairs the economic welfare of the lower classes and the poor. How true is that?
I mean, if you take consumption as a more direct measure of standard of living, in the US inflation-adjusted personal consumption expenditures per household continues to rise decade over decade.
Which is not to argue that there isn’t some low hanging fruit to improve people’s economic security, like universal health care coverage.
Rich people hoarding money is bad for poor people if the world is zero-sum. And unfortunately zero-sum thinking was a self-fulfilling prophecy in the 2010s. Low economic growth encourages zero sum thinking about redistribution, and economic growth was low because of a fixation on balancing the budget. And people had a fixation on balancing the budget because they were in a zero sum mindset.
"economic growth was low because of a fixation on balancing the budget"
Economic growth 2008-2020 was low because of bad monetary policy, specifically, under-target inflation. Balancing the budget (at least to extent of financing all non-high yield investment from taxes) would be pro growth.
Great stuff re: EAs, despair, and coal. But solar and land use:
Jesse Jenkins: ...[In] The most cost-effective of our net-zero scenarios, [wind] spans an area that is equal to Illinois, Indiana, Ohio, Kentucky, and Tennessee put together. And the solar farms are an area the size of Connecticut, Rhode Island, and Massachusetts.
Ezra Klein: Holy crap.
Jesse Jenkins: So these are big, big areas.
https://www.nytimes.com/2022/09/20/opinion/ezra-klein-podcast-jesse-jenkins.html?showTranscript=1
(Not defending fossil fuels, just saying we need nuclear power to give everyone lots of carbon-free energy.)
And geothermal. However efficiency of solar can still go up.
Warning: On topic but also on the nose ... Can AI ever advance to the point where it can put Illinois, Indiana, Ohio, Kentucky, and Tennessee to their highest use and turn everyone in them into wind turbines? 🤔
Can you make wind turbines out of paperclips? If so, then "yes"