Astral Codex Ten Podcast

https://astralcodexten.substack.com/p/mantic-monday-grading-my-trump-predictions I had many opinions on Donald Trump. I tried to back some of those opinions up with predictions about what would happen during his administration. Now that the dust has cleared, it's time to see how I did. The summary: Of 48 specific predictions about Trump, I got 37 directionally right, although this is kind of meaningless. I got an average log error score of -0.48 (where getting everything right is 0 and guessing 50-50 for everything is -0.69) although this is also kind of meaningless. I quadrupled my money on prediction markets, which I think is meaningful. In terms of my more qualitative/implied predictions, got at least one important trend right before anyone else, but also made some embarrassing unforced errors. Going through all my predictions post by post, and giving each a letter grade: 1: 10-23-2015: Trump's base is/will be surprisingly racially diverse (A-) As far as I know, the first post I wrote about Trump was this one, where I argued against the prevailing narrative that Trump was practicing "the politics of white insecurity" or had an unusually white base of support (for a Republican). I wrote that Trump seemed to be doing pretty well (for a Republican) among blacks and Hispanics, and concluded that: There are too few data to say anything for sure. But all of the data that exist suggest that if the Republican primary were held today and restricted to non-whites, Trump would still win. And if Trump were the Republican nominee, he could probably count on equal or greater support from minorities as Romney or McCain before him. In other words, the media narrative that Trump is doing some kind of special appeal-to-white-voters voodoo is unsupported by any polling data. I was right. In the general election a year later, Trump did better than Romney had among non-white voters. He made large gains among blacks, Asians, and Latinos. The only ethnic group where he didn't gain at least five percentage points over Romney's numbers was whites. As I pointed out at the time, the narrative that Trump was especially appealing to white voters was bizarre and not truth-based, motivated primarily by a demand for racist Republicans on the part of increasingly woke narrative-consumers.

https://astralcodexten.substack.com/p/your-book-review-progress-and-poverty [This is the third of many finalists in the book review contest. It's not by me - it's by an ACX reader who will remain anonymous until after voting is done, to prevent their identity from influencing your decisions. I'll be posting about two of these a week for several months. When you've read all of them, I'll ask you to vote for your favorite, so remember which ones you liked. - SA] In 1879, a man asked "How come all this new economic development and industrialized technology hasn't eliminated poverty and oppression?" That man was Henry George, his answer came in the form of a book called Progress & Poverty, and this is a review of that book. Henry George is variously known for leading an early movement that popularized Universal Basic Income, sporting a fancy beard while shouting "The Rent Is Too Damn High!" and inspiring a popular board game that was shamelessly ripped off and repackaged as Monopoly. But he didn't just write a book. He also ran for Mayor of New York city in 1886, beating out some rando Republican named "Theodore Roosevelt," but ultimately losing to the favored candidate of Tammany Hall, who saw George's radical economic ideas as a threat to their well-oiled political machine (Andrew Yang take note). He ran again in 1897 but died just 4 days before the election, prompting a national outpouring of mourning. According to Ralph Gabriel's Course of American Democratic Thought, in New York alone 200,000 people came to see his body lying in repose, half of which had to be turned away. For context, that one crowd was roughly the size of 1% of the entire population of New York at the time. I'm writing this book review for three reasons: George's arguments about land, labor, and capital present a fresh alternative to conventional ideas about "Capitalism" and "Socialism" (and whatever we mean by those on any given day) The book has timeless advice for navigating modern crises such as ever-rising rents, homelessness, and the NIMBY vs. YIMBY wars. This is a golden opportunity to shamelessly over-use the catchy phrase "By George!"

https://astralcodexten.substack.com/p/prospectus-on-prospera Who among us hasn't looked out at the great edifice of human civilization in all its complexity, and thought "Yeah okay but I could do it better"? Centuries of utopian communes, micronations, and seasteads have dreamed of rebuilding society from first principles, free from entrenched interests and the debris of the past. If you got all the laws and values just right, maybe you could prevent poverty and corruption from finding their first footholds. Do the "liberty and justice for all" thing, but for real. And who among us, having had the dream, hasn't entered into multi-year negotiations with the government of Honduras? Taken advantage of a clause in the Honduras-Kuwait Treaty Of Reciprocal Investment guaranteeing them their right to pursue their vision unmolested? Raised millions in venture capital and bought land on a Caribbean island to turn it into a reality? Not Erick Brimen, and not Honduras Próspera Inc. You might have read about them last month in Bloomberg: A Private Tech City Opens For Business In Honduras. Or in NACLA: A Private Government In Honduras Moves Forward. Or FT: An Investor's Prosperity Vision For Honduras. I read all of this and still didn't feel like I quite understood what was going on. Then a fortuitious mistake led me to an email exchange with Trey Goff, Próspera's extremely open and thorough Chief of Staff, who kindly let me grill him on all the stuff I didn't understand. The result is this post. It's all the information I could collect on Próspera from basically every public source, plus some non-public ones. It's about a private tech city and a prosperity vision and all that. But it's also about - - - well, people talk a lot these days about "systemic change". But usually that means something like fiddling with tax rates or ending the filibuster. What if you could actually change the system? Say "this system we have, the one that's letting all these people starve and suffer violence and die of preventable diseases - I don't care for it. Let's try something else"? Yes, this is about startup governments and investment opportunities and blah blah blah, but it's also about trying to fight global poverty by radically changing the rules of the game that makes it possible.

https://astralcodexten.substack.com/p/link-unifying-predictive-coding-with [epistemic status: I know a little about the predictive coding side of this, but almost nothing about backpropagation or the math behind the unification. I am posting this mostly as a link to people who know more.] This is a link to / ad for a great recent Less Wrong post by lsusr, Predictive Coding Has Been Unified With Backpropagation, itself about a recent paper Predictive Coding Approximates Backprop Along Arbitrary Computation Graphs. Predictive coding is the most plausible current theory of how the brain works. I've written about it elsewhere, especially here.

https://astralcodexten.substack.com/p/links-for-april [Remember, I haven't independently verified each link. On average, commenters will end up spotting evidence that around two or three of the links in each links post are wrong or misleading. I correct these as I see them, and will highlight important corrections later, but I can't guarantee I will have caught them all by the time you read this.] 1: A link between childhood "screen time" and attention problems has - say it with me - failed to replicate. The paper is especially interesting for using a "multiverse analysis": We evaluated 848 models, including logistic regression as per the original paper, plus linear regression and twoforms of propensity score analysis. Only 166 models (19.6%) yielded a statistically significant relationship between early TV exposure and later attention problems, with most of these employing problematic analytic choices. If I had the energy to look through 848 models and see which ones got significant findings and which ones didn't, I bet I would become enlightened by the end of it. 2: Seen on architecture Twitter:

[This is the second of many finalists in the book review contest. It's not by me - it's by an ACX reader who will remain anonymous until after voting is done, to prevent their identity from influencing your decisions. I'll be posting about two of these a week for the next few months. When you've read all of them, I'll ask you to vote for your favorite, so remember which ones you liked. - SA] I. If you're looking for the whipping boy for all of medicine, and most of science, look no further than Galen of Pergamon. As early as 1605, in The Advancement of Learning, Francis Bacon is taking aim at Galen for the "specious causes" that keep us from further advancement in science. He attacks Plato and Aristotle first, of course, but it's pretty interesting to see that Galen is the #3 man on his list after these two heavy-hitters. Centuries went by, but not much changed. Charles Richet, winner of the 1913 Nobel Prize in Physiology or Medicine, said that Galen and "all the physicians who followed [him] during sixteen centuries, describe humours which they had never seen, and which no one will ever see, for they do not exist." Some of the 'humors' exist, he says, like blood and bile. But of the "extraordinary phlegm or pituitary accretion" he says, "where is it? Who will ever see it? Who has ever seen it? What can we say of this fanciful classification of humours into four groups, of which two are absolutely imaginary?" And so on until the present day. In Scott's review of Superforecasting, he quotes Tetlock's comment on Galen:

[This is the first of many finalists in the book review contest. It's not by me - it's by an ACX reader who will remain anonymous until after voting is done, to prevent their identity from influencing your decisions. I'll be posting about two of these a week for the next few months. When you've read all of them, I'll ask you to vote for your favorite, so remember which ones you liked. The broken footnotes in this one are either my fault or Substack's, so please don't hold it against this entry. Oh, and I promise not all of them are this long. - SA] Shasta County Shasta County, northern California, is a rural area home to many cattle ranchers.1 It has an unusual legal feature: its rangeland can be designated as either open or closed. (Most places in the country pick one or the other.) The county board of supervisors has the power to close range, but not to open it. When a range closure petition is circulated, the cattlemen have strong opinions about it. They like their range open. If you ask why, they'll tell you it's because of what happens if a motorist hits one of their herd. In open range, the driver should have been more careful; "the motorist buys the cow". In closed range, the rancher should have been sure to fence his animals in; he compensates the motorist.

https://astralcodexten.substack.com/p/metis-and-bodybuilders Fitness researcher Menno Henselmans writes about optimal program design for bodybuilders. His thesis is that peer-reviewed studies prove bodybuilder lore is wrong in lots of places. For example: Traditional bro wisdom holds short rest periods of 1-3 minutes are optimal for bodybuilding. There never seemed to be much of a formal argument for why other than that people traditionally trained this way. The real reason was probably that bodybuilders chased the pump and burn they get from shorter rest periods. Later the idea of chasing the pump was rationalized into the theory of metabolic stress. Yet there wasn't a single study to support that shorter rest periods actually benefit muscle growth.

https://astralcodexten.substack.com/p/two-unexpected-multiple-hypothesis I. Start with Lior Pachter's Mathematical analysis of "mathematical analysis of a vitamin D COVID-19 trial". The story so far: some people in Cordoba did a randomized controlled trial of Vitamin D for coronavirus. The people who got the Vitamin D seemed to do much better than those who didn't. But there was some controversy over the randomization, which looked like this Remember, we want to randomly create two groups of similar people, then give Vitamin D to one group and see what happens. If the groups are different to start with, then we won't be able to tell if the Vitamin D did anything or if it was just the pre-existing difference. In this case, they checked for fifteen important ways that the groups could be different, and found they were only significantly different on one - blood pressure. Jungreis and Kellis, two scientists who support this study, say that shouldn't bother us too much. They point out that because of multiple testing (we checked fifteen hypotheses), we need a higher significance threshold before we care about significance in any of them, and once we apply this correction, the blood pressure result stops being significant. Pachter challenges their math - but even aside from that, come on! We found that there was actually a big difference between these groups! You can play around with statistics and show that ignoring this difference meets certain formal criteria for statistical good practice. But the difference is still there and it's real. For all we know it could be driving the Vitamin D results.

https://astralcodexten.substack.com/p/2020-predictions-calibration-results At the beginning of every year, I make predictions. At the end of every year, I score them (this year I'm very late). Here are 2014, 2015, 2016, 2017, 2018, and 2019. And here are the predictions I made for 2020. Some predictions are redacted because they involve my private life or the lives of people close to me. Usually I use strikethrough for things that didn't happen, but since Substack doesn't let me strikethrough text or change its color or do anything interesting, I've had to turn the ones that didn't happen into links. Italicized are getting thrown out because they were confusing or conditional on something that didn't happen. I can't decide if they're true or not. All of these judgments were as of December 31 2020, not as of now. (Remember, link means something that didn't happen, not something I was wrong about. We have a debate every year over whether 50% predictions are meaningful in this paradigm; feel free to continue it.)