This is kaleberg's Typepad Profile.
Join Typepad and start following kaleberg's activity
Join Now!
Already a member? Sign In
kaleberg
Recent Activity
If slavery had ended in 1783, it would have made little difference with regards to US industrialization. Cotton would still have been planted and grown. In fact, it is more likely that the South would have invested in cotton processing sooner and more intensely to keep more of the profits local. Why export cotton when you can export cloth or garments and get a higher margin while exploiting a local advantage? I doubt slave owners would have invested all that much in manufacturing. It would have been more profitable to invest in more slaves. Why buy a share in factory when you could buy a share in a slave raiding expedition? Why buy a share in a railroad when you could buy slaves, an asset that compounded each generation? If there had been a more profitable place to invest, then slave owners would have invested in it. They might have been evil, but they were also greedy. You would think a factory full of slaves would have been a good bet. As for other slave owner spending, most of it was for prestige goods, often European imports. Poor whites made a lousy market and slaves weren't a market at all. While the North was creating hell-hole jobs in the factories, it was also creating a working class and middle class market for its own goods. The South was just creating a hell-hole.
1 reply
The big change was television. If you look at the time use studies, it was television that massively sopped up hours spent socializing starting in the 1960s. Social media has just taken time away from television viewing. Whether this is a good thing or not is another issue.
1 reply
It's also difficult because there are benefits and certainties of being able to farm even a small plot of land. The goal may be more efficient agriculture, but the displaced small farmers are the ones expected to bear the burden. The usual practice, as history shows, is that the politically connected simply take the land for a token payment and create a marginalized population of refugees. Some would be hired to work as peons, possibly on their own land, but obviously not all of them. A more just approach would be for the displaced to become shareholders in the new agricultural firm. They could get dividends in cash and perhaps a farm-share box. Of course, the capitalist will take the farm into bankruptcy and pauperize everyone shortly thereafter, but it's an idea. I wouldn't be surprised if Haiti does modernize its farms. There used to be very large and efficient plantations there before the revolution, so there is precedent. Whether this leads to the same result as in the 18th century remains to be seen.
1 reply
We've seen the decreased value added. The mechanism is simple. Employers replace higher priced labor with lower priced labor. Instead of a skilled worker making a decent salary, they hire an unskilled worker making minimum wage. That leaves employment unchanged, but channels money from labor to capital. Jobs get deskilled. Wages fall. People lose better paying jobs and are stuck having to take lower wage jobs. Even education has been devalued as jobs requiring a high school education, and had once required a grade school education, now require a college education. Only now are we seeing wholesale job destruction, but surely you have noticed that the recovery from each recession since the 1970s has been weaker and taken longer. We had the keys to the paperless office in the 1980s. Only now is paper use declining. That's about par. It takes 30 to 40 years for a technology to show its real impact. (That was the case for steam engines, automobile engines, electric motors, and now computers. Think of it as the 40 years Moses spent wandering in the desert, but never getting to figure out how to use a smart phone.)
1 reply
Heritability research is all sorts of useful in health management and medical decision making. There are a lot of cases where knowing genetic history can cut the number of false positives and false negatives. Knowing my family history makes a big difference when interpreting the results of a colonoscopy or fasting glucose test. Heritability research is also useful in learning about the human genome in general. Some families have language problems, some are prone to various diseases, some are more likely to gain weight. Knowing that these are heritable implies a possible genetic cause and that might lead to better general understanding and possible treatment. The eyeglass example works because testing for vision problems is relatively cheap and the risks of false positives and negatives are minimal. The typical test is quick and inexpensive, so it can be used repeatedly. Glasses, when not needed, are easy to remove. When the testing is expensive and the risks of misdiagnosis are high, understanding the genetics becomes more important. There may even be social policy benefits of understanding the genetic basis of traits like intelligence. For example, if there are genes for intelligence, there is a good chance they are more broadly distributed across different groups than generally acknowledged. Genetic testing could be used to include as it is often used to exclude.
1 reply
This is pretty hilarious, but aren't they Oxonians? One of the things I like about the Romans in their writing that has survived is that they seem remarkably contemporary. If you swooped up some guy or gal off a hill in Rome and dumped him or her down in a modern city miraculously knowing the local language, they would be surprisingly at home. They'd recognize the status competitions, the unspoken games, the relationships, the businesses and so on. They are more like us than most Europeans from the Middle Ages.
1 reply
There are two left-right axes at play, the social one and the economic one. There isn't much of an economic left at the moment.
1 reply
It was the incomprehensible optimism from 2009 to 2015. Could it have been the audacity of hope?
1 reply
It helps to remember the impact of automation and redesign. Automobiles used to contain 60-80 hours of labor. Now its down to 20-30 and dropping into the 10-20 range. Sure, auto workers could find new jobs, but not manufacturing jobs and definitely not well paying manufacturing jobs. Our current policy is to subsidize lots of poorly paying jobs, that is, jobs pay so little that the government has to make up the difference to provide things like food, housing and medical care. We aren't going to reverse the decline of manufacturing, and we shouldn't want to, but we have to come up with some way of providing jobs that can at least pay for the basics.
1 reply
The spirit is willing, but the flesh is weak.
1 reply
dr2chase sounds quite sensible. If you are going to have to deal with Twitter, you might as well use the Twitter equivalent of a spam filter.
1 reply
The retail apocalypse is more about private equity buying up retail chains and loading them with debt than customers fleeing to Amazon. Retail chains were doing quite well, and they had good cash flow. That made them attractive targets for value extraction. Retailer after retailer has been purchased and the buyouts or M&A debt are now on their books. Their credit lines were maxed out to pay off the investors. All it took was a slight downturn to drive them into bankruptcy. Now, if the business press wasn't so credulous, this might have made the headlines. The problem isn't retail. It is private equity and our financial sector. It is the philosophy of value extraction in which the goal is liquidation, not future cash flow. The good news is that people will still need retail. Online shopping will continue to grow, but I expect the demise of the great chains as opportunities for smaller businesses to grow. Payless may have vanished, but people still need shoes, and they want shoes they can try on to make sure they fit. They want toy stores where they can walk the aisles and see their children's reactions. My guess is that this will make the headlines in the early to mid-2020s. (Call this 2020 foresight.)
1 reply
I've always been rather dubious of the criticisms of the Roman ruling class based on their sexual or other libertine practices. I'm sure that plenty of them were scum, but that sort of accusation was so common as to be meaningless. Look at the accusations against the sexual practices of the 18th century French court. There was a whole pornographic genre. The real critique was political and economic, but sexual deviance sold better. (Yeah, there was cannibalism too. Look at the Cataline Conspiracy, but really? Were all those early Christians cannibals too? As one historian asked, if they were all having sex with their brothers or sisters, could an only child become a proper Christian?) I think Handley has the colonization backwards. It is the coastal and urbanelites who are hostage to the rump of the Confederacy, the neoliberals, and the other right wing scum who have the ear of our "mainstream" politicians and our wealthy, conservative elites. Why else do we have policies with 70% plus nationwide approval being considered outlandish and impossible? Look at abortion rights, look at more progressive taxes, look at gun control, look at single payer healthcare, and so on. Those are all highly popular positions, but the neglected midwest and poor south, the colonizers, have somehow managed to thwart them. It's like the land grant colleges, the railroads, opening the west to colonization and a host of other popular government initiatives that only became possible when the colonizers seceded from the Union. Yes, our nation is run by our colonial masters, but they aren't the people one might think.
1 reply
Labor only gets improved conditions when the elite has its back to the wall. That usually means a famine, a war or a plague. The shooting wars of the 20th century, and even the Cold War worked out pretty well for labor, but when the Cold War ended, there was no more need to offer more than crumbs.
1 reply
ML algorithms are indeed algorithms, but they're simple curve fitting based algorithms. You dump in a training set and do a form of regression and get some parameters that will give you the desired results for your input. If you are reading this blog, you've probably done something just like this in Excel or, perhaps, a more advanced tool. You also know that you've gotten a provisional solution useful for estimating, but with limited value when extrapolating or forecasting. ML curve fits have the same problems, except they are often less robust. They have a topology in a much higher dimension than a simple curve fit and a much more complex set of parameters. It's easy enough to develop some intuition about y=mx+b as m and b vary, but nearly impossible to develop any intuition about an ML curve fit. Try doing an order N polynomial curve fit on your next dataset with N observations. It will match the input cases extremely well, but if you have ever tried to understand an order N polynomial where N is larger than perhaps three, you will realize that you have an answer, but no idea of how useful it is. I can understand why some might not call this kind of solution an algorithm. All the real work is done analyzing the training set to create the curve fit parameters. The actual extrapolation is just simple arithmetic albeit a lot of it. Worse, neither the curve fitting nor the extrapolation seem to have anything to do with the problem domain. I'll argue that it's still an algorithm, but subject to same problem that all such algorithms have. Anyone who has tried to beat the market using an algorithm based on retrospective analysis knows how poorly this works. It's easy to use 20 or 50 years of financial data and create a trading algorithm that produces magnificent results by the magic of 20-20 hindsight. The odds of it producing good results even a few months into the future are minimal. Yes, ML is an algorithm, but we have yet to plumb its limits.
1 reply
Electrical engineering attracted its oddballs. You can see why Tesla couldn't last long on any job or satisfy a backer for very long. He had some great ideas, but he couldn't work with anyone. For all of Edison's flaws, he invented the modern industrial research laboratory and the electric power industry. Tesla could put on a good show, but he was never concerned with the next step. If he had been developing software, he would have released amazing prototypes but never a version one. Nicholas Negroponte called Buckminster Fuller a poet. He expressed ideas about the machine age. Nikolas Tesla was a poet of the era as well. My favorite electrical oddball is Heaviside who produced the modern form of Maxwell's equations, developed much of modern vector calculus, discovered the Heaviside layer of the ionosphere, made the trans-Atlantic cable reliable and put electricity and radio on a sound theoretical footing. He was brilliant, but he never had a degree. He published most of his work in an amateur telegraphy magazine and only formally published when he realized that he wasn't being taken seriously. He may have been the first to have produced the famous sqrt(1-(v^2/c^2)) term in analysis of the electric field, a term central to special relativity. He grew up in Dickensian poverty, in Dicken's old neighborhood, and spent most of his life at the edge of poverty.
1 reply
Was being a woman as bad as being a child? An awful lot of them died too, and not all thanks to political machinations.
1 reply
Did I read that right: "Everyone who opened the first box ended up with 1,000,000, while everyone who opened both boxes ended up with only 1,000. Knowing [all of] this, what do you do?" That's a simple enough game, assuming I would rather have 1,000,000 than 1,000 or vice versa. I assume all that other stuff, including The Predictor, is just stage magic window dressing, sort of like "where do you bury the survivors?" The Dr. Evil conundrum also struck me as rather silly. Dr. Evil knows that there is at least one version of himself capable of destroying the world and getting away with it. If destroying the world is worth a slow painful death, then the solution is obvious. The problem wouldn't really get that far since Dr. Evil still has access to his laboratory and equipment capable of destroying the world. Given that ultimatum, he would rig a deadly suicide bomb to detonate if anyone tried to get at him, then destroy the world. Dr. Evil is evil, not stupid, so he would notice any flaws in his cloned laboratory rather quickly and recognize that it had been cloned and, therefore, that he had been cloned. Even if Dr. Evil's laboratory has been properly cloned, there will still be artifacts. I'm sure that Dr. Evil could figure out the situation by measuring telomeres, timing radioactive decay, performing Bell inequality experiments or doing their super-science equivalents. If he is running in an emulator, there will be artifacts. I'm sure Dr. Evil can come up with a proper analog to Rowhammer or Spectre and track timing and energy discrepancies that would give clues about the nature of the simulator if one is present. https://www.jwz.org/blog/2018/02/case-nightmare-buttcoin/
1 reply
I couldn't follow that. P(t+1) is population density at time t+1. P(t) is population at time t. How does that work? Is it really sensible to have a variable P mean two different things depending on the time? If the model involves both population and population density, shouldn't there be some kind of accounting for the area? Is P* the fixed point population or the fixed point population density? Perhaps one is P* and the other p*? Both were mentioned. Why is the steady state interesting in a dynamic model? I'm lost. So, I decided to look at the original paper. The equations there make a bit more sense. At least population and population density are considered two separate things. The birth / death equation seems designed for a steady state since it assumes that population change is about food surplus or deficit with no regard for the number of women of childbearing age. It also makes some rather unrealistic assumptions about farm productivity, land and labor. Basically it assumes that more people means higher agricultural output with no saturation or falling margin. It's an unrealistic model, but so is F=ma. Is it a useful model for thinking about national wealth, population density and farm productivity? Even the original paper has its doubts about this. It links wealth to urbanization and the related job specialization much more than to population density. It isn't the mathematics that is misleading. It is the glib "Despite these reservations regarding population density, we present results using population density, as well as urbanization, as a proxy for income per capita." Huh? Well, "First, population density data are more extensive." I'm lost again. I think Campbell might have a point.
1 reply
The hilarious thing is that people have been arbitraging discrimination for ages. If you can assign a job a lower social status, you can fill it with lower ranked people for less money and get the same or more work out of them. It's almost the founding principle of the industrial revolution labor market.
1 reply
This looks like a great resource.
1 reply
The problem is that Republican stimulus is not as good for the economy as Democratic stimulus. The Reagan military buildup may have propped up the Soviet Union for three to five additional years, but had surprisingly little effect on the economy. G.W.Bush spent a fortune in Iraq, and it isn't clear that our economy is stronger for it. Republicans are willing to borrow and spend, but they loathe the idea of any of the money going to anyone but a multi-billionaire or professional grifter.
1 reply
You can look at modern globalization two ways. In one view, it turned a lot of impoverished peons into poorly paid workers. In the other view, it turned a lot of poorly paid workers into impoverished peons. It's like a semiconductor. One has to look at the movement of both the electrons and the holes. Before modern globalization, much of the undeveloped world consisted of poor peasants on small farms just getting by or poor workers with marginal jobs that just kept them alive. In contrast, low end workers in countries like the US had access to a certain surplus that gave them better food, better housing, most likely an automobile and better prospects for their children. Now, much of what was once the undeveloped world consists of low end workers with better food, better housing, quite likely a moped or inexpensive car and better prospects for their children. Low end workers in the US, in contrast, now have marginal jobs that just keep them alive and lousy prospects for their children. It was like one of those quantum exchanges where a quantum state here suddenly finds itself teleported somewhere else, though only the state, not the particle was relocated. The world still has what it once had, just in different places. A lot of the arguments against globalization are arguments that we might now have cheaper tee shirts, though this is only because work was transferred rather than automated, but we now have US citizens living in what were once unacceptable third world conditions. If you are one of those particles whose not so bad quantum state was teleported, you might be feeling angry about it.
1 reply