Basically, evidence of shenanigans of samples or workers at the Wuhan lab would change the conclusion, but China has obstructed all attempts at independent investigations that could find such evidence. I think this is why the lab leak theory is still compelling.
Thanks for those links! A couple of points in no particular order. 1) This redit thread is listed as being 11 months old, whereas Wade's writeup is.. i think from May 5th, so dont expect any direct refutation of Wade in the Redit series.
2) the Redit series actually cites (pretty unconvincingly, IMO) the WIV's history of transparency as proof that it didnt accidentally release COVID:
"And I can hear someone out there, shouting into the darkness… “but it happened in 2004!...It happened in 2008!” (117,118)
Yes, but you know what’s interesting about that: those events are part of why this probably wasn’t a lab accident. We know about those events, because scientists (including some Chinese ones) weren’t interested in covering them up (117,118,119).
Why would they cover it up now? Why would they behave differently than they did in 2004 or 2008? Why would these Chinese scientists reverse course on being honest?"
The Redditor doesnt mention the 1977 Russian H1N1 outbreak which was both most likely an accidental release and covered up by scientists and officials at the time.
I too was unconvinced when I read it a few months ago, on balance of evidence. I don't think he was claiming past transparency proves transparency today, merely that it's evidence in favour of.
Then again, past leaks were relatively minor compared to SARS-COV-2, so perhaps by the time this leak was detected it had already spread around the world. Also, past leaks happened under different government administrators, so perhaps the current administrators had a different view on the PR nightmare this would have caused.
Boy, I wish people would just write essays and not hack them up into tweets. That was too hard to wade through.
But to judge from the first half dozen bites, he seems to be rebutting the theory that it was an engineered biological weapon, which Wade concedes from the start.
Since it's long, let me summarize the points of the virologist Duehr that I found most important, and comment on it:
- The largest part of it argues that the virus has not been artificially created from a known backbone like RaTG-13 (i.e., from a virus whose sequence has been published in a previous publication). For example, all the sections of mutation speed etc. hammer this point in. I think both sides agree that this is convincing. But the argument is still weak; the pro-lab counter is that most virus sequences (including most virus sequences from the Wuhan lab) have not been published, and the argument collapses if the backbone was unpublished.
- The strongest argument contra-lab from Duehr is that "Dr. Shi’s projects almost always provide duplicate samples of every bat to both Chinese scientists and laboratories in other countries". This is a strong case since it would make it difficult to conceal the backbone if the virus really came from the lab. I am not sure whether I should read the "almost all" part of the statement literally. Did they really collaborate with other labs on all their projects? We are talking about 400+ viruses and 1500+ strains. (From the link Sandro posted in this thread.) Still, it makes it less likely that the compromising information from the lab hypothesis would only exist in Wuhan, and I personally find a conspiracy implausible if it goes beyond a single lab.
- Duehr claims that the most plausible origin of the pandemic is outside of Wuhan, possibly far away. I haven't heard this argument in recent discussions, so that claim might just be outdated, but I am not confident.
- Duehr also claims that similar experiments happen all the time, at many places in the world. So having an experiment nearby the origin of the pandemic would not be surprising. Other contra-lab paper have also made this claim. As far as I can tell, this is factually wrong if we define "similar" as "gain of function in vivo in humanized animals". Duehr provides four papers as evidence, claiming that similar experiments were done there. Another contra-lab article provides another four papers. I have checked all eight, and "gain of function in vivo in humanized animals" happens in exactly one of these articles, and this is an earlier paper from Dr. Shi's group in Wuhan. Right now my hypothesis is that Wuhan is/was the only lab in the world to perform such experiments. (If anyone knows better, please tell me. How many labs do you estimate were there?) I may seem picky about the definition, but I think it is important. Removing any of the conditions makes a lab accident much less likely: if it's not in vivo, there are no airborne infectious particles; if it's not in humanized animals, then the virus is optimized for a different species than humans.
Good summary, but I have to take issue with the following of Duehr's claims:
* Duehr claims that the most plausible origin of the pandemic is outside of Wuhan, possibly far away: it's plausible if you already assume a natural origin, which is begging the question.
* Duehr also claims that similar experiments happen all the time, at many places in the world. So having an experiment nearby the origin of the pandemic would not be surprising: err, what? It's not "nearby the origin", it's ground zero as far as we can tell. Big difference. Further, even if these experiments were happening all around the world, if an outbreak begins in the same city as one of these labs, you better believe a lab leak is more plausible than if the outbreak happened further away.
Yes, good objections. Though for the second one, there are two numbers floating around: if I get it right, the lab itself was 14km away from ground zero, and there was a building of the institute only 600m away. But if I get it right, the experiments did not happen in the nearby building, so I would count the lab as "same city" instead of "ground zero". But you are right: even if it was 50 labs worldwide and we count it only as "same city", it would still be remarkable.
It still makes a difference to me whether it is 50 labs or whether Wuhan had literally the only lab in the world where such experiments were done.
I believe it was literally the only lab in China where such experiments were done. Or at least that it was the only lab in China where such experiments were supposed to be done. There are other labs elsewhere in the world, most in countries with a better reputation for enforcing safety regulations. And part of the hypothesis is that the research was being outsourced to China, in which case "...because it was faster and cheaper, because Chinese regulators turn a blind eye to the fact that you're not really doing the tedious BSL-4 stuff" is at least plausible.
Hm, but do you actually know a lab or a research paper? I have heard the claim "similar stuff was done elsewhere" before, but whenever I tried to actually find such a lab, I couldn't. As I said, some people gave references, but the ones that I checked were not "similar". At least not in the sense "gain of function in vivo in humanized animals" that you need to have a really dangerous situation.
Actually, according to Duehr, the virologist community thinks that the Wuhan lab was as safe as labs elsewhere, and I buy his arguments. You may want to read his position on this.
The worst thing in that respect is that the experiments were at least formally allowed in BSL-2 environment (if they were derived from some other bat virus than SARS and MERS). Which had nothing to do with the location, it was made by the funding agency in the US.
I'll try and steelman GoF. (which I think should stop.) By studying how viruses change/ evolve to a new host we learn something that helps block against the next pandemic.
It might. I think the point of research is that you don't always know where it will you lead, if you did then you wouldn't call it 'research'. So you through a lot of shit at the wall and sometimes those recombinations get you very effective treatments like (immunotherapy or something).
Anyway, you say that one particular line of research is verboten or restricted and there would be an effect. I'm not saying it would kill the whole field of virology or anything, but if you look at is say there are 100 possible treatments for the next pandemic. If you institute this ban then maybe that goes to 50 or 20 or whatever 100/N is where N is unknown. It may be a relatively low so you still have dozens of effective therapies to choose from (like us right now)
Or it may be higher in which case you might have just one potential, and then things would be not so great. /Steelman off
I started it but still have too much lingering pandemic brain fog to get deeply into it. I find German-translated-to-English takes a lot of mental power. Just oddly heard about it the other week immediately before seeing your Cassirer tweets.
I very much enjoy your twitter mega threads. I think your experts vs populists points are interesting, although I think the distinction that Robin Hanson made between Experts and Elites (https://www.overcomingbias.com/2021/02/experts-versus-elites.html) may be helpful here too, as it reminds me a lot of what I see coming from the grey tribe critiques of a lot of the pandemic response (Balaji Srinivasan's critiques of bailouts and push for DeFi come to mind).
I'd also really encourage considering structuring paragraphs/sentences that end with "2.0" differently as figuring out where full stops should go added some unnecessary mental load. Maybe replace with "v2" instead?
As the unofficial unofficial associate (maybe) podcast of Astral Codex Ten I'd like to mention the latest episode with this blog's readers' favourite Bret Devereaux (of This Isn't Sparta! fame) in conversation with Professor David Abulafia about the ancient Mediterranean. They talked about Alexandria, exactly how may rowers these insanely huge galleys had and much much more. (I see from Zohar Atkins who just beat me to being first that I am not the only one pushing my own stuff . . .)
Excellent! I love Bret's work and have recently hit a gap in my podcast queue, downloaded. If this is going to be the plug-yourself thread, I may as well link my "rational-adjacent, 2 dudes talking" podcast Affix, latest episode discussing Robin Hanson on Experts vs Elites (which I also linked to Zohar earlier in the thread):
I'm going on vacation for a week and could use book recommendations. I'd prefer some lighter non-fiction; I've been doing dense mathematical reading for the past while, and would like to get away from that for a bit. Ideally I'd like something entertaining, well written, and about a non-mainstream topic (one such book I enjoyed was The Deadliest Enemy by epidemiologist Osterholm and writer Olshaker).
Always tempted to recommend _Passions Within Reason_ by Robert H. Frank, but I noticed that I don't really know what a mainstream non-fiction topic is. Does "economics and evolution brought together to explain several human emotions" qualify as mainstream? If it does, this book won't pass your criteria.
I'm not sure either, but I have studied a bit of econ related to that stuff, so while it might be outside the mainstream, I'm not sure it's far enough out of my familiarity. Thank you though
Oof, this is a tough one. It's almost an axiom that all math textbooks are bad (with Spivak's Calculus being the only exception I can think of). Maybe that's only true of the core authors like Apostol or Stein, but I'm also not sure you're looking for something like that.
Personally, I tend to try to read things outside of my area (or just fluff fiction if I'm being honest), but one book I would strongly recommend is Logicomix, which presents an excellent biography of Russell as well as capturing one of the pivotal struggles in mathematics. As the name suggests, it's a comic book, so it's not a heavy read, but it's definitely worth the time.
"It's almost an axiom that all math textbooks are bad"
Giving you the benefit of the doubt here, I suspect you are intending to say "bad light reading". In which case, that's mostly true.
If you're saying most math textbooks are "bad bad" I strongly disagree and would be happy to recommend several "good good" ones, depending on your interest/mathematical maturity
Um, please don’t wait. I am finally getting back to a place in life where I’m going to be able to take mathematics back up. I want to start back over at Algebra. Working my way towards...whatever it ends up I can do. Lay it on me. All of ‘em.
For you I'd recommend you really hack something like Mystik's recommendation of Spivak or maybe Apostol's calculus, if you have the patience. Once you're into the land of rigor and proofs, try Fraleigh or if you loved calc, look under the hood with an analysis book. Here's a repost with my response from a similar question below,
Here's a few I've enjoyed,
Fraleigh - A First Course in Abstract Algebra
[It's on the gentle side, not dry, and provides a good elementary foundation]
Kaplansky - Set theory and metric spaces
[This book... The first thing you'll note is that it's less than 150 pages long. It's essentially a highlight reel of the subject's foundation it also touches on many naively interesting concepts, like cardinals and ordinals. If you've never done anything with ZFC set theory, this is a fun, quick place to start]
Ebbinghaus - Mathematical Logic
[After going through this book, I'm not convinced there isn't a better treatment of the subject (Maybe some logicians can weigh in). That said, it's where I learned what Gödel was actually saying, not the confused pop notions.
If you have a good grounding in linear algebra already, you might enjoy
Lax - Linear Algebra and It's Applications
[Note, it's not "applied" in the common sense of the word. Also, if you don't have prior exp. the book will be... rough sailing. My honors prof used this—it was my first exposure—he advised the entire class to not blink while reading less we miss something vitally important]
I mean I do suppose it depends on our definition of bad. I meant “bad” in the sense of it would be very difficult to teach yourself from them.
I agree that there are some good texts out there, I mentioned it might just be the core texts because Spivak is the only text I’ve had a professor assign that I considered good to learn from. But this was intentional, many of my professors said something along the lines of “it’s good to have a textbook that is not very clear, because then you have to struggle with the material more”
So in some ways I was being a bit hyperbolic I suppose. In other ways this has legitimately been my experience, so I don’t have many experiences with texts I’d actually recommend
What has been your method of learning mathematics? Constructing lots of examples for each concept independent of the text? Reading papers? My general trajectory has been: Read textbook and not understand much -> Read papers and understand even less -> Suddenly understand mathematical concept in a completely different context -> Thinking that I would have understood this concept much faster had I just focused on reading the textbook.
Thankfully I’ve been able to take a lot of mathematics courses at college and post-grad so when I’m enrolled in a class my trajectory is usually: Listen to the lecture -> work through homework -> work through book on stuff I don’t understand for the HW -> Ask the professor for anything I’m stumped on still
When I’m learning something on my own (not often yet as I’m not quite started on my dissertation), I usually pick a topic, come up with questions about it, research and skim papers until I find something with those answers, read through it and physically rewrite the paper so I force myself to confront the material, and then repeat
If you're offering, I'd like to hear your "good good" math textbooks. Upper undergraduate or graduate level. The last math textbooks I loved are Apostol's two number theory / complex function books. They're perfect for me. I've been struggling trying to find something I enjoy as much as those.
I highly recommend Thurston's 3 manifolds (I mentioned it elsewhere in this discussion). Strogatz's Nonlinear Dynamics is beautiful imo -- I wish I could go back in time and tell my younger self to read it. Arnold's ODE book if you've already taken a class/read a book on ODE. Milnor's lecture notes/books. I learned a lot from Lax's lecture notes on functional analysis. I have his book, but haven't read it except for the historical notes. When I was a student, I enjoyed reading Reed & Simon's treatment of unbounded operators, but I am not sure if I am recommending it or not. I also learned a lot from Hirsch & Smale, Introduction to Diff Eq, Dyn Sys and Linear Algebra (1974). The book has been completely rewritten for the 2nd and 3rd editions (with Devaney, and a new title), and I am completely unfamiliar with the newer incarnation.
In a slightly different direction, and perhaps revisiting the elementary topics, a friend whose recommendations I respect highly likes Gelfand's Linear Algebra. He has also recommended Bamberg & Sternberg "A course in mathematics for students of physics" to me multiple times.
Mathematical tastes differ, so take this with a grain of salt.
[It's on the gentle side, not dry, and provides a good elementary foundation]
Kaplansky - Set theory and metric spaces
[This book... The first thing you'll note is that it's less than 150 pages long. It's essentially a highlight reel of the subject's foundation it also touches on many naively interesting concepts, like cardinals and ordinals. If you've never done anything with ZFC set theory, this is a fun, quick place to start]
Ebbinghaus - Mathematical Logic
[After going through this book, I'm not convinced there isn't a better treatment of the subject (Maybe some logicians can weigh in). That said, it's where I learned what Gödel was actually saying, not the confused pop notions.
If you have a good grounding in linear algebra already, you might enjoy
Lax - Linear Algebra and It's Applications
[Note, it's not "applied" in the common sense of the word. Also, if you don't have prior exp. the book will be... rough sailing. My honors prof used this—it was my first exposure—he advised the entire class to not blink while reading less we miss something vitally important]
Contra you: The math textbooks that I've red are all pretty good and I highly them. Here they are, in ~chronological order of reading:
- Probability and Statistical Inference by Nitis Mukhopadhyay
- A First Course in Linear Model Theory by Nalini Ravishanker and Dipak K. Dey
- Introduction to Time Series and Forecasting (second edition) by Peter J. Rockwell and Richard A. Davis
- Discrete Mathematics by Norman L. Biggs
(My mom's a statistician so that's why there's such an emphasis on statistics, and anyways I'd recommend the middle two much less than the first & last.)
I've been reading Wigderson's Mathematics and Computation recently. It's a bit reference-heavy at times but does a great job giving a general survey of a whole lot of the field of Theoretical Computer Science.
I'm glad to see "Visual Complex Analysis" by Needham made your list. I took a graduate class where the professor taught from that book and it was the best math class I ever took, hands down. That book changed my view of what a math textbook could be. Most textbooks refuse to encourage intuition, this book does nothing but.
Thanks for the Needham shout out. I started Penrose's "Road to Reality" last year and got bogged down in chapters 7 (Complex number calculus) and 8 (Rieman surfaces and mappings) Needlam is a referenced by Penrose in chap 7., will it also help me with the Rieman surface? Or something else.
(Physics background. I use complex analysis, but with little deep understanding.)
I really really enjoyed reading Road to Reality a couple of years back. It's a funny book, because you really need to already know most of what is in the book to understand it. What Penrose does is take a lot of complicated mathematical concepts (that the reader is ideally already comfortable with), and give a very hands-on insight into them that textbooks don't necessarily supply. He also of course illuminates connections between them in order to build up the whole edifice of modern Physics.
A good source for Riemannian surfaces is the book by Rick Miranda. Reading perhaps the first chapter of that should suffice for most of what Penrose does
Thanks for the recommendation! Is there value in reading these lectures now, if I have already taken undergrad and some grad courses in Physics. Are there a lot of valuable insights in them that regular textbooks don't necessarily supply?
Oh my. Physics type, Yeah Feynman is great. I'll reread these every ~5-10 years. Most of the physics I use is in there. .. not really so good for a first year student. But it's full of insights as to how Feynman saw the world. They are free online from Cal tech.
A few of us in my hi-tech company did a book club where we worked through a fairly approachable book on Category Theory titled Seven Sketches in Compositionality. Some bits of it were harder going than others but overall it was distinctly enjoyable. (It helps that I love Category Theory pretty much precisely because it spans a lot of different areas of mathematics and finds unexpected connections tying them all together.)
Thanks for the recommendation! I've read some applied category theory on Joan Baez's blog, and didn't get a clear idea of whether using category theory in things like chemical equations, bio systems, etc gives relevant insights that the pre-existing model does not. Does the book address the usefulness of category theory in these domains? I suppose one use could be that we now have a common language to talk about these disparate domains
I really enjoyed _A Distant Mirror_ by Barbara Tuchman, an excellent history of the 14th century. It's not exactly light reading but it is very well written and entertaining (if you like that kind of thing)
It’s different than your example book because it’s more narrative driven, but otherwise it qualifies as light non-fiction, well-written, and very entertaining. It’s a Cold War spy thriller about a KGB agent who became a double agent for Britain in the 70s and 80s.
Semi random author list. (You've probably know some.) Oliver Sachs, E.O. Wilson (I even enjoyed his tome on ants.) James Gleik, R.V. Jones ("Most Secret War", I also liked "instruments and experiences", but it's hard to find.) More physicsy B. Pippard "Physics of Vibration vol I... this is for the experimental type.
Oh and on the same subject, does anyone have a recommendation for a good intro Economics text? Thanks.
I think there's a freely available textbook on the Marginal Revolution website. I think I read the first chapter and it had lots ot helpful illustrations, etc.
If you have any interest in history, then "In The Lion's Court – Power, Ambition and Sudden Death in the Reign of Henry VIII" by Derek Wilson. It dates from 2001 so not the very latest scholarship, but it's not too heavy on the history and scholarship side and takes the approach of picking six men named Thomas who revolved around Henry in his court between 1499 to 1549.
"This book tells for the first time the interlocking stories of six Thomases - Wolsey, More, Cromwell, Cranmer, Howard and Wriothesley - who served as close advisers of Henry VIII - and all suffered for it. Two were beheaded, two were disgraced and narrowly avoided execution, one was burned to death and the other probably took his own life. In the Lion's Court is a revelation of just how perilous it was to be close to England's most tyrannical king."
Wilson definitely has his own opinions as to who the goodies and baddies were, but that makes it a light enough read.
Entangled Life by Merlin Sheldrake. One of the most enjoyable reads of the last year, surprisingly relaxing. It takes multiple interesting approaches to fungi and mycelium. The audiobook is also quite good and read by the author.
Highly recommend “Into Thin Air” by Jon Krakauer, which is about a disaster with a guided Mt Everest climb in the 90s. He participated in the climb, reached the summit, and was involved in the attempted rescue of his fellow climbers.
I agree with the other comments about the difficulty of determining how much of a topic you know with that kind of granularity, but I assume you’re referencing the “80% of the topic/skill in 20% of the time and vice versa” idea, which I think has some real value.
I’m a business transactions lawyer. I think I’ve learned 80% of the practical domain of solar generation project M&A, and it took me (i) probably three of my law schools classes (Contracts + Business Organizations + a drafting course), which I’ll cumulate as six months, plus (ii) 2.5 years or so of licensed practice with close mentorship and long working hours. I think the rest of the 80/20 proposition is true, i.e., it would take me another 10-15 years to learn the rest.
By contrast, I’ve been studying piano and composition since I was seven, and I still don’t think I’m 20% there on any measure.
Do you think most of M&A is related to memorizing laws and past court cases? Is something else involved?
Piano seems hard because it combines moving your body with memorization, so you can't strictly memorize key sequences to become good at piano. Do you play any other instrument? Piano is particularly difficult
That would be in the wheelhouse of an M&A litigator, someone who sues or defends about contested M&A transactions. I’m a transactional attorney, so I negotiate and draft M&A transactions - structuring the deals and writing contracts, basically. I say all that because, as opposed to litigators, what I do is all about process, writing skills and speaking skills, and very little about memorization. I think that’s what made it easy to learn quickly. But writing that now, I realize that maybe it’s cheating a bit not to count the many years of practice writing and speaking persuasively that set me up to do transactional work. It would be tough to draw that line for any skill, e.g., if you’re talking about learning a third language, do you have to credit any time spent learning your second language, which probably made the third one a lot quicker? Etc.
I do play several other instruments, and I agree piano might be the hardest in a certain way. What I’ll say about piano, though, is it makes your body think that pressing a button will get you the sound you want. That translates okay to fretted string instruments like guitar or banjo, but it sets you up very poorly for violin, trumpet, or any other instrument that requires you to make the sound with your body much more directly and finely. I think the skill curve for instruments of different families look a lot different for that kind of reason.
Oh thanks for clarifying. Yeah, writing and speaking probably have spillover effects, although I think those effects in traditional education have been studied and found to be smaller than previously thought.
I'm trying to learn natural language processing right now, and there are tons of steps and tons of decisions to make at each step. Luckily, there isn't a time component like a musical instrument where I have to write things at a certain pace, but the learning curve is steep. I'm trying to figure out if there is a better way than what I am doing now, which is looking at a working model, then reading the documentation on each part until I know what is going on. I think I'll have reached my much hated '80%' mark when I have a taste about which decisions to make and reasons to make those decisions, rather than copying other stuff that already works.
Bon voyage man! Sounds like a very worthwhile effort. Without knowing hardly anything about CS - I have to imagine that what you’re doing (working from a working model, reading documentation on each piece) is the best and most efficient way to go about it. Doesn’t sound like the kind of thing that lends itself to watching lectures or some other way of learning.
One way I like to approach a new topic is to find a good undergrad text on the topic, and then read it like a novel.... not worrying if parts of it are confusing.
Years ago someone in an SSC open thread asked me for the best paper on educational fadeout effects, and in particular the (false imo) claim that they are a measurement artifact. I wasn't really satisfied with the papers that I could point them to at the time, but here's something new and better: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7787577/
Fair warning it's over 40 pages but lots of great stuff; I like the consideration of methods and measurement effects best. (reposted)
My cryptocurrency hedge fund is looking to hire an engineer. PM me if you're interested. Small team + very high potential upside.
Highlights:
* Fully quantitative price prediction and trading (we don't do MM or exchange arbitrage)
* Our returns are uncorrelated to crypto market (pure alpha)
* One of the few funds to offer a BTC share product (we make people more BTC)
* We have been raising additional capital crazy fast this year due to our stellar performance.
Two of the co-founders (me + Satvik Beri) are long term EAs. I've personally donated $$ to MIRI + CFAR way back in 2012-2014. I was also a co-founder of Arbital, in case you heard of it. ;)
We have two fully remote positions: one is for 5+ years of experience, another for 7+ years.
I'm a bit hesitant to add any other requirements, since we mostly just look for smart engineers who get shit done and who would be excited to work with us.
Hey, I'm a longtime reader (thanks for all the great content Scott!) but just made an account upon seeing this. I'm super interested in quant trading stuff, only have about a year of industry experience (in software development and data analytics), but 2+ years working with a startup and in labs, and I've been playing with algo trading since high school. B.S. in mathematics. I know this is a long shot, but let me know if you'd be willing to consider me!
I'm already employed as a research economist. I just feel some affinity to UIUC. But, if you ever have an interest in talking about regulation or working with a university based research group, let me know. This is me: https://www.williamrinehart.com/
hi there alexei, do you have any other way of getting in touch? i might be interested, but do not see any email on your profile (maybe it's a subscriber-only feature?) and do not use linkedin.
Does anyone have a theory as to why thinking about philosophy or other high level concepts can actually feel physically heady? Why does galaxy braining feel this way?
I think this might be what's more commonly known as "insight porn" (the name itself basically explains why it feels heady), unless what you mean by "philosophy or other high level concepts" is much more general.
Does "insight porn" as a phrase imply that the topic has little or no content? I looked it up in urban dictionary and it's defined there as sortof cheap tricks of the "theory of everything" variety, reading with no actual learning going on.
Thank you for the definitions. That "insight porn" exists makes sense. By definition there are other experiences of thought that are not insight porn. I expect that there may be experiences or sensations associated with multiple types of thought.
Ha maybe but idk if it's really an exciting feeling. Just like, spacey and removed from my body when thinking about how wild it is that there's anything at all. Maybe it's just an emotion I haven't named.
Talking in "far mode" is associated with high status, talking in "near mode" is associated with low status. When you think about philosophy, you feel high-status. Talking about philosophy would possibly feel even better, assuming that other people listen to you respectfully.
I don't have an exact model why this is so (maybe Robin Hanson could explain better), but if you ask who is unable to think and talk in far mode, two obvious answers are: stupid people and people who are too busy working. Therefore, philosophising signals that you are neither a moron nor a slave. There is probably a quote from Plato that debating useless and unempirical shit is the most noble action a human soul is capable of.
Once philosophising becomes a known signal of high status, it is further reinforced by social reactions. When a low-status person starts philosophising, they are told to shut up. When a high-status person starts philosophising, people listen respectfully and pretend to be impressed. Knowing the respected philosophical ideas also signals that you are educated.
Philosophising is very popular even among people whose ideas are complete bullshit, because it signals "hey, I consider myself high-status". And if not challenged, then indeed it is a signal of high status; either because the listeners are so stupid that they are impressed, or because they want to avoid conflict with the person who asserts high status.
I think it's not just high and low status, but just signaling in general. People who feel superior for doing philosophy, and people who feel superior who willingly ignore/hate philosophy. But I don't think it's all signaling anyway: you say that "pretend to be impressed", but I feel like they probably *are* impressed (even if on closer inspection they would've realized that the idea wasn't that interesting). And there are definitely a lot of people who think about far mode ideas without socialization: for example, taking hallucinogenics by yourself would probably make you feel spiritual etc.; ie it's not just that it feels good. I think if it can be experienced without other people, then signaling probably isn't going to explain it, at least not entirely. (Although, perhaps drugs replicate the emotions from being in a group; although although, this makes me think otherwise: https://www.youtube.com/watch?v=loCBvaj4eSg)
I do not think any of this actually contradicts the signaling hypothesis.
Things become signals of something, because there is a good reason why they are naturally associated. (Debating abstract thoughts requires intelligence, which is why it is signal of intelligence.) When the signal becomes known as a signal, people will try to countersignal. ("I am so smart that I don't even need to discuss philosophy to make people see how smart I am.") And people are also signaling to themselves.
Max Stirner (The Ego and Its Own) said that having a theory about something can make you feel like you've dominated it. I think he was on to something.
China has been quietly building roads, military bases and villages in Bhutan's territory. The area itself is of no real value (which is probably why it wasn't noticed). But is thought to be motivated by trading for other territory that's more valuable near disputed Indian border,where there have been shows of force from both sides recently.
This is partly interesting for geopolitical implications, but also because the whole idea of stealthily taking over parts of another country by building there is bizarre.
“the whole idea of stealthily taking over parts of another country by building there is bizarre” - Pretty metal. Sounds like a state-level application of adverse possession.
In September 2018, the Finnish police supported by military raided a series of holiday villages / vacation homes in Finnish archipelago that were owned by a single company and had been in construction since 00s, but the business was never operational. The twist is: the company that owned the sites were owned by a Russian and the sites were located near strategic sea routes. There were media claims that that in addition to the luxurious villas, the vacation homes in question had some facilities (helipads, oversized piers, etc) which (according to some) made them look less like a vacation business and more like secretive, plausibly deniable bases for naval operations.
Government never commented on the military aspects, but the police confiscated some millions of euros. According to latest news items, the business is still being investigated for money laundering, with no charges yet. The Russian owner Pavel Melnikov later gave an interview to newspapers, where he claimed that he has a hobby of collecting islands not unlike some other people collect art, and he bought the islands because he liked swimming around them. The sites in question have slowly been sold away.
Why don't Western governments simply make paying ransomware illegal? This seems like a classic collective action problem- it may certainly be rational for an individual actor to pay a ransom, but it feeds more & more cybercrime by creating a huge market. While ransomware will probably never go away entirely, taking revenue away from large criminal organizations is generally considered a good thing. Realistically the government probably can't stop individuals from paying a ransom, but large corporations tend to be fairly conservative and 'hey Mr. Executive, federal prison is not a nice place' is a fairly convincing argument. This law doesn't have to deter 100% of ransom payments to just make the practice overall much less profitable.
Not even sure you'd need to pass a new law- for the US, anyways, I am very confident that somewhere in our massive list of federal laws 'making cryptocurrency payments to organized crime' could be covered by something already in there. The entire finance/payments industry is already very highly regulated.
Agreed, but it seems like enforcement has not been a priority for the US so far. There has to be a really good reason that my dumb (non-attorney) brain hasn't thought of.
I would imagine that deterring bad guys now before they encrypt, like, JP Morgan's servers seems like a good idea. Or Raytheon or Microsoft or Mastercard or someone
Ah interesting. But would paying a ransom be considered a financial crime or something more in the area of theft, robbery etc? Also, after the first couple of high-profile cases, wouldn't it become unrealistic for an accountant working at a large company to claim ignorance of the illegal-to-pay-ransom law?
Acting willfully does not require an intent to break the law. "Someone commits an act willfully when he or she does it willingly or on purpose. It is not required that he or she intend to break the law, hurt someone else, or gain any advantage" https://www.justia.com/criminal/docs/calcrim/800/960/
Ditto re acting intentionally: "A person acts with wrongful intent when he or she intentionally does a prohibited act[or fails to do a required act]; however, it is not required that he or she intend to break the law." https://www.justia.com/criminal/docs/calcrim/200/252/
I really have no idea. But suppose a large bank was paralysed and hundreds of thousands of accounts inaccessible. A ransom of say $1m would solve the problem. What should the law be in such a case?
I have no solid answer, but maybe the blast radius is one of the reasons? Ransomware can bankrupt a company if they can't get their data back, which would put its employees out on the street. Though I suppose there's no reason it couldn't be made illegal to pay ransomware and you'd also be guaranteed financial support; but *then* people might infect their computer systems with ransomware to get the financial aid without having to work (would really up insider risk - as someone's IT admin, would you rather work 9-to-5, or get the same compensation for a comfortable amount of time after your company was infected with ransomware)?
To stress: I am just thinking out loud. This topic is way more complex than I could wrap my head around. Grain of salt.
Couldn't such a law be made pretty much arbitrarily difficult to enforce by opening multiple Bitcoin wallets and breaking the payment into multiple transactions? Assuming the criminals are well-hidden and reporting rather than paying means your data are gone forever, reporting means you're out of business and paying secretly still has a pretty low chance of serious repercussions.
> large corporations tend to be fairly conservative and 'hey Mr. Executive, federal prison is not a nice place' is a fairly convincing argument
There is a solution for that: "data recovery firms" that pretend to recover the data, but actually just pay the ransom, giving the customer plausible deniability.
Of course David Friedman has an analogy from 18th century England: "Jonathan Wild, self-appointed Thieftaker General," who got "rewards for the recovery of stolen property, and income from the large-scale employment of thieves".
It doesn't matter whether they are illegal or not; the American people are not so coldheartedly rational as to be willing to put crime victims in jail just because the consequentialist math says it generates more net utils, so for something like this they simply won't. Prosecutors will find that, gosh, they're very busy and there's ten more important things on their agenda this week. Juries will decide that "reasonable doubt" extends to the slightest fig leaf of bare possibility. And if you somehow pass a no-exceptions mandatory-prosecution law, then someday the law will fall on a telegenic minority woman running a small business who's just trying to make payroll for her equally telegenic working-class staff, and This Is An Outrage and Something Must Be Done.
It seems like a minimum first step would be to make the ransom itself unspendable. With Bitcoin that’s easy, just make the coins themselves the equivalent of registered stolen goods—subject to seizure even from a bona fide purchaser for value.
I wouldn’t say I want to attack Bitcoin as an ultimate goal, but if forced to choose between attacking Bitcoin and not taking steps to deter ransomware attacks on critical infrastructure I’d definitely pick attacking Bitcoin.
I’ve been feeling stuck in a solipsistic frame of mind for a year or two - constantly wondering how I can have any belief that anything is real outside of my own mind. Has anyone here found a satisfying/permanent way out of that? I used to rely on a kind of soft religion, but it wore off. Would appreciate any pointers.
By realizing that there is nothing you can really do about it. If your senses are feeding complete garbage about the world and you cannot stop them from feeding you nonsense, then there is nothing else you can do. I do think that there is a real, external world and my senses can provide, at the very least, a few clues on how it actually works, but I doubt you can 100% prove it is so, it is best stop thinking about it too much and just roll with it.
I appreciate that. I think you’re getting at what I’ve started to find uncomfortable about it - it’s like a tick or an intrusive thought, i.e. something that I know is doing me harm (in terms of distraction, frustration, brief fits of fatalism) and that I know I can’t think my way out of but that I can’t stop thinking about. I could describe some of the ruminating habits I used to have when I was depressed the same way, I wonder if maybe employing some of the same mental tricks that helped with depression would help with this. Again, thanks for your thoughts.
It's not the kind of thing one can think one's way out of, and it's so compellingly grim that it might not work to tiptoe away from it via attending to gentler alternatives (eg, "soft religion"). What about plunging into some things that provide strong sensations and emotions -- rock climbing, other thrill sports, aggressively fighting injustice, helping the desperate in a direct and physical way?
So I think you hit on something here! A lot of the onset of the solipsist thinking coincided with finally finishing school, getting my “real” job, getting married, getting a nice place and a dog, etc. - settling down. And I think that a lot of the wild emotional ups and downs and the variance of experience inherent in daily life as a teenager and an early-twenty-something helped to keep me grounded in reality. Now that I spend every day doing almost the same thing, I’m just spending a lot more time in my head (in a bad way). It might be that I’m just not mentally cut out for the daily grind and need a lifestyle with more excitement and variance to keep myself level. Thanks for your thoughts!
Bit odd that he creates the world just for you and doesn’t communicate, for sure. Another argument against solipsism is to compare your life with an actual simulation. Ever play a video game where you were living an ordinary ho hum life? Sims excepted I suppose. Imagine buying a video game where you were nothing special but the world was vast and plenty of the NPCs were doing great. You could watch this in your low rent apartment on TV. You need to work to own this tv or you get kicked out if the house and starve. Eventually you die doing not much. That’s it. That’s the game.
Lol that’s so bleak but you’re right. I don’t know if you like $5 PC games but Dave Man might give you a laugh, it’s just like what you described. I’m hung up on what you pointed out, that everything your senses give you is an adapted representation anyway, at best. I’m sure that plays out on a lot of different levels (e.g., I interpret others’ behaviors in a way that suits my survival; also, I perceive colors, objects, and the rest on those emergent levels without knowing for sure whether what they seem to represent really exists to another observer).
For what it’s worth, the idea of God does consistently present itself as the “out,” I just wish I could flesh out an idea of him/her/them beyond “another observer.”
It's funny, because I'm listening to "The Master and His Emissary" right now, which goes into a lot of detail to discredit Solipsism. Also, what Nancy Lebowitz said
Total non-professional opinion - I think this may be a form of OCD. It might be counterproductive to try thinking your way out and revisiting the possibility. Resisting can make intrusive thoughts even more disturbing and frequent. The way out might be to learn to live with the possibility. Again , I have no real expertise on the matter, just throwing out a possibility. Best of luck!!
It depends on your underlying epistomology. I think most commonly accepted theories of justified belief will allow you to accept that the external world exists (even classical foundationalists, who I think also include the most serious idealists). For a simplified example, suppose you accept a coherence theory of justification. Then it's far more coherent to add your set of beliefs that "the external world is real" than to add "the external world is fake." Obviously it's more complicated (and fun to think about) than that and would recommend any introductory epistemology book. More personally I find that it's better to give up on gaining ultimate knowledge and focus more on what you are justified in believing. Also, consider seriously what your standard for knowledge is. Generally to have knowledge of something we don't need to be certain it is true.
Have you sat down and tried to define, really really thoroughly, what precisely you mean by "my own mind" and "outside" and "belief" and "thing"? My experience of contemplating solipsism is that, once I've defined (or failed to define) the terms it uses to express itself, it stops seeming like a coherent problem. Put another way: when I express, as truly and completely and precisely as I can, what solipsism is trying to refer to when it says "my own mind", I find that it's not something that could conceptually do the thing solipsism is suggesting it might do.
A third way of trying to say it is that solipsism exists because we're married to terms that exist in a very different web of relationships than the entities the terms are supposed to correspond to. The terms are just wrong. My actual mind doesn't behave the way the English-language "mind" does. Real things don't behave the way the English-language "thing" does. So solipsism is a logical statement when used of those terms, but it doesn't actually make sense for the entities themselves.
This. The problem with solipsism is that it seems like a meaningful position, but it’s actually incoherent since it denies the basis on which words and thoughts mean things.
Nah, that’s a terrible argument, just like his naturalistic argument is. Both covertly smuggle in the conclusion, distract you with an irrelevant thought experiment, and then declare victory.
You're just asserting that knowledge requires proof (like knowledge of hands), but this clearly isn't true. This is touched on in the article I linked.
That’s not the problem with Moore’s argument. The problem is that it assumes that a hand is an external object, which is the very thing in question, then presents the fact that hands exist — which was never in question — as an argument for external objects.
It’s the exact same problem with kicking a stone to refute Berkeley’s idealism, or, from another angle, trying to refute a trivialist by making an argument that trivialism is false: the existence of hands, the ability to kick stones, and the falsehood of trivialism are all entirely compatible with the view being refuted.
I've been amused by solipsism and subjective idealism framework in general in my salad days. I still think it's a valid default position for a young philosopher. For instance, I used to claim that materialism imply a huge extra entity (Objective Reality) so it must be more complex in the Occamian sense and therefore less probable.
However, you learn more things, it turns out to be the opposite. While materialism does imply the existence of objective reality it delivers on all its predictions. Unlike solipsism which doesn't have any predictive power. Moreover, if your consciousness is fundamental, why don't you have psychic powers? Why does your experience seems stable? Conservation of energy and matter seems much less likely in idealistic universe than in materialistic one. While this by no means disprove solipsism and idealistic framework in a strict sense, it makes it more complex in Occamian sense and therefore less probable explanation.
It's complicated, and some of his sources claim that Australia is not, in fact, unusually high in venomous species. But the shortest answer is: Cold-blooded species can only thrive in warm climates, and are unsuited to hunting strategies with high energy expenditure like chasing or overpowering. As a result, they are more likely to evolve venom to hunt with.
I wouldn't call that video "good". The guy gets uncomfortably in-your-face, and the signal-to-noise ratio is low. The same information could be packaged in half the time.
Also, is number of venomous species per country a good way to look at the situation? Country borders are arbitrary, and countries are of widely varying sizes.
However, the number of venomous species in Australia might be a matter of chance. I'm not sure how much they're in the same environments so it might not be an arms race.
Hmm, venomous is associated with hot climates. (not sure why) I wonder if it's also associated with dry? I'll also observe that Australia spans a temperate climate in the south, (lots of rain and huge gum trees.) to desert, and then tropical climates in the north. So lots of different habitats. There is ring of life around the edges, and desert in the middle, maybe that makes it harder for species to find new habitat. (Sorry just speculating wildly.)
I recently bought a new house. The disclosure mentioned that subterranean termites had been found in some wood in the basement when they looked for them, but that the seller hadn't done anything about it and I needed to decide how to react. My options seem to be to do nothing, or to pay a very large amount of money to an exterminator and wait for them to do stuff before I move in.
When I Googled termites, I found many exterminator company websites saying you needed to exterminate them right away, but also some articles saying that about 50%-60% of houses in California have the kind of termites my house has. If my house has a completely average number of termites, this sounds like less of an emergency.
Does anyone know more about how I should think about this?
+1 is it not normal to get someone to inspect the house *before* you buy it, so you can renegotiate the price if a lot of expensive work needs to be done?
Termites need wood and moisture. You can reduce moisture with a de-humidifier. Also, cover up any spots where outside humid air can enter the basement. Obviously, check the wood in the basement for any damp, rotting wood. If you see that you probably need extermination. Assuming you don't have that, keep the moisture low and get an annual inspection by a reliable professional.
>cover up any spots where outside humid air can enter the basement
Maybe it depends on construction techniques and the local climate and water table, but I thought that builders deliberately put in air bricks so that inside humid air can leave the basement? and that house owners covering these air bricks over or letting them clog up was a common cause of damp problems.
Indeed. I've renovated houses in England and France and without fail houses in both those countries are more humid inside than outside. A central feature of modern house building (at least in Europe..) is to design the structure such that moisture can leave the house. And also to do so without it condensing on a cold surface on the way out.
You should pay to have them exterminated. Ultimately, you will need to sell the house. If termites have been eating your foundation, many buyers will not be willing to buy. If they destroy your foundation completely or damage it, you won't know until your floor starts sagging and you have to get that replaced.
Source: had to remove a bug infestation recently from my foundation before selling my house
If this wasn't an insane seller's market, an active termite infestation would be a hard pass for me. There is no easy way to determine the severity, it could be anything from "nothing" to "the structural components of your house are now compromised and it will take massive amounts of time and money to make it safe." How long do you plan to live there? Do you plan to do major renovations that involve opening up walls? If you plan to flip it in a few years you might get away with doing nothing, especially if the market stays hot and buyers don't have much leverage. If you're there for the long haul, or you are going to expose the wood and risk seeing thousands of dollars of needed repairs, you should contact a trusted pest control professional and follow their recommendations.
It's a case of "ignore small problem now, develop big problem later". The seller ignored it and left it on the buyer (in this case, you) to deal with it. Deal with it now, because if you intend to live in that house for any reasonable period of time, you will eventually have to deal with it. Even if you only intend to do up the house and hope to sell it on in a couple of years' time, a prospective buyer is going to have the same problem as you do now, only worse by a couple of years' neglect, and they may well decide either not to buy or not to buy at the price you are asking.
Honestly, *anything* to do with a house? Fix it NOW before you are *forced* to fix it because by then it will be bigger problem, cost more, and be an even huger pain to deal with.
Looking 'em up on this University of California website, you should deal with them now while you have the chance and the least inconvenience. http://ipm.ucanr.edu/PMG/PESTNOTES/pn7415.html
These ones like damp, so that may be an indication of a further problem (leaky pipes?) with the basement, and since they prefer rotting wood to sound as it's easier to chew, yeah you should have this looked at and don't ignore it. And it seems like fumigation doesn't work so pesticides have to be sprayed, another reason to get this done *before* you move in:
"Moisture Requirements
Subterranean termites require moist environments. To satisfy this need, they usually nest in or near the soil and maintain some connection with the soil through tunnels in wood or through shelter tubes. Furthermore, because of the moisture requirements of subterranean termites, they are often found in wood that has been slightly decayed. Soil serves as a source of moisture that protects termites from desiccation, shields them from predators, and can be used as a building material for shelter tubes and carton nests (Formosan subterranean termites) above ground. Termites can also excavate passageways through the soil to reach additional food sources.
Feeding
Termites do not like all wood species, but the condition of the wood is more important in determining the probability of infestation. Decayed wood is eaten faster and preferred over sound wood. Digestion of wood, in this case, really begins before the termites take their first bite, since decay fungi in the wood break down cellulose into smaller units. Termites can digest sound wood, but decay fungi make their work much easier.
MANAGEMENT OF SUBTERRANEAN TERMITES
It is unlikely that homeowners will be able to execute subterranean termite control on their own. However, it is important for homeowners to have some familiarity with inspection procedures, reduction of conducive conditions, and treatment strategies. Successful termite management requires special skills and knowledge, including a working knowledge of building construction. An understanding of termite biology and identification can help a homeowner understand and select a suitable method of control. Of course, homeowners can replace termite damaged wood and correct conditions conducive to subterranean termite infestation on their own; however, applications of registered pesticides are highly regulated and require a licensed pest control professional to carry out the inspection and control program.
Multiple colonies of the same termite species or several different species can infest a building. A professional inspection and an integrated approach to control are required. A combination of methods, such as habitat modification, elimination of excess moisture, removal of infested wood from the structure, exclusion of termites from the building by physical and/or chemical means, and the use of chemical methods to destroy existing colonies will probably be necessary.
Most subterranean termite species consume wood at about the same rate, but three factors can make some species potentially more voracious and damaging than others. These factors include the environment in which they live (termites eat more wood when conditions are optimal over a longer period of time), the size of the insects (larger insects eat more wood), and the number of insects (larger colonies eat more wood).
Controlling Subterranean Termites
Subterranean termites in structures cannot be controlled using techniques that are appropriate for drywood termites, such as fumigation, heat treatment, freezing, and termite electrocution devices, because the reproductives and a large majority of the termites are concentrated in nests near or below ground level out of reach of these control methods. The primary methods of controlling these termites are insecticides, either applied to the soil adjacent to the structure, directly to nests via shelter tubes, or through bait stations. To facilitate control of subterranean termites, destroy their shelter tubes whenever possible to interrupt access to wooden substructures.
Insecticides
Liquid applications of pesticides are most often used for subterranean termite control and applied to the soil either in drenches or by injection. There are no reliable over-the-counter termite control products available for the public in California; all effective products are for professional use only.
Pest management professionals are provided special training because of the hazards involved in applying insecticides to the soil around and under buildings. Applications in the wrong place can cause insecticide contamination of heating ducts and/or damage to radiant heat pipes or plumbing used for water or sewage under the treated building. Soil type, weather, and application techniques influence the mobility of insecticides in the soil; soil-applied insecticides must not leach through the soil profile to contaminate groundwater or run off to contaminate surface water."
Here are my words of wisdom and experience, first in bullet points:
-- Termites are fucking Satan's evil incarnate and they never stop.
-- Fumigation is a Band-aid that is done for insurance reasons and because a crew can just show up and do it, just costs $$$ and helps you not see termites for maybe a year. All these termite companies do the same thing and the market is huge because of CA house prices. Think of them as the Earl Scheib option, an auto paint company that just sprays a new coat on but does not do any prep or repair and misses a lot of stuff, but it looks ok from 30' away.
-- The only way to really solve the problem is the Nuclear Option, and it is a huge amount of work and disruption requiring you or a trusted contractor to touch and treat every single cubic inch of your house, drill holes all over the yard and even the hardscape. It is the only way you can be sure that your house will not collapse two weeks from now.
Here is a link to a Tech Bulletin from the only Company I trust.
I'm sorry you let this slip by you and now have a termite problem. Of course, all wooden structures in California have a termite problem, so fumigation status would seem to me to be a very important part of due diligence.
Over the last ten years in LA, I have developed a standard visual spotter's guide to gentrification:
1) Fumigation tent/enclosure goes over the house
2) Redwood fence installed. (You can tell everything about the new owners by how the fence is executed: Did they do it as cheap as possible? (house flipper) Did they ask for something they saw on Pinterest? (Clueless first-time-owner.) Or did they do it like it's going to actually last and function for 25 years? (These people plan to LIVE here.))
3) Prius or similar EV appears in driveway/curb.
I suppose upon reflection that the reason fumigation happens post-sale is that it is impossible with people and their stuff actually living there.
OK, I will wrap this up and give you the news: Everything must be treated with Disodium Octaborate Tetrahydrate. The stuff is not expensive in bulk if you want to DIY, and Thank God it is still sold in California, you must act fast before it too is banned.
If you dropped one AI into a game with nine other humans it would of course do quite badly, but if you dropped one human a game with nine AIs, I bet the AIs outdo the human. I have a hard time defining what a "fair" structure would be for testing humans vs AIs here: the asymmetric nature of the game means you can't even do like DOTA 2 and make it five humans vs five AIs, and putting humans on the same teams as AIs will just drag the team down with communication problems.
I guess I would have the criteria be to have the ai master any scenario: 1 human vs all ai or all ai vs 1 human and every other combo. The interaction with other players required in this game would be similar to a Turing test. So, yeah having the ai master this seems to me like it might have to be a general ai - which much more advanced than I was thinking of. I had games like Civ or the paradox games in mind, which I would suspect making a dominant ai would be easier than Among Us.
I've encountered groups of humans who would all cluster in exactly one place (pixel perfect) waiting out other people. An AI that figures out that cooperation mode with other AIs always wins.
I think whatever it is needs to be realtime, the possibility space is just so much bigger and therefore more resistant to brute force (or in modern ML, slightly smart force) searches. Starcraft 2 shows this in action: the early version of AlphaStar had obviously superhuman micromanagement but failed to grasp basic tactics like "don't pick fights by attacking through a chokepoint and up a ramp". Even the final version seemed to run on really good economy management shoring up strategy ranging from simple to stupid (it lost one notable game by building a bunch of can't-attack-air units, running into an enemy air army, and building more can't-attack-air units until it died).
Right, brute forcing isn’t much of a solution for the more complicated games. Age of Empires 2 definitive edition has a very challenging AI (as long as you don’t rush the ai) But that ai’s effectiveness is due to the fact there is an optimal path of play for the game especially in the opening minutes of the game. I don’t *think* Civ, EU4, CK3, etc have such a single optimal gameplay that an AI could be programmed to to be able to almost always win.
Beating arbitrary levels in Mario Maker 2 would be tough. It's a time-limited, constantly changing environment, and the AI would have to master a wide variety of situations. It would be like putting a self driving car on a track built by a hostile designer that wants to cause crashes.
If you really push the limits of the game engine, this becomes a reading comprehension test. Mario Maker doesn't have much in the way of explicit scripting language, but you can still kludge together an astounding number of "if player does arbitrary X, do arbitrary Y to the game world" mechanisms through clever use game systems. Imagine a level where sixty seconds in, the floor will disappear unless at the very start you walk three steps to the right and jump at empty air.
You can make a level that does basically arbitrary unfair things that have no relation to traditional platformer mechanics, then explains those things via English text (so that human players can easily beat the level) and now you have a level the AI can't beat until it learns to read. Repeat for whatever other tests of human information processing you build in.
The original X-COM: UFO Defense from back in the 90's. Still actively modded and updated; check out the Open X-Com project. Very difficult tactical combat system with a ton of OOC prep, research, and equipment decisions that make or break you.
Also, sometimes the aliens are just overpowering when you show up, and you have to flee or get TPK'd.
In Open XCOM there's an ironman difficulty setting, with no save scumming. If you can make an AI that does a credible playthrough of XCOM UFO Defense on ironman, hardest difficulty, then I want to see the playthrough video. It's fricken tough.
UFO: Enemy Unknown is not actually a very hard game *except* when played blind and/or without deep knowledge of the mechanics.
Most missions can be beaten with a pretty rote plan:
1) on turn 1, use two Smoke Grenades, one dropped inside the Skyranger and one thrown to the top part of the ramp, then end turn. (This stops the aliens seeing into the Skyranger on their first turn, or getting reaction shots as you disembark.)
2) Disembark, shooting at any aliens that you spot from soldiers *further back in the smoke cloud* who can't see the aliens (and which, therefore, the aliens can't see for reaction fire).
3) Set about 1/3 of your soldiers as snipers (kneel them down in a firing line, maybe two spaces apart, and don't move them except to turn) and the rest as scouts. The highest-Firing-Accuracy soldiers should be snipers.
4) Sweep the map with the scouts, shooting at any aliens with the snipers (who are beyond visual range and therefore can't be reacted to). Stop sweeping for a turn when either your scouts are all down to about half TU, or you're running out of TU to take shots with your snipers.
4a) If there are buildings in the way of your snipers' sightlines, shoot at them with Laser Rifles until they're mostly destroyed. This also goes for jungle terrain and orchards. Sand dunes and hills are trickier; you may have to relocate your snipers to the top of them. Forests are the worst because you can't really just level them, but you'll usually have a couple of snipers with clear sightlines so it's not too bad.
5) Once you find the UFO (if there is a UFO; terror sites won't have one), camp a bunch of soldiers off ~15 squares side-on to the entrance (that way, an alien who comes out will not be able to see them, and your soldiers will be able to take reaction shots. If it's a Medium Scout, just camp there (with the majority of your troops) until all the aliens come out and get side-swiped. If it's bigger, camp with ~5-6 soldiers while assembling the rest (including snipers) to go in the entrance.
6) Sweep the UFO, room by room, with 3-man squads along each passageway (if there are too many passageways, just hold a chokepoint while clearing out one section and then bring over the freed-up soldiers from the cleared section to assist). Never move a soldier while he's in an alien's sightline unless you're moving him out of the sightline or you're sacrificing him. Prefer points for troops at end-turn that are to the side of a door, so as to get reaction shots.
The only really tricky missions are psi missions, alien bases, and Cydonia, and the latter two can be put off painlessly until you get psi tech for yourself (which makes the game boringly easy). With psi, the moment a soldier is targetted for psi-attack, have him drop everything and just passively scout for the rest of the mission; he has low psi-strength and will probably be targetted again. The trickiest bit is that you're looking to capture the psionic alien alive, which is risky and may not be achievable.
Now, the plan I outline above is not *trivial* to convert into code, but I don't see any outright judgement calls there except in the psi missions. I'm pretty sure you could code an outright dumb program to do it (and certainly you could for research order and most of the Geoscape). If a computer game's rote for a human, it's almost certainly programmable.
Now, an AI that had to compete against a human on a modded two-player X-Com would have to be significantly better. But the actual vanilla game is pretty easy (though TFTD's quite a bit harder).
Thanks for the writeup. Have you played Xpiratez mod? This strat set won't work; there's a ot of alternate vision modes, more psi enemies and varied enemy types, damage resistances, and a ridiculous tech tree.
One thing that a computer would be really good at in the original XCOM is farming stat improvements. Once you realize how it works, you can create super-soldiers if you are willing to do a bunch of boring things. https://www.ufopaedia.org/index.php/Experience_Training
An AI is also patient enough to wait out the computer AI.
There's a resource in xpiratez called "Glamour" that you only get for finishing missions quickly, and which you need for some improvements/tech. Also, having fully maxxed out soldiers in xpiratez is nowhere near as invincible as it is in vanilla. 1 shot kills happen all the time.
Seriously, there's a ton of depth and tactical requirements for xpiratez that would make it very interesting to see an AI tackle it.
1-shot kills of maxed-out soldiers happen quite a lot in vanilla UFO. The Heavy Plasma Gun does 0-230 damage per shot and can fire up to 6 times per turn; maximum armour is the front plate of a Flying Suit at 110 and maximum health is 61.
TFTD is a mixed bag; your armour will stop ordinary alien weapons, but the "armour doesn't help" stuff gets more common as well.
Apocalypse definitely has near-invincible soldiers, though, both at the start and the end (though not the middle). I usually don't lose a single soldier until Devastator Cannons show up, and a shielded/disruptor-armoured/cloaked soldier takes a minimum of three shots to bring down, which have to be from different weapons in a specific order.
EU4 and a few of Paradox's other grand strategy games get weird when examining the effects of AI micromanagement. For the uninitiated: the state of the board is recalculated once per day of the in-game calendar, which will generally be every second or three depending on game speed. Orders given by the players are input between updates, and many are immediately visible on the board to all players. This can occasionally lead to a situation where on March 23rd 1446 I want my armies marching toward Oxford, but then my opponent will see that and flank me by ordering a march towards Sussex, so I see that and change directions to head them off at Sussex, and then they choose to go back to Oxford, so *I* switch back to Oxford, and then they go back to Sussex, and....
In human multiplayer games, the typical way of solving the issue is just to have the game run at a defined speed and let the chips fall where they may, with the last set of orders at the update going through. Against the computer, it doesn't come up excessively often since the AI tends to issue fairly broad directives without too much micromanagement. But when the AI does choose to micro, it *will* beat the human through reaction time or just sheer persistence.
Right now the AI is mediocre at operational-level thinking, but already has an insurmountable edge in the finest-grain tactics. Paradox have taken some steps to reduce the effect of the latter (moves that take multiple days 'lock-in' after 50% of the time elapses), but I'm not sure it's fully soluble without a significant change to the game dynamics.
(Stellaris largely resolves that issue by making combat and a few of the most micro-critical actions all real-time, but has other AI compute time problems so bad they keep breaking the game trying to avoid them. No, I did not care for the recent pop growth changes, how could you tell.)
> showing that a common pattern is good growth until ~1980, poor growth until ~2002, then good growth until ~2012, then poor growth again.
I'm not convinced it's that common, from the graphs. It fits Sub-Saharan Africa. In LATAM and ME/NA there's a blip in ~1980, but it's not obvious that growth after is slower than growth before. And in South Asia, neither ~1980 nor ~2012 seem like anything happens.
Does anybody knows good books with a lot presented and solved analytics cases. Examples of good cases:
1. estimate average cookie lifetime for your site userbase
2. for offline store chains test impact of new interior design
3. extrapolate results of survey for all population (in survey there is x1% people from cities, y1% liberals, ... while in population there is x2% people from cities, y2% liberals, ...)
I'm not sure I see that pattern in the GDP graphs - S Asia does not show a dip in 1980, and LA/Caribbean growth restarts in about 1986 after about a 5-year dip -- but more importantly, the numbers are in US dollars (and current dollars - not adjusted for inflation). Wouldn't the numbers be affected by the value of the US dollar? I might be wrong, but would think we would want to see the numbers in purchasing power parity to see if there is an actual pattern. PS: Eg: See World Bank GNI data here https://data.worldbank.org/indicator/NY.GDP.PCAP.CD. From 1990-2019, China's per capita GDP has gone from 729 to 8242 (11x) in constant PPP int't dollars, but from 982 to 168040 (17x) in constant US dollars (and from 317 to 10216 in current US dollars, which the posted graphs use).
Current dollars *specifically is* adjusted for inflation, and won't get artifacts from the historical USD exchange rate (as the historical USD doesn't enter the picture; the only exchange rate is other-currency-then vs. the constant USD-now).
I'm not exactly a genetics expert, but I've got a PhD in molecular biology and bred mutant mice, so I might be up to the task. Are you looking for a quick check of the jargon, or a deep peer review style feedback?
Thank you for sharing that story, it was thoroughly enjoyable. Definitely agree that it’s rather Eliezer-ish, whether from actually being him, or just having a similar style.
I get the impression that there've been three main style of mainstream fantasy over the last thirty years or so, though I'm not sure if it's really a culture trend or just what I've ended up reading.
In the early nineties, you get unironic adventure fantasy set in a huge fantasy world (Think classic D&D, Dragon Lance or Wheel of Time). the focus is more on the generic gist of the world than on specific mechanics. Lots of magic and adventures, not too much time sitting around contemplating.
late 90s-2000s seems more the "gritty fantasy" era (think Game of Thrones). Lots of random brutality, occasional graphic sex, more "historical" sense in that there's less old-timey fantasy adventure values (less honour, more realpolitik and murders). Some magic, but usually more as a tool than the main story focus.
2010s-now seems more historical fiction and hard mechanics style (think Brandon Sanderson or Naomi Novik). Feels like the authors have done a lot of historical research to make their worlds historically realistic and have spent a lot of time designing formal magic systems with strict rules (not that earlier authors didn't spend a lot of time on worldbuilding, but earlier worldbuilding tends to feel more brainstormy and less processed - which one I prefer changes by mood, TBH).
Curious how accurate this feels to people and, if accurate, what you'd guess the next trend will be (seems like we're just about due for one).
Urban fantasy/paranormal romance., possibly starting with The War for the Oaks (1987). Starts with the contemporary world., and may spend the whole book there.
I believe Laurel Hamilton's Anita Blake books were the first with public knowledge of vampires, werewolves, etc., in the contemporary world rather than supernatural creatures having to live in hiding.
These books can have various degrees of grittyness, though I don't know of any that are full grim dark.
A fair number of them are in first person from a single point of view.
The thing that disappeared, rather surprisingly, was the picaresque fantasy. One or two adventurers, perhaps with loosely attached ethics, having separate adventures (not much of a story arc) in a large world of supernatural wonders. This was completely supplanted by more complex worldbuilding.
You are right, I don't think Vance characters were that likely to be wanderers, but for me "One or two adventurers, perhaps with loosely attached ethics, having separate adventures (not much of a story arc) in a large world of supernatural wonders" is a very good fit for many of his books, the cycle of the dying earth for example.
It seems fine (although I'd put the old high fantasy as having its heyday earlier) to me, although I feel your definition of fantasy is a bit restrictive (although it does match what bookstores would mostly shelve things as).
I would add in magical realism, which I think started getting popular in the late 90s (? Not sure), as fantasy that was more "literary", usually very soft and concerned with small stories. Often from translated works ala Jorge Luis Borges, but there appears to be a burgeoning amount of native English works (at least from my perspective buying ebooks)
Also the "new weird", which near as I can tell grew out of Lovecraftian horror into its own thing. It seems to have porous boundaries and straddles soft sci-fi at times (as in much of Jeff Vandermeer's work), but books like Perdido Street Station are unquestionably fantasy. Usually very dark in tone, obsessed with apocalyptic/devolving imagery, and again often more "literary" in theme and presentation.
Well, in the bookstore it'd probably be shelved as YA or romance, but I could see it counting as magical realism, although from what I've gleaned from what's trickled into the mainstream (having not read the books myself), it doesn't carry many of the hallmarks of the genre. But genres tend to be messy, porous things.
As the genre matures, it'll take on more and more of the good craftsmanship techniques that mainstream literary fiction already uses. Fantasy writers will get less and less slack for mistakes (overlong physical description, bad pacing and stereotyped characters being three major widespread flaws in fantasy). Brandon Sanderson has better craft than Robert Jordan; sorry, he just does. Each generation both learns from and builds on the previous one.
The future trend is that fantasy will begin to use modern and post-modern literary techniques. Vandermeer is leading the charge there (this is most obvious in City of Saints and Madmen), but there are others. Kelly Link is an obvious example. But you might talk about Haruki Murakami, Jonathan Lethem, China Mieville, Mark Danielewski. Stephen King's Hearts In Atlantis or David Mitchell's Utopia Avenue. The walls between fantasy and mainstream will become blurry.
"As the genre matures, it'll take on more and more of the good craftsmanship techniques that mainstream literary fiction already uses. Fantasy writers will get less and less slack for mistakes"
This is very much not how it's working out in reality. Robert E Howard repeated an adjective inelegantly here and there, but he was a far better prose writer than the stodgy brick-trilogy emitters we've been enduring since the mid-'80s. (Most likely, those repetitions were the product of him basically publishing first drafts written very rapidly for pulps, and any amount of allowance for editing would have seen him shift them himself.) Dragonlance wowed a generation by being offensively bland milquetoast, and now the rise of online fandom gives us loud, angry and voluminous apologetics for the most godawful printed-fanfic-level writing imaginable. Sanderson may be a better craftsman than Jordan, but that's only because it's such a low bar to clear. He is in no wise a better craftsman than Dunsany, Eddison, Leiber, Smith.
It makes sense to me that we would disagree about this, given that you seem to be including Eddison and Dunsany in the fantasy genre. I haven't read them, but they seem more like precursors or ancestors. It doesn't seem quite right to compare them to Sanderson; it's like asking him to stand up to Elena Ferrante.
I suspect that, as a genre emerges out of the mainstream literary works that inspired/defined it, the literary quality dips significantly. Which would make sense if fans don't read very much outside their chosen genre. And the pocket paperback glut of the 70s and the current Internet fanfiction flood wouldn't help. But despite that, I think there's going to be a trend toward advancement in any genre. I think you can see it happening in fantasy, horror, sci-fi, mystery and experimental fiction. And I think you can see the quality dip happening right now in YA, which is a much more recent genre.
I hadn't read him before, but I sampled a few pages of Robert E Howard, and I don't think his craft stands up very well next to, say, Jemisin or Rothfuss. He trounces Dragonlance, though.
But these are matters of taste. Maybe I just prefer the contemporary style, and you prefer the older style. I like my account, but I'm not too attached to it.
I can't speak to Sanderson since, while I have a vague feeling I may have read something by him, I can't call it to mind.
But the "gritty fantasy" era in the original comment is really the same genre as Low or Picaresque Fantasy, which goes back to at least the 40s with Fritz Leiber's "Fafhrd and the Grey Mouser"; maybe less overt sex'n'gore than "Game of Thrones" but the same spiced with cynicism and down among the rogues and thieves and fighting men for hire milieu than the High or Heroic Fantasy of Tolkien or William Morris (has anyone else on here read "The Well At The World's End"?)
So I don't think that it's so much "new sub-types arise" (although those do happen), but rather that the times of the real world favour different sub-genres; sometimes we want heroes who are truly heroic, other times the taste is for anti-heroes. Sometimes it's highly coloured sword-and-sorcery or planetary romance, other times it's sliderule physics hard science.
Fafhrd and the Gray Mouser were cynical about people in general and about the gods, but I don't remember whether the stories were cynical about governments.
Conan is sort of an odd example of grittiness. As a king, he's a very good king who protects his people from greedy aristocrats and protects a religious minority from persecution. But most of the stories are before he becomes king, when he's a thief and murderer and looks heroic only because his enemies tend to be even worse (and because his own murders are mostly offstage).
Conan does get a degree of character development over the entire series of stories, and he does have his own ethical code that he sticks to. He's also not a total grimdark hero or rather anti-hero, as he does show unexpected sensitivity at times, e.g. in "The Tower of the Elephant":
"Tears rolled from the sightless eyes, and Conan's gaze strayed to the limbs stretched on the marble couch. And he knew the monster would not rise to attack him. He knew the marks of the rack, and the searing brand of the flame, and tough-souled as he was, he stood aghast at the ruined deformities which his reason told him had once been limbs as comely as his own. And suddenly all fear and repulsion went from him, to be replaced by a great pity. What this monster was, Conan could not know, but the evidences of its sufferings were so terrible and pathetic that a strange aching sadness came over the Cimmerian, he knew not why. He only felt that he was looking upon a cosmic tragedy, and he shrank with shame, as if the guilt of a whole race were laid upon him.
'I am not Yara,' he said. 'I am only a thief. I will not harm you.'"
Very happy to see The Well At The World's End mentioned here! I read it a few years ago. The language and style were tough to get used to but it gave me the most Tolkien-esque feelings of anything that wasn't Tolkien. (To be clear, I mean that as a massive compliment! And it might be worth mentioning that I *didn't* have the same reaction to other works that get cited as pre-Tolkien examples of genre fantasy-- Phantastes, The King of Elfland's Daughter, or The Worm Ouroboros.)
I love "Phantastes" and "Lilith", but MacDonald is his own style and I think you can definitely see the "Scots devotee of German metaphysics" at work there as with Carlyle. I constantly have the feeling "okay, there is something going on here that is going over my head" but that is not a bad feeling.
"The Worm Ouroboros" and all the Zimiavian novels are again sui generis. I think you need a high tolerance for cod-Elizabethan (which luckily enough I have) and it is very much its own moral/ethical system, which is not what Nietzsche would call "slave morality": some people just *are* better than others, they are higher, nobler, stronger, and the values of that world are Beauty, War, Honour, and Love. Peasants gotta be in the background tilling the ground and serving the lords and ladies, but this is not a meritocracy or a Christian worldview. Everybody loves Lord Gro! I love Lord Gro, C.S. Lewis loves Lord Gro, but he's a back-stabbing schemer who switches sides at the drop of a hat, has no problem with literal dishonourable back-stabbing or poisoning or otherwise pragmatically getting rid of a rival, does not seem to have any discernible principles, and the one scrap of justification given is that he always switches sides to the *loser*, not the winner, because he has some kind of inclination towards the underdog. Which means that he may faithfully serve you as long as you are on the way to power, but once you get power, he's as likely to change over to your enemies as to stick around.
And Dunsany again is somebody different. So yeah, pre-Tolkien, it was all over the place; post-Tolkien, there is a definite model for High Fantasy, whether you follow it or try to subvert it (GRRM is subverting Tolkien, whatever else he's trying to do, it's pretty clear he's not trying to subvert Morris or Dunsany or Eddison or Branch Cabell).
I think your eras might be too short. Gritty urban SFF extends back to the 80s in its mainstream, with titles like Neuromancer. I'm also not totally sure what I would say is the main variant right now or over the last five years. I don't find that the Hugo winners align very closely with Sanderson, and neither the Hugo's nor Sanderson align all that well with my personal taste, but I still get lots of new fantasy books to read all the time. Fantasy also, of course, is experiencing the same loss of gatekeepers as everything else, I read plenty of self published authors and am no longer dialed in to publishers like Tor.
It feels like you're overfitting on too few datapoints to me. To me it feels more like the fantasy genre has a shitload of different subgenres and that occasionally an author that specializes in one such genre becomes really popular than that the genre as a whole goes through different phases.
Unironic fantasy adventures: I don't particular like this subgenre and therefore haven't read many modern examples myself, but wikipedia tells me the last Dragonlance book came out in 2008 and the last Riftwar Cycle book in 2013, so hardly early nineties.
Gritty Fantasy: Don't know if there are any good examples of this genre from before the 90's, but it's definitely still going after the zeroes. The Books of the Ancestor trilogy somewhat overlaps with hard fantasy, but the last book in that series came out in 2019, the Tethered Mage came out in 2017 and The Traitor Baru Cormorant is from 2015. Also technically A Song of Ice and Fire is still ongoing.
Hard and Historic Fantasy: I love Brandon Sanderson, but he's hardly the first person who came up with a detailed magic system. Assassins Apprentice came out in 1995 and while it may not have had as detailed a magic system as Mistborn, it does have clearly established rules and mechanics. Plus, I understand that D&D's Vancian magic system is based on the Dying Earth series, the first book of which Wikipedia tells me came out in 1950. As for well researched historical fantasy, see, for instance, the Deverry cycle (first book 1986).
To be clear, I'm not saying different genres are never created and that genres don't become more or less popular over time. It's just that the fantasy genre is really diverse and as it has grown it has become more capable of catering to different audiences.
Re gritty fantasy: depending on your value of gritty, much of Barbara Hambly's stuff is from the eighties and has a lot of death and nastiness. The degree in medieval history is definitely visible. It's hinted at to a large degree, rather than being portrayed outright, but it's there. The Guardians of the Flame series is in the eighties and makes it clear that a real fantasy world is not a nice place - rape and people dying nastily are things that very much happen. Also an example of the "college students visit their RPG world" subgenre.
I agree with you about my extrapolating from too few days points. I also agree about the "shitload of different styles" part, but it does feel like there are fashions (although I'm not sure and it might just be what I ended up reading from each period). E.g. there might have been a dragon lance book in 2008, but few people were still talking about it.
I also kind of disagree about the specific point of assassin's apprentice. You're right in it having a more rigid rule-based magic system in some ways, but I think their approach to magic is still more old-fashioned unironic adventure. E.g. the Skill isn't so much something with specific rules that characters try to take advantage of, it's kind of "magical telepathic connection that's also in some ways a metaphor for the protagonist's ability to connect to other people", and while the wit has a bunch of specific things it can or can't do, its role in the story is more about the characters connections to animals/nature vs civilization. (Compare mistborn or stormlight, where Sanderson intentionally came up with a technical magic system and then tried to think of how people would use it in practice).
(I like both approaches depending on my mood, but it does feel like fundamentally different styles).
I think mainstream fantasy is dependent on what prior knowledge people have.
If the reading public doesn't have a lot of fantasy knowledge, you get a lot of diverse works.
Once Tolkien's concepts became the one common reference that almost all fantasy readers could be expected to have in common, it becomes hard to deviate without going out of your way to ensure that people know you are deviating. If everyone thinks of Gimli when they hear 'dwarf', you need to go out of way to differentiate if your dwarves are not like Tolkien's (aside from the places where Tolkien doesn't venture, such as the classic subject of dwarven women and beards or lack thereof).
The next generations of fantasy literature (the "gritty era" and beyond) I think derive from a need by authors to do something new to sell the story. It's a lot easier to subvert the reader's expectations by changing the story than by changing the object concepts (such as fantasy races). It also helps that repeated use of the unironic adventure fantasy exposed plenty of holes in the concept for people to use as starting points for decontruction, and once those holes are opened, there's no easy way to ignore them.
To give an example of how changing knowledge impacts storytelling, there's an early D&D adventure that takes place in a dungeon that is actually a long crashed and buried space ship. If this is a new idea that the players are not familiar with, it's probably a fun adventure as the players slowly figure out what is going on. However, once the idea gets used enough, it becomes common background knowledge, and the assumption needs to be that the players may recognize what is going on and how to abuse it right off the bat, especially if the players have given more thought to the concept than the GM.
Aside from Pratchett, I've never been an enthusiastic reader of western fantasy. However, it's fascinating to see the parallels between the development of western fantasy as you've outlined it and the development of the Japanese Isekai subgenre ("Isekai is a subgenre of fantasy in which a character is suddenly transported from their world into a new or unfamiliar one"). While the basic Isekai idea has been around for a long time ("A Conneticut Yankee in King Arthur's Court" and "The Wizard of Oz" both fit the definition), it's something that as a subgenre took off relatively recently.
From my understanding, the ones written when it started to be popular enough to be broken out as a subgenre are akin to the unironic adventure fantasy. I suspect that the main point which drove the popularity of the subgenre originally was that the protagonist was a blatant self-insert vessel for the reader, which made it easy to justify the reader's knowledge of what was going on matching up with the protagonist's knowledge; anything the reader needs explained, the protagonist also needs explained.
Once the reading have enough Isekai stories down, you need to differentiate your stories to make them stand out, and hence you get "dark and gritty" as one way to do this. Part of this is that at some point, since your protagonist is a reader proxy, the assumption is that the protagonist should recognize that he is in an Iseaki. Since the unironic adventure fantasy Isekai cast the protagonist as the hero, one way to challenge this is to have him either depicted as or end up supporting the villain (if not an outright villain protagonist).
The other development that has to go along over time is that Isekai authors will introduce story elements involving modern knowledge (visible back to "A Conneticut Yankee in King Arthur's Court"), and once introduced enough that the audience remembers it, they may start asking about it when it doesn't appear if it would be applicable. For example, someone from the modern era should have at least a basic idea of disease theory (what bacteria and germs are, basic sanitary principles to prevent infections, etc); it's easy enough to bypass needing to worry about this with magic healing to ignore disease entirely, but if some sort of disease is going to play a role in the story, then any modern era character is going to be able to at least make some constructive suggestions (and I've seen several fascinating takes on this very subject). Thus, the stories gets harder over time.
I'm still coming up to speed on everything between Tolkien and Sanderson but I don't find this classification very satisfying.
One trend I've noticed, and enjoyed, in recent years is authors leaving the high-fantasy paradigm and getting their structure straight from earlier sources like myth and fairy-tale. Neil Gaiman does this a lot, and so does Novik's Spinning Silver which is my favorite recent fantasy book. Her Temeraire series, on the other hand, I'd class with Game of Thrones as a setting that's more historical-fiction than fantasy: beyond the basic conceit there's not much interest in incorporating magic with the plot, so it's presenting itself as what's likely to happen in more or less "realistic" terms (though in practice still heavily influenced by the author's preconceptions). "Hard" settings like Sanderson's, or the Eragon series, use magic more effectively but they face a similar problem of sometimes straying into territory that's more SF than F. To my mind, the distinction is that in fantasy the worldbuilding serves the plot, while in SF the plot serves the worldbuilding (and yes, this would classify, say, Star Wars as more "fantasy" than many fantasy novels).
We've had some great recent guests on the futurati podcast.
Episode 29 with Jason 'the progress guy' Crawford was really illuminating. He did a good job of pushing back on my criticisms of progress, talked about how he came around to the stagnation hypothesis, and sang the glories of concrete:
Episode 30 had us interviewing one of the world's foremost experts on innovation in education policy, Andreas Schleicher, and how we can apply those lessons to the U.S.
In Episode 31 we spoke with the unexpectedly funny Brad Templeton, a renowned futurist who told us why we're completely wrong about autonomous vehicles:
Episode 32 had us branching out a little bit by talking to non-futurist Elaine Pofeldt, author of a book on businesses with one employee worth at least $1 million:
I've been running data-analysis/decisionmaking challenges on LW for the last few months, to a small but intensely positive response. I'm posting here in case any SSC readers with a Data Science/Analysis background want to try their hands.
One thing that changed after 1980 is that current-dollar growth rates started having a low US inflation rate tacked onto them instead of a high US inflation rate. Switch to a constant-dollar view and things look a little different.
Talking of solipsism, as somebody is below, reminds me of the story about the professor of solipsism in Cambridge university. I forget the name. He lived to a fine old age, 95 or so, and never married so he had an apartment to himself on campus. The students and faculty took really good care of him, visiting him every day in his retirement, keeping him company and making sure that any illness was promptly treated.
“After all”, they surmised, “ when he dies it’s curtains for the rest of us”.
I wonder if someone can shed some light on why physicist Richard Feynman relied on older mathematical techniques and didn't trust the newer techniques. My question is prompted by this quote from Stephen Wolfram (who was a physics grad student at Caltech while Feynman was a professor there): "It’s kind of interesting to look at [Feynman's handwritten note with calculations related to a Feynman diagram]. His style was always very much the same. He always just used regular calculus and things. Essentially nineteenth-century mathematics. He never trusted much else. But wherever one could go with that, Feynman could go. Like no one else." I assume that Feynman understood how to use techniques like complex analysis and tensor calculus. Anybody know why Feynman didn't trust them? Wolfram's article is here: https://www.wired.com/2016/07/my-time-with-richard-feynman/
I think you have to filter that assertion through the fact that Wolfram is making it, and if there is anyone who has staked his reputation and world-view on New Math -- and especially math that has to be done on a computer -- it would be he.
Complex analysis is squarely 19th century, I think. From context, my guesses about what Wolfram meant are that Feynman didn't like using computers to do calculations, or possibly that Feynman did not appreciate Wolfram's "new kind of science" schtick and its usefulness to physics (i.e. this is just Wolfram's "unappreciated genius" self-narrative).
You might want to bear in mind Feynman invented the most original and fruitful mathematical physics technique to have appeared in the last century or so (his path integral reformation of quantum mechanics).
Feynman mentions in Surely You're Joking, Mr Feynman that he learned calculus from an old textbook his physics teacher gave him, and as a result he approached problems with a "different box of tools." Which got him a great reputation for solving problems because if someone came to him for help with a math problem they'd already tried everything they could think of, so Feynman just had to try something in his toolbox that they didn't know about.
He doesn't mention anything that he didn't trust, but I imagine his approach to solving problems would look unusual to other mathematicians if he commonly uses methods he learned out of an old textbook.
What are the best resources for a balanced, thoughtful perspective on whether an undecided person should have kids? I've seen arguments for both sides - I can make arguments for both sides - but I've never seen a rigorous, evidence-based approach.
Thanks, Ayush. I'd read that one before and thought it was one of Paul's best. He acknowledges his bias in the essay but still offers one of the most thoughtful approaches to the topic.
What kind of rigourous, evidence-based approach would you like?
There's certainly studies looking at (e.g.) self-reported happiness of parents vs non-parents, but the results you get depend heavily on exactly what question you ask and when you ask it. Having kids is a matter of trading off most of the good things in your life, in exchange for other good things.
Thanks, Melvin. Having read some of the research that compares self-reported happiness of parents and non-parents, parents appear to be less happy than those without kids but report higher levels of meaning.
This seems like a mixed bag - as you'd probably expect - yet the social expectation is that most families will want or have kids. That feels heavily skewed for what I perceive as a fairly balanced risk/reward profile.
I guess I'm trying to find something that can thoughtfully connect these two facts and offer a holistic view of whether the benefits are worth the drawbacks with something less subjective than "When you have kids, you'll understand they're the greatest," or "I pity all the families having to corral their kids on a hot summer's day while I eat ice cream and have infinite freedom."
I have found meaning far more important than happiness in my life. Before getting married, I was honestly just drifting for lack of motivation. I played a lot of video games and probably would have said I was happy, but looking back I wouldn't take that again at all. I am probably slightly less likely to say I'm happy now, but I much prefer life with a family.
I guess what I'm saying is make sure you're asking the right question, and understand the implications.
You're asking the wrong question. The main argument for having kids is that it's what we're hardwired for in terms of maximizing our personal life meaning. I guess I could try to kludge together some objective measure of meaningness, but that's besides the point.
Having kids is the most subjectively rewarding experience in life. Demanding rigor and evidence is completely missing the point.
Fair enough, Nick. I suppose the most important thing I'm trying to find is a balanced position - because everyone is necessarily on one side of the fence or the other, folks usually pretty tribal on this topic. If rigor and evidence aren't the order of the day, I'd at least like to see a balanced argument. Do you know of any of those out there?
I have a tough time imagining one. Some people don't seem to derive much meaning from kids, or don't pay attention to them and then end up disappointed. There's a lot of room for positivity and growth, but a lot of room for negativity and failure. Your experiences will vary dramatically from everyone else you ask; I sincerely doubt the question is answerable in a way that, after having kids, you would look back on as being useful.
On the meaning side, I can't really overstate it though. Prior to having kids I had occasional moments in my life that touched me deeply. Campfire moments, I think of them as, because a lot of them occurred around fires with friends. Probably a few dozen altogether over my first 33 years.
Since having kids it's a constant, daily and nightly stream of moments that put all those to shame. When I say it's what we're hardwired for I mean that in a way that doesn't make sense until you experience it. There's nothing else to compare it to.
All the rest of the pluses and minuses are things that can be written down on a pros and cons table, and compared, and balanced, and argued about. You're probably already familiar with everything that can go on that list. Financial costs. Teenagers. Grandkids around your bed when you die. Stuff like that. The one piece that I think gets missed is the meaning piece, because it's impossible to convey from people with kids to people without them.
The good news is that, whatever you decide, you'll be happy. In one case it'll be because you tapped a deep well of human experience that you never would have otherwise known, and in the other, it's because you didn't and you won't miss what you never experienced.
As a counterpoint, I haven't found life with kids any more meaningful than life without. It's more fun, interesting, etc. and I wouldn't trade it for the world, but I haven't experienced a perceptible increase in meaningful moments the way you describe.
Maybe an interesting way to rephrase the decision - and make it a bit more tractable - might be ‘how will having kids be for ME’?
That way you have 2 ‘smaller’ but maybe more objective questions to answer: what is my personality/preferences, how do kids affect one’s life typically, and so to what will that ‘add up’?
Like (hey, just thinking out loud :)): try taking the free mbti/personality test at 16personalities.com and then read until the ‘parenting’ section, which might make the kids question a bit more tangible to reason about?
From what I hear (and extrapolate from when I was a kid) they're great at first, then terrible, and then once they're 26+ they become great again, if you raised 'em OK.
(I wasn't "a teenager", have gotten on very well with my parents my entire life, and was raised by people who viewed kids as miniature adults, and treated us as maturely as we acted. Anecdote isn't data, but...)
I’m considering quitting my job and going to a bootcamp to learn data analysis/ML/AI. Anyone know of a good program in the NYC area? Preferably with an I person option.
Would love to read a deep dive into the prospects for GLP-1 receptor agonists as weight loss medications. Semaglutide 2.4 mg generated some buzz recently, and now apparently there is Tirzepatide. Has anyone tried semaglutide or liraglutide, or prescribed them to a patient? Lots of questions about side effects, long-term use, cost/insurance coverage, etc. But these trial results seem remarkable. Are we on the verge of having obesity treatments-other than bariatric surgery-that actually work?
Some pillows last longer than others, and lifespan correlates reasonably well with price. But if you're asking what is the most comfortable pillow, I'm firmly of the opinion there's no such thing. What's comfortable depends entirely on your preferred sleep position and what feels good to you. I've had all kinds of pillows -- down, form, fiber-filled -- cheap and pricey -- soft and firm -- and my current set-up is the most comfortable I've ever had. It's a fat couch throw pillow, topped with an ancient, deflated-looking pillow that seems to be filled with crumbed foam. For me it's perfect, and I'm prone to neck cricks and general pillow discontent, so that's saying a lot. Just experiment around.
I've had a pretty good experience with Allerease down alternative pillows. They're the right mix of firm and soft for my head.
One caveat is that I've noticed the down alternative pillows don't keep "cool" as well as regular down ones. If you like having a cool pillow against your head or face when you lie down and try to sleep, the down alternative pillows will warm up faster.
Let me be more specific: anyone have a recommendation on a thick but soft pillow that whisks heat away from my head and stays cool without being too firm?
OP, I suggest describing your perfect pillow and seeing if anyone can recommend a model. Cool idea though.
Almost orthogonal to your question but hopefully useful: my wife got some silk pillowcases and they are wonderful. They are a little more work to wash, but we each have two pillows and only the "top" one of each pair is wrapped such, reducing the workload.
Thanks for the replies just wanted to see what recommendations people threw out. It just seems one of those products where there’s loads of options and it’s hard to pick.
Agree that it’s pretty individual but that down/down alternative is tried and true. And apparently I should dress it up in silk ;)
I am extremely happy with it. Best purchase I've made since my Allbirds. I sleep through the night much more often now. After I got one, my wife tried it for one night, and we immediately bought one for her as well. If I thought they were no longer going to be made, I would try to buy enough of them to last for the rest of my life.
There, he acknowledges that the response to Covid appears to be a counterexample to his thrive/survive model. However, in the comments on that open thread, several posters point out that since Covid mostly endangers the weak/elderly and isn't an existential threat to humanity, it actually fits in the model that Republicans would be willing to sacrifice the weak and elderly so that the rest of us can continue to thrive, while Liberals are more willing to let the economy take a hit in order to save them.
(Also, it's worth noting that Republicans tend to live in rural areas and are more likely to have jobs that cannot be done from home, so Liberals are both more affected by the pandemic and less affected by the lockdowns.)
There's two different definitions of "urban" here. The definition that yields the US as 82.5% urbanized counts basically anything where residences are organized (possibly even everything within an incorporated community as well).
But the definition that lines up well with the political divisions is tripartite, into "urban", "suburban", and "rural". The way I've heard this divide jokingly described is: if you can step outside naked and no one sees you, you're rural; if you can step outside naked and the neighbors call the cops, you're suburban; if you can step outside naked and the neighbors ignore you, you're urban.
Everything "urban" or "suburban" counts as "urbanized" in the former definition. "Urban" areas usually vote 80% for Democrats (though some parts of many cities are as low as 60%, and there's often a few rich neighborhoods that vote for Republicans). "Rural" areas usually vote 80% for Republicans (though there are a few regions of Black or indigenous populations that vote for Democrats). "Suburban" areas can go either way (often depending on population density).
(1) Survive vs Thrive only claims to be maybe 80% accurate and this is one of the times it fails.
There might be evidence for this internationally. Trump and Bolsonaro made the news for having lax covid response and bring right wing. Abe is right wing and had a strong covid response, as did some central European countries. If there isn't a correlation between how right wing a country's government is and how lax their covid response is, then this is probably coincidence. The alternative history where Ben Carson became president instead of Trump is also interesting to consider here.
(2) The Right looked at a 1% mortality rate and thought it was nothing to worry about. The Left looked at a 1% mortality rate and decided it was a crisis. I'm not sure if the people making the decision were numerate enough to think this way. Would people's responses be different if covid had a 5% mortality rate?
(3) The Left's support for and the Right's opposition to government agencies was more important than Survive vs Thrive. This also seems to be what's going on with school choice. Maybe Survive vs Thrive works less well in the short term or for particular issues and better for grand strategy questions or issues that persist for decades.
Has anyone had any experience playing a digital piano keyboard (one of the 88 key models with a weighted key action such as Yamaha, Casio, Roland, Kawai etc make) on top of a regular office desktop with the height adjusted down to 24 inches or so, rather than on a dedicated stand? I'm debating whether to try and refactor my sit/stand desk setup to include my piano-- I compose with piano and computer frequently and it'd be nice to do it with both of them at the same adjustable height. And it seems like the key height should be OK in "sit mode" if I adjust it all the way down to a 24 inch desktop height, but I don't know whether I need to worry about instability of the keyboard on top of the desk, or resonances with the desktop, or some other such complication.
> I compose with piano and computer frequently and it'd be nice to do it with both of them at the same adjustable height.
Do you have your piano literally right next to your computer? If so I can't foresee any harm from making them the same height if you do it right. (As for resonances or whatever, if the piano + supports and computer + supports aren't touching each other, there shouldn't be much of an issue.) Otherwise you should keep them as is because your hands will have plenty of time to adjust to a different height while you move and changing the height of a piano or computer desk is generally hard.
The tax code was much more favorable to you if you started a company in the Postwar Era, versus getting rich off regular income. Same time that the top tax bracket was 91%, the capital gains tax rate was 25% AFTER excluding 50% of the gains from taxation altogether. You'd see Hollywood types try to turn their income into that to reduce taxation, by using stuff like setting yourself up as a corporation selling shares of stock (with said company "leasing" your services) or the Oil Depletion Allowance (IE you exclude something like 28% of the income from investing in Texas oil wells from taxation).
Regarding the growth slowdown, a book that I think does a good job of painting the big picture is 'Fully Grown' by Dietrich Vollrath. According to him: mostly demographics as well as the shift to services where productivity growth is inherently slower. His biggest takeaway is that none of this is inherently a bad thing.
Anyone looking for a 'Priest' to the Rationalist 'Wizard' are encouraged to check out Mr. Tom Murphy at https://dothemath.ucsd.edu/ An Astrophysics professor come.... neomalthusian (?), he is back to blogging in support of his new textbook on economics and the physical world which can be found here. https://escholarship.org/uc/item/9js5291m#article_main
Longtime rationalists will enjoy spotting the motte and bailey arguments, the strawmen, assuming the conclusion, and a host of other illogical and sometimes bizarre reasoning.
On the plus side, his blog does allow comments and he frequently engages his readers, so if you are interested in trying to argue the Wizard side, you might get some traction there.
When publishing predictions, Scott always mentions that some people consider 50% predictions meaningless as they indicate lack of information. Could someone please explain the reasoning behind assuming it is meaningless? Because, suppose I have a playing die, and believe it is biased so I expect it to roll up the number 6 with 50% chance (and have no idea about relative chances for the numbers from 1 to 5). But this clearly means I have some information, and if it is true, I could benefit from betting on it against someone who e.g. thinks the die is fair. So what's the reasoning behind the opposite view?
I think it's based on an idea that everything is either true or false, and so there's a natural two-element partition for every proposition. (I may not be entirely charitable here, because I'm not sympathetic.)
But I think the particular discussion at hand is based on the fact that he evaluates his predictions based on "calibration". Instead of measuring how close each *particular* prediction is to the truth (and perhaps, comparing to someone else's prediction on the same proposition), "calibration" says to group together all your X% predictions and see how close you are to X% true. Scott usually does it by grouping together the X% predictions *and* the negations of the (100-X)% predictions, to get a larger dataset. But when you do this, you automatically group together the 50% predictions and the negations of the 50% predictions, and thus automatically get 50% correct. (Though I think for this one particular bin, he just chooses the 50% predictions and not their negations, which then invites the question of how he chose which one of the two to be the prediction.)
Scott does yes-or-no predictions, so a 50% prediction does no better at helping others find the correct answer than random chance. Even if you have information saying the chance is 50%, you perform no better than someone with no information at all.
This really depends what your priors are. If you have literally no information (and use a uniform distribution as your default prior), then a 50% probability gives you no additional information. But most people have some information already, like a prior that a 6-sided die is at least somewhat fair. In this case, a prediction that it lands on 6 with 50% probability is meaningful.
He also sometimes fakes a probability distribution with predictions, like the chance that covid deaths are above 10/100/1000k. In those cases, 50% acts as another data point.
This isn't the same thing. Scott is predicting binary outcomes with a probability of 50%. The equivalent on a six-side die would be predicting 6 comes up with 1/6 certainty. But note that, in this case, you can still calibrate. If you make this prediction across six different die and they all all come up six, there is a decent chance you're wrong about die in general and need to update that some common manufacturing might exist in all of them making 6 more likely.
The issue Scott runs into is really specific to binary outcomes. Say he makes 4 predictions. He's 50% certain that Covid deaths will be under 500,000, 50% certain that the US murder rate will be under its 30 year low, 50% certain that Biden wins the election, and 50% certain that GDP growth is less than 5%.
How does he calibrate? By this phrasing, he got 100% correct, not 50%, so he's miscalibrated. But phrase two of those the other way (50% certain Covid deaths will be over 500,000 and 50% certain Trump wins) and now he got 50% of them. Perfectly calibrated.
When you can be either perfectly calibrated or perfectly miscalibrated making exactly equivalent predictions simply depending on how you choose to phrase them, it isn't that the predictions are meaningless. It's that the attempt to use them for calibration can't work.
Minor complaint, when I post I'll sometimes not be able to continue on ACX, and I'll have to reload it. (Which means I lose my place in the thread stream.) I think this is related to my keeping my mouse over the post button after posting. A little slashed red circle appears. And if I leave it there, it stops working. Does this happen to others? I've got laggy satellite internet.
I'm reading the latest Michael Lewis book, *The Premonition*; it's a post-mortem of the US covid response. I've only just started, but so far it's great.
His last book, *The Fifth Risk*, was published in 2018, which among other things was about how the Trump administration had crippled the United States' ability to deal with catastrophic risks. Turns out he was right.
But, as he says in the introduction of this book, it turns out that the rot spreads much deeper; "Trump was a comorbidity".
Right now it's just introducing some of the characters and what they were doing before the pandemic. One of them is a public health official in Santa Barbara, Dr. Dean, and the book goes into some detail about what that is like — and what the CDC is like. One example of many: Once there was a patient, a college student, who had symptoms of meningitis B and was in a state of shock (they had to amputate both his legs to save his life). Dr Dean set out to control the outbreak — and it's astounding just how useless the CDC was throughout all of it. I don't think blockquotes markup works on Substack, sadly, but here:
> She got on the phone with the main guy at the CDC and his silent crowd. The guy strongly disagreed with her doing anything. "What he actually said," recalled Charity, "was, 'That decision is not supported by the data.' I said, "Oh really — there is no data." She outlined a plan she'd created: thin out the dorms by moving some of the students into hotel rooms; shut down the intramural sports teams; and administer a vaccine that had been approved in Europe but that the FDA had not yet signed off on. "The CDC guy said, 'We're not going to do any of that, and if you do that, we're going to put it in writing that it was your decision and we disagreed with it,'" Charity recalled.
> [...] But in the end the campus ignored the CDC and did everything Dr. Dean recommended. From start to finish, [...] "the CDC wasn't pleased with her. The CDC kept saying 'There is no evidence to back it up.' They didn't have any evidence, because there is only one case every four years."
> The root of the CDC's behavior was simple: fear. They didn't want to take any action for which they might later be blamed. "The message they send is, We're better than you and smarter than you, but we're letting you stick your neck out to take the risk," said Charity.
> Charity never would know which of the measures she took had controlled the disease; she knew only that all of them together had. [...]
> Two years after the UCSB meningitis outbreak, the CDC finally published a report on how to deal with a meningitis outbreak on a college campus. On its list of best practices were most of the things Charity had done at UCSB. [...]
> Charity had washed her hands of the CDC. "I banned their officers from my investigations," she said. The CDC did many things. It published learned papers on health crises, after the fact. It managed, very carefully, public perception of itself. But when the shooting started, it leapt into the nearest hole, while others took the fire. "In the end I was like, 'Fuck you,'" said Charity. "I was mad they were such pansies. I was mad that the man behind the curtain ended up being so disappointing."
Seeing as "listen to the experts" is a common refrain, it's important to remember that *credentialed, educated, powerful, and authoritatively titled* is not the same thing as *expertise*. Those men at the CDC were credentialed and titled. Dr. Dean is an expert.
It's strange, people are doing to the CDC what they did to FEMA about 10-15 years ago. Neither agency was ever meant as a first-line responder to emergencies, for the simple reason that (domestic) emergency response is a state and local responsibility, that being the nature of a federal republic. There's nothing in the Constitution that gives the Federal Government the responsibility *or* power to do emergency response (except for external threats, e.g. invasion and war, which the military handles). So it's been a state and local responsibility for ages.
Until Katrina. Then all of a sudden people thought this Federal agency which existed to disburse emergency loans and write down best practises for sharing among agencies (FEMA) should *be* a front-line responder, and were shocked that when asked to be that, they were awful.
Same thing is happening now to the CDC. They were never meant as a first-line responder to pandemic, that has always been understood to be a state and local responsibility, like all other emergency response. The CDC collects data, mostly, does some research, hands out grants for research, and writes down best practices (long after the fact) for sharing. They were never meant as a first-line responder, and that of course is reflected in who they hire and their institutional reflexes (well illustrated by the anecdote).
And yet...people are now acting as if they *should* be a first-line responder, and finding it outrageous that when pressed into the role, they're terrible. It's strange, it's almost like people have forgotten that we live in a federal republic, with some pretty strict restrictions on what the Federal Government *can* do.
There's certainly an argument to be made that for pandemics we *ought* to have a national-level first responder, since disease doesn't respect state boundaries, but, tricky Constitutional questions of legal authority aside, this is not what the CDC is, or ever was.
(1) The CDC should provide accurate information about the control and prevention of diseases. It instead provides inaccurate information.
(2) The CDC sure thinks they should be a first-line responder. Remember when they made a coronavirus test kit, and the FDA banned every test kit (including the WHO's) except theirs — and then the CDC's was defective? That's not unfair societal and political expectations being *imposed upon* the federal government, that's the federal government fucking things up and accidentally killing people.
In the recent book review "The Wizard and the Prophet" I noticed almost no pushback in the comments from people with a more radical environmental perspective, i.e. people who take a non-human centered view of the environment.
Is this a perspective that is just not common in the rationalist community? Because it seems like a very important addition to the conversation. The 'Wizard' community has many successes if you take a human-centered approach, but if you expand your scope to the natural world as a whole I think it is clear we are not innovating our way out of environmental damage.
Examples:
1. Deforestation continues on a massive scale to this day. The rate has declined some in the past couple decades, but rainforests and other habitats are still being destroyed daily at an astounding pace.
2. Evidence of population drops for insects appeared last year. The data was likewise astounding, claiming upward of 1-2% population drops per year, or 30-40% declines since the mid 20th century. There is not enough data to be confident about how this is taking place, whether it is concentrated in certain areas/ecosystems or more widespread, but that a problem exists seems fairly undeniable.
3. Overfishing is still occurring worldwide. Some specific commercial populations have recovered when managed, and some countries are practicing generally responsible fishing, but overall it is still a large issue.
4. Climate Change, which was cited as perhaps the 'Prophet' side's best issue, will likely continue to exacerbate all the issues which we are already creating throughout this century.
5. Plastic and other pollutants in the ocean are having other large effects on marine ecosystems.
I suppose a 'Wizard' point of view could point out that many of these problems do have technical solutions in theory and even in practice in some locations, but I think the 'Prophet' would counter that just because solutions exist does not mean they will be applied, and that they in fact will not be until a large amount of damage is done. An irreversible amount of damage.
I just really didn't like that book review, or it's lens for approaching our interactions with the planet. We are having a profoundly negative effect if you look beyond the human view.
As far what is or isn't common in the rationalist community, I think this adversarial collaboration Scott hosted on his old blog is a good sample of how a lot of rationalists approach the problem of valuing life generally.
The main takeaway is that the paper takes a long time trying to determine what animals are conscious, and how conscious they might be. One takeaway quote:
"Of particular relevance is the conscious experience of pain and suffering, which we regard as morally undesirable when it occurs in ourselves or others. Most animals have damage sensors, but triggering these may not result in the subjective experience of ‘suffering’ if the animal is not conscious or the stimulus is a constant presence that has been accustomed to.
And this is the absolute crux of this investigation; if animals suffer under current farming standards to the point of preferring non-existence then there is a moral burden on meat eaters to justify eating them."
It seems to me (I'm not a rationalist, I just like to hang out with them) that most rationalists are utilitarians. If they seem human-centric it's likely because humans are the only creatures whose happiness or suffering we are intimately acquainted with. As such, a topic such as overfishing is only really relevant if you can show that it would hurt humans, since as far as we can tell fish likely aren't conscious and might not even feel pain in a way that is meaningful for a utilitarian calculation.
I'm not a utilitarian, so I can't really speak further on the specifics.
I wonder if we can get any of the Substack people to explain their expectations for what it would be like? Their Help page really isn't; it's geared towards writers (for how to publish on Substack) and the section for readers is all about "how to pay for a subscription".
It doesn't appear to be the case that they anticipated lively debate in the comments sections, so no edit, link, formatting, etc. functions. That's why I'd like to know what the mental model coming over was - did they think of it more as a newspaper online page, where subscribers can leave comments below an article or opinion piece, but nothing more and certainly no kind of conversation (and where the paper or media owners can delete or close down comments they don't like)?
No joke--I've been wanting to visit India and Brazil. Based on the likeliest trends, when will COVID-19 levels decline enough in those countries for me to be able to visit them and not find half of the tourist attractions and restaurants closed?
Had mild success with posting job ads in an open thread before. I know back then this was accepted, not sure if it is anymore (if not, please delete, and my bad). Anyway:
I'm curios if anyone here would like to lead an infrastructure/ops/devops team (starting as 1 man job, will hopefully soon grow into a 2-3 people team if all goes well).
The product is something called mindsdb, fully open source "generic" machine learning. The thing you'd be working on is our cloud offering (so not open source, but running 98% the same code as the open source version)... which is rather important, since it's how we can actually afford to develop a fully open source product long-term.
Fully remote, offers start 80k/year but the cap is much higher (basically low pay for the US and insanely high pay for literally anywhere else in the world except for the trade-eco hubs), work contract is whatever (employee via local company branch, contractor, working via a 3rd party company... we don't care really, unless you're US or UK based, in which case I think it has to be a standard employment contract). 3-5 meetings a week (total 2-4 hours) + slack for other communication, people aren't sticklers for immediate answers but ideally you an be kind-of online around EU and/or US friendly working timezones.
Please ping me if it sounds interesting, it's honestly much harder to find people that can to devops tham I thought :)
If the Theory is right, then there is still a way for one alien species to take over the galaxy without attracting attention. It would involve a "two-phased attack" with self-replicating nanomachines. An alien species could, over a long time and with great secrecy, seed every solar system in the galaxy with its own Von Neumann probe, which would contain self-replicating macro- and nano-machines. Once every solar system had a probe, the aliens would send out a signal, and all of the probes would start self-replicating. They wouldn't just make "Gray Goo" copies of themselves--they might make soldiers, weapons, and other advanced technology. Any enemy aliens would be overwhelmed, or at least forced to reveal themselves to fight back.
This could be done without revealing the planet the probes trace their origins to. One day, it would seem like the whole galaxy suddenly came alive at once.
If this theory is right, then we should also not expect to hear back from any alien civilizations, and also we should be even *more* scared about the prospect of there being any.
(Honestly, given the multitude of proposed solutions to the Fermi paradox and our extreme lack of concrete data about it, I'd put very low weight on any hypothesis. The solution I like the best, though, is still https://arxiv.org/abs/1806.02404.)
Cryptocurrency/DeFi/etc has been on my metaphorical radar for quite some time, but I feel hesitant about investing/working in it. Here are some of my worries, in (very roughly) descending order:
1. Coinbase/Binance/etc doesn't allow people under 18 to join and trade (which is think is for legal reasons), and I'm one of them. (And, more generally, the legality of using cryptocurrencies/DeFi/etc.)
2. Even if I don't use a company like that, I might need to install special software in order to do this properly, and I'm afraid of messing up my computer beyond repair in the process.
3. Not everyone accepts cryptocurrency as payment for commerce.
4. I'll have to pay taxes on it, which I don't know how to do properly.
5. There are a lot of criminals involved in this area.
6. Someone recommended (https://www.youtube.com/watch?v=sEtj34VMClU) that I use relatively extreme security measures before even beginning (such as writing passwords in a paper notebook, using 2FA everywhere), and I have no sense of calibration on whether this is too much or not enough.
7. Hackers/scammers/ordinary market fluctuations could wipe out any profits that I make.
8. Mining cryptocurrency takes a lot of energy, which produces a lot of environmental damage. Renewable/zero-emissions energy sources could mitigate a lot of it, but it's not exactly straightforward to find or obtain.
9. I don't understand any economics.
Can you clarify whether these issues are as important as I think and recommend ways I could take care of the important ones? Thanks.
I too am interested in this. FYI, here is a link to the article in question:
https://thebulletin.org/2021/05/the-origin-of-covid-did-people-or-nature-open-pandoras-box-at-wuhan/
This was the best summary from a virologist that I've seen:
https://www.reddit.com/r/science/comments/gk6y95/covid19_did_not_come_from_the_wuhan_institute_of/
Of course, none of the points raised are slam dunks as they later admitted by posting what it would take to re-examine the conclusion:
https://www.reddit.com/r/science/comments/gk6y95/covid19_did_not_come_from_the_wuhan_institute_of/fqpci3n/
Basically, evidence of shenanigans of samples or workers at the Wuhan lab would change the conclusion, but China has obstructed all attempts at independent investigations that could find such evidence. I think this is why the lab leak theory is still compelling.
Thanks for those links! A couple of points in no particular order. 1) This redit thread is listed as being 11 months old, whereas Wade's writeup is.. i think from May 5th, so dont expect any direct refutation of Wade in the Redit series.
2) the Redit series actually cites (pretty unconvincingly, IMO) the WIV's history of transparency as proof that it didnt accidentally release COVID:
"And I can hear someone out there, shouting into the darkness… “but it happened in 2004!...It happened in 2008!” (117,118)
Yes, but you know what’s interesting about that: those events are part of why this probably wasn’t a lab accident. We know about those events, because scientists (including some Chinese ones) weren’t interested in covering them up (117,118,119).
Why would they cover it up now? Why would they behave differently than they did in 2004 or 2008? Why would these Chinese scientists reverse course on being honest?"
The Redditor doesnt mention the 1977 Russian H1N1 outbreak which was both most likely an accidental release and covered up by scientists and officials at the time.
I too was unconvinced when I read it a few months ago, on balance of evidence. I don't think he was claiming past transparency proves transparency today, merely that it's evidence in favour of.
Then again, past leaks were relatively minor compared to SARS-COV-2, so perhaps by the time this leak was detected it had already spread around the world. Also, past leaks happened under different government administrators, so perhaps the current administrators had a different view on the PR nightmare this would have caused.
This published works covers all the details of how SARS-COV-2 presents many challenges to claims of natural origins: https://link.springer.com/article/10.1007/s10311-021-01211-0
That would be my assumption as well. A leak that amounts to nothing much is substantially different than one that kills millions.
Side note, this link discusses known accidential lab leaks, worth a read for those interested.
https://thebulletin.org/2014/03/threatened-pandemics-and-laboratory-escapes-self-fulfilling-prophecies/
Boy, I wish people would just write essays and not hack them up into tweets. That was too hard to wade through.
But to judge from the first half dozen bites, he seems to be rebutting the theory that it was an engineered biological weapon, which Wade concedes from the start.
Since it's long, let me summarize the points of the virologist Duehr that I found most important, and comment on it:
- The largest part of it argues that the virus has not been artificially created from a known backbone like RaTG-13 (i.e., from a virus whose sequence has been published in a previous publication). For example, all the sections of mutation speed etc. hammer this point in. I think both sides agree that this is convincing. But the argument is still weak; the pro-lab counter is that most virus sequences (including most virus sequences from the Wuhan lab) have not been published, and the argument collapses if the backbone was unpublished.
- The strongest argument contra-lab from Duehr is that "Dr. Shi’s projects almost always provide duplicate samples of every bat to both Chinese scientists and laboratories in other countries". This is a strong case since it would make it difficult to conceal the backbone if the virus really came from the lab. I am not sure whether I should read the "almost all" part of the statement literally. Did they really collaborate with other labs on all their projects? We are talking about 400+ viruses and 1500+ strains. (From the link Sandro posted in this thread.) Still, it makes it less likely that the compromising information from the lab hypothesis would only exist in Wuhan, and I personally find a conspiracy implausible if it goes beyond a single lab.
- Duehr claims that the most plausible origin of the pandemic is outside of Wuhan, possibly far away. I haven't heard this argument in recent discussions, so that claim might just be outdated, but I am not confident.
- Duehr also claims that similar experiments happen all the time, at many places in the world. So having an experiment nearby the origin of the pandemic would not be surprising. Other contra-lab paper have also made this claim. As far as I can tell, this is factually wrong if we define "similar" as "gain of function in vivo in humanized animals". Duehr provides four papers as evidence, claiming that similar experiments were done there. Another contra-lab article provides another four papers. I have checked all eight, and "gain of function in vivo in humanized animals" happens in exactly one of these articles, and this is an earlier paper from Dr. Shi's group in Wuhan. Right now my hypothesis is that Wuhan is/was the only lab in the world to perform such experiments. (If anyone knows better, please tell me. How many labs do you estimate were there?) I may seem picky about the definition, but I think it is important. Removing any of the conditions makes a lab accident much less likely: if it's not in vivo, there are no airborne infectious particles; if it's not in humanized animals, then the virus is optimized for a different species than humans.
Good summary, but I have to take issue with the following of Duehr's claims:
* Duehr claims that the most plausible origin of the pandemic is outside of Wuhan, possibly far away: it's plausible if you already assume a natural origin, which is begging the question.
* Duehr also claims that similar experiments happen all the time, at many places in the world. So having an experiment nearby the origin of the pandemic would not be surprising: err, what? It's not "nearby the origin", it's ground zero as far as we can tell. Big difference. Further, even if these experiments were happening all around the world, if an outbreak begins in the same city as one of these labs, you better believe a lab leak is more plausible than if the outbreak happened further away.
Yes, good objections. Though for the second one, there are two numbers floating around: if I get it right, the lab itself was 14km away from ground zero, and there was a building of the institute only 600m away. But if I get it right, the experiments did not happen in the nearby building, so I would count the lab as "same city" instead of "ground zero". But you are right: even if it was 50 labs worldwide and we count it only as "same city", it would still be remarkable.
It still makes a difference to me whether it is 50 labs or whether Wuhan had literally the only lab in the world where such experiments were done.
I believe it was literally the only lab in China where such experiments were done. Or at least that it was the only lab in China where such experiments were supposed to be done. There are other labs elsewhere in the world, most in countries with a better reputation for enforcing safety regulations. And part of the hypothesis is that the research was being outsourced to China, in which case "...because it was faster and cheaper, because Chinese regulators turn a blind eye to the fact that you're not really doing the tedious BSL-4 stuff" is at least plausible.
Hm, but do you actually know a lab or a research paper? I have heard the claim "similar stuff was done elsewhere" before, but whenever I tried to actually find such a lab, I couldn't. As I said, some people gave references, but the ones that I checked were not "similar". At least not in the sense "gain of function in vivo in humanized animals" that you need to have a really dangerous situation.
Actually, according to Duehr, the virologist community thinks that the Wuhan lab was as safe as labs elsewhere, and I buy his arguments. You may want to read his position on this.
The worst thing in that respect is that the experiments were at least formally allowed in BSL-2 environment (if they were derived from some other bat virus than SARS and MERS). Which had nothing to do with the location, it was made by the funding agency in the US.
I'll try and steelman GoF. (which I think should stop.) By studying how viruses change/ evolve to a new host we learn something that helps block against the next pandemic.
It might. I think the point of research is that you don't always know where it will you lead, if you did then you wouldn't call it 'research'. So you through a lot of shit at the wall and sometimes those recombinations get you very effective treatments like (immunotherapy or something).
Anyway, you say that one particular line of research is verboten or restricted and there would be an effect. I'm not saying it would kill the whole field of virology or anything, but if you look at is say there are 100 possible treatments for the next pandemic. If you institute this ban then maybe that goes to 50 or 20 or whatever 100/N is where N is unknown. It may be a relatively low so you still have dozens of effective therapies to choose from (like us right now)
Or it may be higher in which case you might have just one potential, and then things would be not so great. /Steelman off
Google Scholar is always worth checking; no need for institutional access or LG, as GS points right to https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7787577/
Ah thanks! Will delete and repost.
For folks who like philosophy, I’m writing a new feature every two weeks on a thinker I think matters today. https://mobile.twitter.com/ZoharAtkins/status/1372675033336778755
Did you recently read Time of the Magicians?
No. Good? I liked Peter Gordon on Heidegger and Cassirer debate. I wrote my dissertation on Heidegger
I started it but still have too much lingering pandemic brain fog to get deeply into it. I find German-translated-to-English takes a lot of mental power. Just oddly heard about it the other week immediately before seeing your Cassirer tweets.
Seeing as you like both Heidegger and Wittgenstein, have you read "The Master and the Emissary" by Iain McGilchrist?
Also curious what folks make of my hypothesis that experts vs. populists is an investiture conflict 2.0 https://whatiscalledthinking.substack.com/p/investiture-conflict-20-experts-vs?r=8nz8&utm_campaign=post&utm_medium=web&utm_source=twitter
I very much enjoy your twitter mega threads. I think your experts vs populists points are interesting, although I think the distinction that Robin Hanson made between Experts and Elites (https://www.overcomingbias.com/2021/02/experts-versus-elites.html) may be helpful here too, as it reminds me a lot of what I see coming from the grey tribe critiques of a lot of the pandemic response (Balaji Srinivasan's critiques of bailouts and push for DeFi come to mind).
I'd also really encourage considering structuring paragraphs/sentences that end with "2.0" differently as figuring out where full stops should go added some unnecessary mental load. Maybe replace with "v2" instead?
As the unofficial unofficial associate (maybe) podcast of Astral Codex Ten I'd like to mention the latest episode with this blog's readers' favourite Bret Devereaux (of This Isn't Sparta! fame) in conversation with Professor David Abulafia about the ancient Mediterranean. They talked about Alexandria, exactly how may rowers these insanely huge galleys had and much much more. (I see from Zohar Atkins who just beat me to being first that I am not the only one pushing my own stuff . . .)
https://www.buzzsprout.com/207869/8481869-abulafia-and-devereaux-the-ancient-mediterranean
Excellent! I love Bret's work and have recently hit a gap in my podcast queue, downloaded. If this is going to be the plug-yourself thread, I may as well link my "rational-adjacent, 2 dudes talking" podcast Affix, latest episode discussing Robin Hanson on Experts vs Elites (which I also linked to Zohar earlier in the thread):
https://www.affix.live/1518706/8469814-episode-18-elites-and-experts-we-definitely-manage-to-be-neither
I'm going on vacation for a week and could use book recommendations. I'd prefer some lighter non-fiction; I've been doing dense mathematical reading for the past while, and would like to get away from that for a bit. Ideally I'd like something entertaining, well written, and about a non-mainstream topic (one such book I enjoyed was The Deadliest Enemy by epidemiologist Osterholm and writer Olshaker).
Thanks in advance!
Oh to note, I read Deadliest enemy a couple years pre-covid, so before epidemiology was such a big talking point
Batavia's Graveyard by Mike Dash. Eye popping stuff. Tom Holland's Rubicon is his best book and he is a very fine historian!
Skimmed the blurbs, they look interesting. I’ll git them a try, thanks!
I enjoy Erik Larson’s books. They seemed suited for what you’re looking for here.
I checked out his stuff, it does look about like what I'm looking for. Thank you
That’s great. Enjoy! You can’t go wrong with any of his books.
Strongly recommend The Splendid and the Vile. A page turner even though you know how it turned out 😊
Always tempted to recommend _Passions Within Reason_ by Robert H. Frank, but I noticed that I don't really know what a mainstream non-fiction topic is. Does "economics and evolution brought together to explain several human emotions" qualify as mainstream? If it does, this book won't pass your criteria.
I'm not sure either, but I have studied a bit of econ related to that stuff, so while it might be outside the mainstream, I'm not sure it's far enough out of my familiarity. Thank you though
Memoirs of Hadrian by Marguerite Yourcenar. A truly great piece of writing that somehow never makes the lists.
I'll check it out, thank you
On an unrelated note, would you have recommendations for books that are math-heavy?
Oof, this is a tough one. It's almost an axiom that all math textbooks are bad (with Spivak's Calculus being the only exception I can think of). Maybe that's only true of the core authors like Apostol or Stein, but I'm also not sure you're looking for something like that.
Personally, I tend to try to read things outside of my area (or just fluff fiction if I'm being honest), but one book I would strongly recommend is Logicomix, which presents an excellent biography of Russell as well as capturing one of the pivotal struggles in mathematics. As the name suggests, it's a comic book, so it's not a heavy read, but it's definitely worth the time.
Thanks for the recommendation! Speaking of textbooks, I also like Bott and Tu a lot
Thurston's Three dimensional topology and geometry
Yes I've read a part of it, and loved it
"It's almost an axiom that all math textbooks are bad"
Giving you the benefit of the doubt here, I suspect you are intending to say "bad light reading". In which case, that's mostly true.
If you're saying most math textbooks are "bad bad" I strongly disagree and would be happy to recommend several "good good" ones, depending on your interest/mathematical maturity
Um, please don’t wait. I am finally getting back to a place in life where I’m going to be able to take mathematics back up. I want to start back over at Algebra. Working my way towards...whatever it ends up I can do. Lay it on me. All of ‘em.
I really liked “Abstract Algebra, a Comprehensive Treatment.” It’s what I used when brushing up for my GREs
I found https://github.com/GleasSpty/MATH-104-----Introduction-to-Analysis/blob/master/Gleason%2C%20Jonathan%20-%20Introduction%20to%20Analysis.pdf to be a relatively easy read, as far as math textbooks go – at times it made me laugh out loud.
For you I'd recommend you really hack something like Mystik's recommendation of Spivak or maybe Apostol's calculus, if you have the patience. Once you're into the land of rigor and proofs, try Fraleigh or if you loved calc, look under the hood with an analysis book. Here's a repost with my response from a similar question below,
Here's a few I've enjoyed,
Fraleigh - A First Course in Abstract Algebra
[It's on the gentle side, not dry, and provides a good elementary foundation]
Kaplansky - Set theory and metric spaces
[This book... The first thing you'll note is that it's less than 150 pages long. It's essentially a highlight reel of the subject's foundation it also touches on many naively interesting concepts, like cardinals and ordinals. If you've never done anything with ZFC set theory, this is a fun, quick place to start]
Ebbinghaus - Mathematical Logic
[After going through this book, I'm not convinced there isn't a better treatment of the subject (Maybe some logicians can weigh in). That said, it's where I learned what Gödel was actually saying, not the confused pop notions.
If you have a good grounding in linear algebra already, you might enjoy
Lax - Linear Algebra and It's Applications
[Note, it's not "applied" in the common sense of the word. Also, if you don't have prior exp. the book will be... rough sailing. My honors prof used this—it was my first exposure—he advised the entire class to not blink while reading less we miss something vitally important]
I mean I do suppose it depends on our definition of bad. I meant “bad” in the sense of it would be very difficult to teach yourself from them.
I agree that there are some good texts out there, I mentioned it might just be the core texts because Spivak is the only text I’ve had a professor assign that I considered good to learn from. But this was intentional, many of my professors said something along the lines of “it’s good to have a textbook that is not very clear, because then you have to struggle with the material more”
So in some ways I was being a bit hyperbolic I suppose. In other ways this has legitimately been my experience, so I don’t have many experiences with texts I’d actually recommend
What has been your method of learning mathematics? Constructing lots of examples for each concept independent of the text? Reading papers? My general trajectory has been: Read textbook and not understand much -> Read papers and understand even less -> Suddenly understand mathematical concept in a completely different context -> Thinking that I would have understood this concept much faster had I just focused on reading the textbook.
Thankfully I’ve been able to take a lot of mathematics courses at college and post-grad so when I’m enrolled in a class my trajectory is usually: Listen to the lecture -> work through homework -> work through book on stuff I don’t understand for the HW -> Ask the professor for anything I’m stumped on still
When I’m learning something on my own (not often yet as I’m not quite started on my dissertation), I usually pick a topic, come up with questions about it, research and skim papers until I find something with those answers, read through it and physically rewrite the paper so I force myself to confront the material, and then repeat
If you're offering, I'd like to hear your "good good" math textbooks. Upper undergraduate or graduate level. The last math textbooks I loved are Apostol's two number theory / complex function books. They're perfect for me. I've been struggling trying to find something I enjoy as much as those.
I highly recommend Thurston's 3 manifolds (I mentioned it elsewhere in this discussion). Strogatz's Nonlinear Dynamics is beautiful imo -- I wish I could go back in time and tell my younger self to read it. Arnold's ODE book if you've already taken a class/read a book on ODE. Milnor's lecture notes/books. I learned a lot from Lax's lecture notes on functional analysis. I have his book, but haven't read it except for the historical notes. When I was a student, I enjoyed reading Reed & Simon's treatment of unbounded operators, but I am not sure if I am recommending it or not. I also learned a lot from Hirsch & Smale, Introduction to Diff Eq, Dyn Sys and Linear Algebra (1974). The book has been completely rewritten for the 2nd and 3rd editions (with Devaney, and a new title), and I am completely unfamiliar with the newer incarnation.
In a slightly different direction, and perhaps revisiting the elementary topics, a friend whose recommendations I respect highly likes Gelfand's Linear Algebra. He has also recommended Bamberg & Sternberg "A course in mathematics for students of physics" to me multiple times.
Mathematical tastes differ, so take this with a grain of salt.
Thanks very much! I haven't read any of those, although I've heard of some of them. I'll check them all out.
Those sound like great recommendations. Thanks!
Here's a few I've enjoyed,
Fraleigh - A First Course in Abstract Algebra
[It's on the gentle side, not dry, and provides a good elementary foundation]
Kaplansky - Set theory and metric spaces
[This book... The first thing you'll note is that it's less than 150 pages long. It's essentially a highlight reel of the subject's foundation it also touches on many naively interesting concepts, like cardinals and ordinals. If you've never done anything with ZFC set theory, this is a fun, quick place to start]
Ebbinghaus - Mathematical Logic
[After going through this book, I'm not convinced there isn't a better treatment of the subject (Maybe some logicians can weigh in). That said, it's where I learned what Gödel was actually saying, not the confused pop notions.
If you have a good grounding in linear algebra already, you might enjoy
Lax - Linear Algebra and It's Applications
[Note, it's not "applied" in the common sense of the word. Also, if you don't have prior exp. the book will be... rough sailing. My honors prof used this—it was my first exposure—he advised the entire class to not blink while reading less we miss something vitally important]
Enjoy!
I bought Logicomix the day it was released, and loved it all the way through.
Contra you: The math textbooks that I've red are all pretty good and I highly them. Here they are, in ~chronological order of reading:
- Probability and Statistical Inference by Nitis Mukhopadhyay
- A First Course in Linear Model Theory by Nalini Ravishanker and Dipak K. Dey
- Introduction to Time Series and Forecasting (second edition) by Peter J. Rockwell and Richard A. Davis
- Discrete Mathematics by Norman L. Biggs
(My mom's a statistician so that's why there's such an emphasis on statistics, and anyways I'd recommend the middle two much less than the first & last.)
I've been reading Wigderson's Mathematics and Computation recently. It's a bit reference-heavy at times but does a great job giving a general survey of a whole lot of the field of Theoretical Computer Science.
Thanks for the recommendation!
I've been curating a math reading list for a few years, I hope this is the sort of thing you were looking for: https://math.mit.edu/~notzeb/rec.html
I'm glad to see "Visual Complex Analysis" by Needham made your list. I took a graduate class where the professor taught from that book and it was the best math class I ever took, hands down. That book changed my view of what a math textbook could be. Most textbooks refuse to encourage intuition, this book does nothing but.
Thanks for the Needham shout out. I started Penrose's "Road to Reality" last year and got bogged down in chapters 7 (Complex number calculus) and 8 (Rieman surfaces and mappings) Needlam is a referenced by Penrose in chap 7., will it also help me with the Rieman surface? Or something else.
(Physics background. I use complex analysis, but with little deep understanding.)
I really really enjoyed reading Road to Reality a couple of years back. It's a funny book, because you really need to already know most of what is in the book to understand it. What Penrose does is take a lot of complicated mathematical concepts (that the reader is ideally already comfortable with), and give a very hands-on insight into them that textbooks don't necessarily supply. He also of course illuminates connections between them in order to build up the whole edifice of modern Physics.
A good source for Riemannian surfaces is the book by Rick Miranda. Reading perhaps the first chapter of that should suffice for most of what Penrose does
thank you
Thanks for the very comprehensive list!
The Feynman Lectures on Physics
Thanks for the recommendation! Is there value in reading these lectures now, if I have already taken undergrad and some grad courses in Physics. Are there a lot of valuable insights in them that regular textbooks don't necessarily supply?
Oh my. Physics type, Yeah Feynman is great. I'll reread these every ~5-10 years. Most of the physics I use is in there. .. not really so good for a first year student. But it's full of insights as to how Feynman saw the world. They are free online from Cal tech.
Ah I see, thanks!
Strongly recommend The Non-Euclidean Revolution by Richard J. Trudeau. Suitable for math undergraduate who never got properly into hyperbolic geometry
Thanks for the recommendation!
A few of us in my hi-tech company did a book club where we worked through a fairly approachable book on Category Theory titled Seven Sketches in Compositionality. Some bits of it were harder going than others but overall it was distinctly enjoyable. (It helps that I love Category Theory pretty much precisely because it spans a lot of different areas of mathematics and finds unexpected connections tying them all together.)
Thanks for the recommendation! I've read some applied category theory on Joan Baez's blog, and didn't get a clear idea of whether using category theory in things like chemical equations, bio systems, etc gives relevant insights that the pre-existing model does not. Does the book address the usefulness of category theory in these domains? I suppose one use could be that we now have a common language to talk about these disparate domains
Andy Weir’s new book, Hail Mary, came out a little while ago. If you enjoyed The Martian you’ll probably enjoy this one too.
Ooh, I did like the martian. Thanks
I really enjoyed _A Distant Mirror_ by Barbara Tuchman, an excellent history of the 14th century. It's not exactly light reading but it is very well written and entertaining (if you like that kind of thing)
The Spy and the Traitor by Ben Macintyre.
It’s different than your example book because it’s more narrative driven, but otherwise it qualifies as light non-fiction, well-written, and very entertaining. It’s a Cold War spy thriller about a KGB agent who became a double agent for Britain in the 70s and 80s.
sounds interesting, thank you
Semi random author list. (You've probably know some.) Oliver Sachs, E.O. Wilson (I even enjoyed his tome on ants.) James Gleik, R.V. Jones ("Most Secret War", I also liked "instruments and experiences", but it's hard to find.) More physicsy B. Pippard "Physics of Vibration vol I... this is for the experimental type.
Oh and on the same subject, does anyone have a recommendation for a good intro Economics text? Thanks.
I think there's a freely available textbook on the Marginal Revolution website. I think I read the first chapter and it had lots ot helpful illustrations, etc.
If you have any interest in history, then "In The Lion's Court – Power, Ambition and Sudden Death in the Reign of Henry VIII" by Derek Wilson. It dates from 2001 so not the very latest scholarship, but it's not too heavy on the history and scholarship side and takes the approach of picking six men named Thomas who revolved around Henry in his court between 1499 to 1549.
"This book tells for the first time the interlocking stories of six Thomases - Wolsey, More, Cromwell, Cranmer, Howard and Wriothesley - who served as close advisers of Henry VIII - and all suffered for it. Two were beheaded, two were disgraced and narrowly avoided execution, one was burned to death and the other probably took his own life. In the Lion's Court is a revelation of just how perilous it was to be close to England's most tyrannical king."
Wilson definitely has his own opinions as to who the goodies and baddies were, but that makes it a light enough read.
Entangled Life by Merlin Sheldrake. One of the most enjoyable reads of the last year, surprisingly relaxing. It takes multiple interesting approaches to fungi and mycelium. The audiobook is also quite good and read by the author.
https://marginalrevolution.com/marginalrevolution/2020/07/what-are-fungi.html
Also a MR review.
Highly recommend “Into Thin Air” by Jon Krakauer, which is about a disaster with a guided Mt Everest climb in the 90s. He participated in the climb, reached the summit, and was involved in the attempted rescue of his fellow climbers.
What is the fastest you ever learned 80% a topic? What was the topic? How long did it take you?
How would know able to determine that you've learned 80% of a topic, or indeed 100%, or 50% or any other percentage of a topic?
Did you enable skill XP bar in your HUD setting?
I agree with the other comments about the difficulty of determining how much of a topic you know with that kind of granularity, but I assume you’re referencing the “80% of the topic/skill in 20% of the time and vice versa” idea, which I think has some real value.
I’m a business transactions lawyer. I think I’ve learned 80% of the practical domain of solar generation project M&A, and it took me (i) probably three of my law schools classes (Contracts + Business Organizations + a drafting course), which I’ll cumulate as six months, plus (ii) 2.5 years or so of licensed practice with close mentorship and long working hours. I think the rest of the 80/20 proposition is true, i.e., it would take me another 10-15 years to learn the rest.
By contrast, I’ve been studying piano and composition since I was seven, and I still don’t think I’m 20% there on any measure.
Do you think most of M&A is related to memorizing laws and past court cases? Is something else involved?
Piano seems hard because it combines moving your body with memorization, so you can't strictly memorize key sequences to become good at piano. Do you play any other instrument? Piano is particularly difficult
That would be in the wheelhouse of an M&A litigator, someone who sues or defends about contested M&A transactions. I’m a transactional attorney, so I negotiate and draft M&A transactions - structuring the deals and writing contracts, basically. I say all that because, as opposed to litigators, what I do is all about process, writing skills and speaking skills, and very little about memorization. I think that’s what made it easy to learn quickly. But writing that now, I realize that maybe it’s cheating a bit not to count the many years of practice writing and speaking persuasively that set me up to do transactional work. It would be tough to draw that line for any skill, e.g., if you’re talking about learning a third language, do you have to credit any time spent learning your second language, which probably made the third one a lot quicker? Etc.
I do play several other instruments, and I agree piano might be the hardest in a certain way. What I’ll say about piano, though, is it makes your body think that pressing a button will get you the sound you want. That translates okay to fretted string instruments like guitar or banjo, but it sets you up very poorly for violin, trumpet, or any other instrument that requires you to make the sound with your body much more directly and finely. I think the skill curve for instruments of different families look a lot different for that kind of reason.
What are you learning?
Oh thanks for clarifying. Yeah, writing and speaking probably have spillover effects, although I think those effects in traditional education have been studied and found to be smaller than previously thought.
I'm trying to learn natural language processing right now, and there are tons of steps and tons of decisions to make at each step. Luckily, there isn't a time component like a musical instrument where I have to write things at a certain pace, but the learning curve is steep. I'm trying to figure out if there is a better way than what I am doing now, which is looking at a working model, then reading the documentation on each part until I know what is going on. I think I'll have reached my much hated '80%' mark when I have a taste about which decisions to make and reasons to make those decisions, rather than copying other stuff that already works.
Bon voyage man! Sounds like a very worthwhile effort. Without knowing hardly anything about CS - I have to imagine that what you’re doing (working from a working model, reading documentation on each piece) is the best and most efficient way to go about it. Doesn’t sound like the kind of thing that lends itself to watching lectures or some other way of learning.
One way I like to approach a new topic is to find a good undergrad text on the topic, and then read it like a novel.... not worrying if parts of it are confusing.
That's interesting. Do you think you retain very much of the material from reading, or do you take notes? What textbook did you read and why?
Years ago someone in an SSC open thread asked me for the best paper on educational fadeout effects, and in particular the (false imo) claim that they are a measurement artifact. I wasn't really satisfied with the papers that I could point them to at the time, but here's something new and better: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7787577/
Fair warning it's over 40 pages but lots of great stuff; I like the consideration of methods and measurement effects best. (reposted)
My cryptocurrency hedge fund is looking to hire an engineer. PM me if you're interested. Small team + very high potential upside.
Highlights:
* Fully quantitative price prediction and trading (we don't do MM or exchange arbitrage)
* Our returns are uncorrelated to crypto market (pure alpha)
* One of the few funds to offer a BTC share product (we make people more BTC)
* We have been raising additional capital crazy fast this year due to our stellar performance.
Two of the co-founders (me + Satvik Beri) are long term EAs. I've personally donated $$ to MIRI + CFAR way back in 2012-2014. I was also a co-founder of Arbital, in case you heard of it. ;)
We have two fully remote positions: one is for 5+ years of experience, another for 7+ years.
I'm a bit hesitant to add any other requirements, since we mostly just look for smart engineers who get shit done and who would be excited to work with us.
Tech stack: Python + Julia, AWS
I'm friends-of-friends with Alexei and Satvik and can vouch for them being good people.
Thank you! 🙏
Is there a way to PM on Substack? Clicking the email option on your profile doesn't lead to something that looks formatted correctly.
So I sent you a message over LinkedIn, does that work?
Hmm, not sure why the email wouldn’t work. Are there settings for it?
But yeah LinkedIn works.
Hey, I'm a longtime reader (thanks for all the great content Scott!) but just made an account upon seeing this. I'm super interested in quant trading stuff, only have about a year of industry experience (in software development and data analytics), but 2+ years working with a startup and in labs, and I've been playing with algo trading since high school. B.S. in mathematics. I know this is a long shot, but let me know if you'd be willing to consider me!
Quite possible if your algo skills are strong enough to pass our interview. Feel free to email me: my first name at temple dot capital
Crypto, Julia, UIUC grad, and follows rationalist bloggers—what a rare find.
I'm not in a position to change jobs at just this moment; may I contact you when that changes?
Haha, sure thing!
Late to the game, but I'm another UIUC grad, though I work in R. :(
Feel free to send me you resume: my first name @ temple dot capital
I'm already employed as a research economist. I just feel some affinity to UIUC. But, if you ever have an interest in talking about regulation or working with a university based research group, let me know. This is me: https://www.williamrinehart.com/
hi there alexei, do you have any other way of getting in touch? i might be interested, but do not see any email on your profile (maybe it's a subscriber-only feature?) and do not use linkedin.
my first name at temple dot capital
Does anyone have a theory as to why thinking about philosophy or other high level concepts can actually feel physically heady? Why does galaxy braining feel this way?
I think this might be what's more commonly known as "insight porn" (the name itself basically explains why it feels heady), unless what you mean by "philosophy or other high level concepts" is much more general.
Does "insight porn" as a phrase imply that the topic has little or no content? I looked it up in urban dictionary and it's defined there as sortof cheap tricks of the "theory of everything" variety, reading with no actual learning going on.
> I looked it up in urban dictionary
As of when I last checked it (at https://www.urbandictionary.com/define.php?term=Insight%20Porn), here's only one entry for "insight porn", written by "metamusoul". And anyways I wouldn't trust urban dictionary that much.
Also, related: https://mindlevelup.wordpress.com/2016/10/28/insight-porn/
Thank you for the definitions. That "insight porn" exists makes sense. By definition there are other experiences of thought that are not insight porn. I expect that there may be experiences or sensations associated with multiple types of thought.
Ha maybe but idk if it's really an exciting feeling. Just like, spacey and removed from my body when thinking about how wild it is that there's anything at all. Maybe it's just an emotion I haven't named.
ego death, mindfulness, nirvana
The human mind can't help responding to ~imagery
Talking in "far mode" is associated with high status, talking in "near mode" is associated with low status. When you think about philosophy, you feel high-status. Talking about philosophy would possibly feel even better, assuming that other people listen to you respectfully.
I don't have an exact model why this is so (maybe Robin Hanson could explain better), but if you ask who is unable to think and talk in far mode, two obvious answers are: stupid people and people who are too busy working. Therefore, philosophising signals that you are neither a moron nor a slave. There is probably a quote from Plato that debating useless and unempirical shit is the most noble action a human soul is capable of.
Once philosophising becomes a known signal of high status, it is further reinforced by social reactions. When a low-status person starts philosophising, they are told to shut up. When a high-status person starts philosophising, people listen respectfully and pretend to be impressed. Knowing the respected philosophical ideas also signals that you are educated.
Philosophising is very popular even among people whose ideas are complete bullshit, because it signals "hey, I consider myself high-status". And if not challenged, then indeed it is a signal of high status; either because the listeners are so stupid that they are impressed, or because they want to avoid conflict with the person who asserts high status.
I think it's not just high and low status, but just signaling in general. People who feel superior for doing philosophy, and people who feel superior who willingly ignore/hate philosophy. But I don't think it's all signaling anyway: you say that "pretend to be impressed", but I feel like they probably *are* impressed (even if on closer inspection they would've realized that the idea wasn't that interesting). And there are definitely a lot of people who think about far mode ideas without socialization: for example, taking hallucinogenics by yourself would probably make you feel spiritual etc.; ie it's not just that it feels good. I think if it can be experienced without other people, then signaling probably isn't going to explain it, at least not entirely. (Although, perhaps drugs replicate the emotions from being in a group; although although, this makes me think otherwise: https://www.youtube.com/watch?v=loCBvaj4eSg)
I do not think any of this actually contradicts the signaling hypothesis.
Things become signals of something, because there is a good reason why they are naturally associated. (Debating abstract thoughts requires intelligence, which is why it is signal of intelligence.) When the signal becomes known as a signal, people will try to countersignal. ("I am so smart that I don't even need to discuss philosophy to make people see how smart I am.") And people are also signaling to themselves.
Max Stirner (The Ego and Its Own) said that having a theory about something can make you feel like you've dominated it. I think he was on to something.
This is from memory of Stirner.
China has been quietly building roads, military bases and villages in Bhutan's territory. The area itself is of no real value (which is probably why it wasn't noticed). But is thought to be motivated by trading for other territory that's more valuable near disputed Indian border,where there have been shows of force from both sides recently.
This is partly interesting for geopolitical implications, but also because the whole idea of stealthily taking over parts of another country by building there is bizarre.
Tweet thread outlining main points: https://twitter.com/BeijingPalmer/status/1390759269696421897?s=19
I link that as main article https://foreignpolicy.com/2021/05/07/china-bhutan-border-villages-security-forces/ is semi paywalled.
“the whole idea of stealthily taking over parts of another country by building there is bizarre” - Pretty metal. Sounds like a state-level application of adverse possession.
In September 2018, the Finnish police supported by military raided a series of holiday villages / vacation homes in Finnish archipelago that were owned by a single company and had been in construction since 00s, but the business was never operational. The twist is: the company that owned the sites were owned by a Russian and the sites were located near strategic sea routes. There were media claims that that in addition to the luxurious villas, the vacation homes in question had some facilities (helipads, oversized piers, etc) which (according to some) made them look less like a vacation business and more like secretive, plausibly deniable bases for naval operations.
Government never commented on the military aspects, but the police confiscated some millions of euros. According to latest news items, the business is still being investigated for money laundering, with no charges yet. The Russian owner Pavel Melnikov later gave an interview to newspapers, where he claimed that he has a hobby of collecting islands not unlike some other people collect art, and he bought the islands because he liked swimming around them. The sites in question have slowly been sold away.
Links in English and Finnish:
https://corporalfrisk.com/2018/09/23/a-dawn-raid-in-the-archipelago/
https://yle.fi/uutiset/3-10976716
Forgot to add the reason I shared the story: Maybe a secret military base building is a trend?
Why don't Western governments simply make paying ransomware illegal? This seems like a classic collective action problem- it may certainly be rational for an individual actor to pay a ransom, but it feeds more & more cybercrime by creating a huge market. While ransomware will probably never go away entirely, taking revenue away from large criminal organizations is generally considered a good thing. Realistically the government probably can't stop individuals from paying a ransom, but large corporations tend to be fairly conservative and 'hey Mr. Executive, federal prison is not a nice place' is a fairly convincing argument. This law doesn't have to deter 100% of ransom payments to just make the practice overall much less profitable.
Not even sure you'd need to pass a new law- for the US, anyways, I am very confident that somewhere in our massive list of federal laws 'making cryptocurrency payments to organized crime' could be covered by something already in there. The entire finance/payments industry is already very highly regulated.
So- why aren't ransom payments illegal already?
Agreed, but it seems like enforcement has not been a priority for the US so far. There has to be a really good reason that my dumb (non-attorney) brain hasn't thought of.
I would imagine that deterring bad guys now before they encrypt, like, JP Morgan's servers seems like a good idea. Or Raytheon or Microsoft or Mastercard or someone
"This means the government would have to prove the person knew paying the ransom was illegal"
Really? What about "ignorance of the law is no excuse"?
Ah interesting. But would paying a ransom be considered a financial crime or something more in the area of theft, robbery etc? Also, after the first couple of high-profile cases, wouldn't it become unrealistic for an accountant working at a large company to claim ignorance of the illegal-to-pay-ransom law?
Acting willfully does not require an intent to break the law. "Someone commits an act willfully when he or she does it willingly or on purpose. It is not required that he or she intend to break the law, hurt someone else, or gain any advantage" https://www.justia.com/criminal/docs/calcrim/800/960/
Ditto re acting intentionally: "A person acts with wrongful intent when he or she intentionally does a prohibited act[or fails to do a required act]; however, it is not required that he or she intend to break the law." https://www.justia.com/criminal/docs/calcrim/200/252/
Wouldn't it be rather hard to prove a company paid a ransom? Or would you need whistle-blowers?
I really have no idea. But suppose a large bank was paralysed and hundreds of thousands of accounts inaccessible. A ransom of say $1m would solve the problem. What should the law be in such a case?
I have no solid answer, but maybe the blast radius is one of the reasons? Ransomware can bankrupt a company if they can't get their data back, which would put its employees out on the street. Though I suppose there's no reason it couldn't be made illegal to pay ransomware and you'd also be guaranteed financial support; but *then* people might infect their computer systems with ransomware to get the financial aid without having to work (would really up insider risk - as someone's IT admin, would you rather work 9-to-5, or get the same compensation for a comfortable amount of time after your company was infected with ransomware)?
To stress: I am just thinking out loud. This topic is way more complex than I could wrap my head around. Grain of salt.
Couldn't such a law be made pretty much arbitrarily difficult to enforce by opening multiple Bitcoin wallets and breaking the payment into multiple transactions? Assuming the criminals are well-hidden and reporting rather than paying means your data are gone forever, reporting means you're out of business and paying secretly still has a pretty low chance of serious repercussions.
> large corporations tend to be fairly conservative and 'hey Mr. Executive, federal prison is not a nice place' is a fairly convincing argument
There is a solution for that: "data recovery firms" that pretend to recover the data, but actually just pay the ransom, giving the customer plausible deniability.
https://slatestarcodex.com/2019/06/02/ot129-opaque-thread/#comment-758072
https://features.propublica.org/ransomware/ransomware-attack-data-recovery-firms-paying-hackers
Of course David Friedman has an analogy from 18th century England: "Jonathan Wild, self-appointed Thieftaker General," who got "rewards for the recovery of stolen property, and income from the large-scale employment of thieves".
It doesn't matter whether they are illegal or not; the American people are not so coldheartedly rational as to be willing to put crime victims in jail just because the consequentialist math says it generates more net utils, so for something like this they simply won't. Prosecutors will find that, gosh, they're very busy and there's ten more important things on their agenda this week. Juries will decide that "reasonable doubt" extends to the slightest fig leaf of bare possibility. And if you somehow pass a no-exceptions mandatory-prosecution law, then someday the law will fall on a telegenic minority woman running a small business who's just trying to make payroll for her equally telegenic working-class staff, and This Is An Outrage and Something Must Be Done.
It seems like a minimum first step would be to make the ransom itself unspendable. With Bitcoin that’s easy, just make the coins themselves the equivalent of registered stolen goods—subject to seizure even from a bona fide purchaser for value.
This would be interpreted as an attack on Bitcoin itself (which may be what you wanted).
I wouldn’t say I want to attack Bitcoin as an ultimate goal, but if forced to choose between attacking Bitcoin and not taking steps to deter ransomware attacks on critical infrastructure I’d definitely pick attacking Bitcoin.
How does the government find out which bitcoins are dirty?
The victims would report them after getting control back.
I’ve been feeling stuck in a solipsistic frame of mind for a year or two - constantly wondering how I can have any belief that anything is real outside of my own mind. Has anyone here found a satisfying/permanent way out of that? I used to rely on a kind of soft religion, but it wore off. Would appreciate any pointers.
Being surprised. I played with solipsism briefly, but I was adequately sure I didn't write Goedel, Escher, Bach.
Haha thanks. Well put.
By realizing that there is nothing you can really do about it. If your senses are feeding complete garbage about the world and you cannot stop them from feeding you nonsense, then there is nothing else you can do. I do think that there is a real, external world and my senses can provide, at the very least, a few clues on how it actually works, but I doubt you can 100% prove it is so, it is best stop thinking about it too much and just roll with it.
Your mind is interpreting the world as you need it. At that level what you are seeing is in fact a true representation.
What do you mean by “as you need it”? In an evolutionary/survival sense?
Yes.
I appreciate that. I think you’re getting at what I’ve started to find uncomfortable about it - it’s like a tick or an intrusive thought, i.e. something that I know is doing me harm (in terms of distraction, frustration, brief fits of fatalism) and that I know I can’t think my way out of but that I can’t stop thinking about. I could describe some of the ruminating habits I used to have when I was depressed the same way, I wonder if maybe employing some of the same mental tricks that helped with depression would help with this. Again, thanks for your thoughts.
It's not the kind of thing one can think one's way out of, and it's so compellingly grim that it might not work to tiptoe away from it via attending to gentler alternatives (eg, "soft religion"). What about plunging into some things that provide strong sensations and emotions -- rock climbing, other thrill sports, aggressively fighting injustice, helping the desperate in a direct and physical way?
So I think you hit on something here! A lot of the onset of the solipsist thinking coincided with finally finishing school, getting my “real” job, getting married, getting a nice place and a dog, etc. - settling down. And I think that a lot of the wild emotional ups and downs and the variance of experience inherent in daily life as a teenager and an early-twenty-something helped to keep me grounded in reality. Now that I spend every day doing almost the same thing, I’m just spending a lot more time in my head (in a bad way). It might be that I’m just not mentally cut out for the daily grind and need a lifestyle with more excitement and variance to keep myself level. Thanks for your thoughts!
Try having kids or murdering a hobo.
I’ll flip a coin.
Sam, it's the billable hours.
The world is far too complex to have been made for just you. Who do you think did that. God. Or your own mind?
I mean, I really hope it’s God, but he’s a terrible communicator.
Bit odd that he creates the world just for you and doesn’t communicate, for sure. Another argument against solipsism is to compare your life with an actual simulation. Ever play a video game where you were living an ordinary ho hum life? Sims excepted I suppose. Imagine buying a video game where you were nothing special but the world was vast and plenty of the NPCs were doing great. You could watch this in your low rent apartment on TV. You need to work to own this tv or you get kicked out if the house and starve. Eventually you die doing not much. That’s it. That’s the game.
Lol that’s so bleak but you’re right. I don’t know if you like $5 PC games but Dave Man might give you a laugh, it’s just like what you described. I’m hung up on what you pointed out, that everything your senses give you is an adapted representation anyway, at best. I’m sure that plays out on a lot of different levels (e.g., I interpret others’ behaviors in a way that suits my survival; also, I perceive colors, objects, and the rest on those emergent levels without knowing for sure whether what they seem to represent really exists to another observer).
For what it’s worth, the idea of God does consistently present itself as the “out,” I just wish I could flesh out an idea of him/her/them beyond “another observer.”
It's funny, because I'm listening to "The Master and His Emissary" right now, which goes into a lot of detail to discredit Solipsism. Also, what Nancy Lebowitz said
I’m going to check that out. Thanks!
That would be a good book to see a review of. I found it very interesting content-wise but not well written - tedious and repetitive.
Total non-professional opinion - I think this may be a form of OCD. It might be counterproductive to try thinking your way out and revisiting the possibility. Resisting can make intrusive thoughts even more disturbing and frequent. The way out might be to learn to live with the possibility. Again , I have no real expertise on the matter, just throwing out a possibility. Best of luck!!
It depends on your underlying epistomology. I think most commonly accepted theories of justified belief will allow you to accept that the external world exists (even classical foundationalists, who I think also include the most serious idealists). For a simplified example, suppose you accept a coherence theory of justification. Then it's far more coherent to add your set of beliefs that "the external world is real" than to add "the external world is fake." Obviously it's more complicated (and fun to think about) than that and would recommend any introductory epistemology book. More personally I find that it's better to give up on gaining ultimate knowledge and focus more on what you are justified in believing. Also, consider seriously what your standard for knowledge is. Generally to have knowledge of something we don't need to be certain it is true.
Have you sat down and tried to define, really really thoroughly, what precisely you mean by "my own mind" and "outside" and "belief" and "thing"? My experience of contemplating solipsism is that, once I've defined (or failed to define) the terms it uses to express itself, it stops seeming like a coherent problem. Put another way: when I express, as truly and completely and precisely as I can, what solipsism is trying to refer to when it says "my own mind", I find that it's not something that could conceptually do the thing solipsism is suggesting it might do.
A third way of trying to say it is that solipsism exists because we're married to terms that exist in a very different web of relationships than the entities the terms are supposed to correspond to. The terms are just wrong. My actual mind doesn't behave the way the English-language "mind" does. Real things don't behave the way the English-language "thing" does. So solipsism is a logical statement when used of those terms, but it doesn't actually make sense for the entities themselves.
This. The problem with solipsism is that it seems like a meaningful position, but it’s actually incoherent since it denies the basis on which words and thoughts mean things.
Moore's common sense proof of the external world. This seems like a decent overview: http://micahcobb.com/blog/g-e-moores-proof-of-an-external-world/
Nah, that’s a terrible argument, just like his naturalistic argument is. Both covertly smuggle in the conclusion, distract you with an irrelevant thought experiment, and then declare victory.
Honestly I think Moore was kind of a hack.
You're just asserting that knowledge requires proof (like knowledge of hands), but this clearly isn't true. This is touched on in the article I linked.
That’s not the problem with Moore’s argument. The problem is that it assumes that a hand is an external object, which is the very thing in question, then presents the fact that hands exist — which was never in question — as an argument for external objects.
It’s the exact same problem with kicking a stone to refute Berkeley’s idealism, or, from another angle, trying to refute a trivialist by making an argument that trivialism is false: the existence of hands, the ability to kick stones, and the falsehood of trivialism are all entirely compatible with the view being refuted.
Reality *is* just the patterns in my experience
I've been amused by solipsism and subjective idealism framework in general in my salad days. I still think it's a valid default position for a young philosopher. For instance, I used to claim that materialism imply a huge extra entity (Objective Reality) so it must be more complex in the Occamian sense and therefore less probable.
However, you learn more things, it turns out to be the opposite. While materialism does imply the existence of objective reality it delivers on all its predictions. Unlike solipsism which doesn't have any predictive power. Moreover, if your consciousness is fundamental, why don't you have psychic powers? Why does your experience seems stable? Conservation of energy and matter seems much less likely in idealistic universe than in materialistic one. While this by no means disprove solipsism and idealistic framework in a strict sense, it makes it more complex in Occamian sense and therefore less probable explanation.
Why is wildlife in Australia unusually venomous and dangerous? Everyplace is subject to the evolution.
Veritasium has a good video about this: https://www.youtube.com/watch?v=myh94hpFmJY
It's complicated, and some of his sources claim that Australia is not, in fact, unusually high in venomous species. But the shortest answer is: Cold-blooded species can only thrive in warm climates, and are unsuited to hunting strategies with high energy expenditure like chasing or overpowering. As a result, they are more likely to evolve venom to hunt with.
I wouldn't call that video "good". The guy gets uncomfortably in-your-face, and the signal-to-noise ratio is low. The same information could be packaged in half the time.
Also, is number of venomous species per country a good way to look at the situation? Country borders are arbitrary, and countries are of widely varying sizes.
However, the number of venomous species in Australia might be a matter of chance. I'm not sure how much they're in the same environments so it might not be an arms race.
Hmm, venomous is associated with hot climates. (not sure why) I wonder if it's also associated with dry? I'll also observe that Australia spans a temperate climate in the south, (lots of rain and huge gum trees.) to desert, and then tropical climates in the north. So lots of different habitats. There is ring of life around the edges, and desert in the middle, maybe that makes it harder for species to find new habitat. (Sorry just speculating wildly.)
I recently bought a new house. The disclosure mentioned that subterranean termites had been found in some wood in the basement when they looked for them, but that the seller hadn't done anything about it and I needed to decide how to react. My options seem to be to do nothing, or to pay a very large amount of money to an exterminator and wait for them to do stuff before I move in.
When I Googled termites, I found many exterminator company websites saying you needed to exterminate them right away, but also some articles saying that about 50%-60% of houses in California have the kind of termites my house has. If my house has a completely average number of termites, this sounds like less of an emergency.
Does anyone know more about how I should think about this?
What does the inspector that you hired say?
+1 is it not normal to get someone to inspect the house *before* you buy it, so you can renegotiate the price if a lot of expensive work needs to be done?
Termites need wood and moisture. You can reduce moisture with a de-humidifier. Also, cover up any spots where outside humid air can enter the basement. Obviously, check the wood in the basement for any damp, rotting wood. If you see that you probably need extermination. Assuming you don't have that, keep the moisture low and get an annual inspection by a reliable professional.
>cover up any spots where outside humid air can enter the basement
Maybe it depends on construction techniques and the local climate and water table, but I thought that builders deliberately put in air bricks so that inside humid air can leave the basement? and that house owners covering these air bricks over or letting them clog up was a common cause of damp problems.
Indeed. I've renovated houses in England and France and without fail houses in both those countries are more humid inside than outside. A central feature of modern house building (at least in Europe..) is to design the structure such that moisture can leave the house. And also to do so without it condensing on a cold surface on the way out.
If NorCal is anything like SoCal, there's no basement.
Um if the sale hasn't closed knock whatever you can off the price. If it has closed get a specialist to work at once.
You should pay to have them exterminated. Ultimately, you will need to sell the house. If termites have been eating your foundation, many buyers will not be willing to buy. If they destroy your foundation completely or damage it, you won't know until your floor starts sagging and you have to get that replaced.
Source: had to remove a bug infestation recently from my foundation before selling my house
If this wasn't an insane seller's market, an active termite infestation would be a hard pass for me. There is no easy way to determine the severity, it could be anything from "nothing" to "the structural components of your house are now compromised and it will take massive amounts of time and money to make it safe." How long do you plan to live there? Do you plan to do major renovations that involve opening up walls? If you plan to flip it in a few years you might get away with doing nothing, especially if the market stays hot and buyers don't have much leverage. If you're there for the long haul, or you are going to expose the wood and risk seeing thousands of dollars of needed repairs, you should contact a trusted pest control professional and follow their recommendations.
It's a case of "ignore small problem now, develop big problem later". The seller ignored it and left it on the buyer (in this case, you) to deal with it. Deal with it now, because if you intend to live in that house for any reasonable period of time, you will eventually have to deal with it. Even if you only intend to do up the house and hope to sell it on in a couple of years' time, a prospective buyer is going to have the same problem as you do now, only worse by a couple of years' neglect, and they may well decide either not to buy or not to buy at the price you are asking.
Honestly, *anything* to do with a house? Fix it NOW before you are *forced* to fix it because by then it will be bigger problem, cost more, and be an even huger pain to deal with.
Looking 'em up on this University of California website, you should deal with them now while you have the chance and the least inconvenience. http://ipm.ucanr.edu/PMG/PESTNOTES/pn7415.html
These ones like damp, so that may be an indication of a further problem (leaky pipes?) with the basement, and since they prefer rotting wood to sound as it's easier to chew, yeah you should have this looked at and don't ignore it. And it seems like fumigation doesn't work so pesticides have to be sprayed, another reason to get this done *before* you move in:
"Moisture Requirements
Subterranean termites require moist environments. To satisfy this need, they usually nest in or near the soil and maintain some connection with the soil through tunnels in wood or through shelter tubes. Furthermore, because of the moisture requirements of subterranean termites, they are often found in wood that has been slightly decayed. Soil serves as a source of moisture that protects termites from desiccation, shields them from predators, and can be used as a building material for shelter tubes and carton nests (Formosan subterranean termites) above ground. Termites can also excavate passageways through the soil to reach additional food sources.
Feeding
Termites do not like all wood species, but the condition of the wood is more important in determining the probability of infestation. Decayed wood is eaten faster and preferred over sound wood. Digestion of wood, in this case, really begins before the termites take their first bite, since decay fungi in the wood break down cellulose into smaller units. Termites can digest sound wood, but decay fungi make their work much easier.
MANAGEMENT OF SUBTERRANEAN TERMITES
It is unlikely that homeowners will be able to execute subterranean termite control on their own. However, it is important for homeowners to have some familiarity with inspection procedures, reduction of conducive conditions, and treatment strategies. Successful termite management requires special skills and knowledge, including a working knowledge of building construction. An understanding of termite biology and identification can help a homeowner understand and select a suitable method of control. Of course, homeowners can replace termite damaged wood and correct conditions conducive to subterranean termite infestation on their own; however, applications of registered pesticides are highly regulated and require a licensed pest control professional to carry out the inspection and control program.
Multiple colonies of the same termite species or several different species can infest a building. A professional inspection and an integrated approach to control are required. A combination of methods, such as habitat modification, elimination of excess moisture, removal of infested wood from the structure, exclusion of termites from the building by physical and/or chemical means, and the use of chemical methods to destroy existing colonies will probably be necessary.
Most subterranean termite species consume wood at about the same rate, but three factors can make some species potentially more voracious and damaging than others. These factors include the environment in which they live (termites eat more wood when conditions are optimal over a longer period of time), the size of the insects (larger insects eat more wood), and the number of insects (larger colonies eat more wood).
Controlling Subterranean Termites
Subterranean termites in structures cannot be controlled using techniques that are appropriate for drywood termites, such as fumigation, heat treatment, freezing, and termite electrocution devices, because the reproductives and a large majority of the termites are concentrated in nests near or below ground level out of reach of these control methods. The primary methods of controlling these termites are insecticides, either applied to the soil adjacent to the structure, directly to nests via shelter tubes, or through bait stations. To facilitate control of subterranean termites, destroy their shelter tubes whenever possible to interrupt access to wooden substructures.
Insecticides
Liquid applications of pesticides are most often used for subterranean termite control and applied to the soil either in drenches or by injection. There are no reliable over-the-counter termite control products available for the public in California; all effective products are for professional use only.
Pest management professionals are provided special training because of the hazards involved in applying insecticides to the soil around and under buildings. Applications in the wrong place can cause insecticide contamination of heating ducts and/or damage to radiant heat pipes or plumbing used for water or sewage under the treated building. Soil type, weather, and application techniques influence the mobility of insecticides in the soil; soil-applied insecticides must not leach through the soil profile to contaminate groundwater or run off to contaminate surface water."
Scott,
Here are my words of wisdom and experience, first in bullet points:
-- Termites are fucking Satan's evil incarnate and they never stop.
-- Fumigation is a Band-aid that is done for insurance reasons and because a crew can just show up and do it, just costs $$$ and helps you not see termites for maybe a year. All these termite companies do the same thing and the market is huge because of CA house prices. Think of them as the Earl Scheib option, an auto paint company that just sprays a new coat on but does not do any prep or repair and misses a lot of stuff, but it looks ok from 30' away.
-- The only way to really solve the problem is the Nuclear Option, and it is a huge amount of work and disruption requiring you or a trusted contractor to touch and treat every single cubic inch of your house, drill holes all over the yard and even the hardscape. It is the only way you can be sure that your house will not collapse two weeks from now.
Here is a link to a Tech Bulletin from the only Company I trust.
https://drive.google.com/file/d/1tdyrqy0xq0cUQrpL4cj0HTUPAP504ySY/view?usp=sharing
________________________________
I'm sorry you let this slip by you and now have a termite problem. Of course, all wooden structures in California have a termite problem, so fumigation status would seem to me to be a very important part of due diligence.
Over the last ten years in LA, I have developed a standard visual spotter's guide to gentrification:
1) Fumigation tent/enclosure goes over the house
2) Redwood fence installed. (You can tell everything about the new owners by how the fence is executed: Did they do it as cheap as possible? (house flipper) Did they ask for something they saw on Pinterest? (Clueless first-time-owner.) Or did they do it like it's going to actually last and function for 25 years? (These people plan to LIVE here.))
3) Prius or similar EV appears in driveway/curb.
I suppose upon reflection that the reason fumigation happens post-sale is that it is impossible with people and their stuff actually living there.
OK, I will wrap this up and give you the news: Everything must be treated with Disodium Octaborate Tetrahydrate. The stuff is not expensive in bulk if you want to DIY, and Thank God it is still sold in California, you must act fast before it too is banned.
Act swiftly and immediately.
Best,
BR
This is probably a dumb question but I'm on a roll so I'll ask it anyway.
Is there any hope of replacing (some of?) the wood with something else that isn't wood? Metal, fiberglass, plastic?
What fairly well-known pc/console game would be the toughest to train an AI to master? My guess would be EU4.
Among Us is probably quite tricky.
Good choice. I hadn’t even thought of that one.
If you dropped one AI into a game with nine other humans it would of course do quite badly, but if you dropped one human a game with nine AIs, I bet the AIs outdo the human. I have a hard time defining what a "fair" structure would be for testing humans vs AIs here: the asymmetric nature of the game means you can't even do like DOTA 2 and make it five humans vs five AIs, and putting humans on the same teams as AIs will just drag the team down with communication problems.
I guess I would have the criteria be to have the ai master any scenario: 1 human vs all ai or all ai vs 1 human and every other combo. The interaction with other players required in this game would be similar to a Turing test. So, yeah having the ai master this seems to me like it might have to be a general ai - which much more advanced than I was thinking of. I had games like Civ or the paradox games in mind, which I would suspect making a dominant ai would be easier than Among Us.
I've encountered groups of humans who would all cluster in exactly one place (pixel perfect) waiting out other people. An AI that figures out that cooperation mode with other AIs always wins.
I think whatever it is needs to be realtime, the possibility space is just so much bigger and therefore more resistant to brute force (or in modern ML, slightly smart force) searches. Starcraft 2 shows this in action: the early version of AlphaStar had obviously superhuman micromanagement but failed to grasp basic tactics like "don't pick fights by attacking through a chokepoint and up a ramp". Even the final version seemed to run on really good economy management shoring up strategy ranging from simple to stupid (it lost one notable game by building a bunch of can't-attack-air units, running into an enemy air army, and building more can't-attack-air units until it died).
Right, brute forcing isn’t much of a solution for the more complicated games. Age of Empires 2 definitive edition has a very challenging AI (as long as you don’t rush the ai) But that ai’s effectiveness is due to the fact there is an optimal path of play for the game especially in the opening minutes of the game. I don’t *think* Civ, EU4, CK3, etc have such a single optimal gameplay that an AI could be programmed to to be able to almost always win.
Beating arbitrary levels in Mario Maker 2 would be tough. It's a time-limited, constantly changing environment, and the AI would have to master a wide variety of situations. It would be like putting a self driving car on a track built by a hostile designer that wants to cause crashes.
If you really push the limits of the game engine, this becomes a reading comprehension test. Mario Maker doesn't have much in the way of explicit scripting language, but you can still kludge together an astounding number of "if player does arbitrary X, do arbitrary Y to the game world" mechanisms through clever use game systems. Imagine a level where sixty seconds in, the floor will disappear unless at the very start you walk three steps to the right and jump at empty air.
You can make a level that does basically arbitrary unfair things that have no relation to traditional platformer mechanics, then explains those things via English text (so that human players can easily beat the level) and now you have a level the AI can't beat until it learns to read. Repeat for whatever other tests of human information processing you build in.
The original X-COM: UFO Defense from back in the 90's. Still actively modded and updated; check out the Open X-Com project. Very difficult tactical combat system with a ton of OOC prep, research, and equipment decisions that make or break you.
Also, sometimes the aliens are just overpowering when you show up, and you have to flee or get TPK'd.
In Open XCOM there's an ironman difficulty setting, with no save scumming. If you can make an AI that does a credible playthrough of XCOM UFO Defense on ironman, hardest difficulty, then I want to see the playthrough video. It's fricken tough.
UFO: Enemy Unknown is not actually a very hard game *except* when played blind and/or without deep knowledge of the mechanics.
Most missions can be beaten with a pretty rote plan:
1) on turn 1, use two Smoke Grenades, one dropped inside the Skyranger and one thrown to the top part of the ramp, then end turn. (This stops the aliens seeing into the Skyranger on their first turn, or getting reaction shots as you disembark.)
2) Disembark, shooting at any aliens that you spot from soldiers *further back in the smoke cloud* who can't see the aliens (and which, therefore, the aliens can't see for reaction fire).
3) Set about 1/3 of your soldiers as snipers (kneel them down in a firing line, maybe two spaces apart, and don't move them except to turn) and the rest as scouts. The highest-Firing-Accuracy soldiers should be snipers.
4) Sweep the map with the scouts, shooting at any aliens with the snipers (who are beyond visual range and therefore can't be reacted to). Stop sweeping for a turn when either your scouts are all down to about half TU, or you're running out of TU to take shots with your snipers.
4a) If there are buildings in the way of your snipers' sightlines, shoot at them with Laser Rifles until they're mostly destroyed. This also goes for jungle terrain and orchards. Sand dunes and hills are trickier; you may have to relocate your snipers to the top of them. Forests are the worst because you can't really just level them, but you'll usually have a couple of snipers with clear sightlines so it's not too bad.
5) Once you find the UFO (if there is a UFO; terror sites won't have one), camp a bunch of soldiers off ~15 squares side-on to the entrance (that way, an alien who comes out will not be able to see them, and your soldiers will be able to take reaction shots. If it's a Medium Scout, just camp there (with the majority of your troops) until all the aliens come out and get side-swiped. If it's bigger, camp with ~5-6 soldiers while assembling the rest (including snipers) to go in the entrance.
6) Sweep the UFO, room by room, with 3-man squads along each passageway (if there are too many passageways, just hold a chokepoint while clearing out one section and then bring over the freed-up soldiers from the cleared section to assist). Never move a soldier while he's in an alien's sightline unless you're moving him out of the sightline or you're sacrificing him. Prefer points for troops at end-turn that are to the side of a door, so as to get reaction shots.
The only really tricky missions are psi missions, alien bases, and Cydonia, and the latter two can be put off painlessly until you get psi tech for yourself (which makes the game boringly easy). With psi, the moment a soldier is targetted for psi-attack, have him drop everything and just passively scout for the rest of the mission; he has low psi-strength and will probably be targetted again. The trickiest bit is that you're looking to capture the psionic alien alive, which is risky and may not be achievable.
Now, the plan I outline above is not *trivial* to convert into code, but I don't see any outright judgement calls there except in the psi missions. I'm pretty sure you could code an outright dumb program to do it (and certainly you could for research order and most of the Geoscape). If a computer game's rote for a human, it's almost certainly programmable.
Now, an AI that had to compete against a human on a modded two-player X-Com would have to be significantly better. But the actual vanilla game is pretty easy (though TFTD's quite a bit harder).
Thanks for the writeup. Have you played Xpiratez mod? This strat set won't work; there's a ot of alternate vision modes, more psi enemies and varied enemy types, damage resistances, and a ridiculous tech tree.
No, I haven't engaged with OXC at all and Piratez' fluff never seemed very relatable to me.
One thing that a computer would be really good at in the original XCOM is farming stat improvements. Once you realize how it works, you can create super-soldiers if you are willing to do a bunch of boring things. https://www.ufopaedia.org/index.php/Experience_Training
An AI is also patient enough to wait out the computer AI.
There's a resource in xpiratez called "Glamour" that you only get for finishing missions quickly, and which you need for some improvements/tech. Also, having fully maxxed out soldiers in xpiratez is nowhere near as invincible as it is in vanilla. 1 shot kills happen all the time.
Seriously, there's a ton of depth and tactical requirements for xpiratez that would make it very interesting to see an AI tackle it.
There are also several mission types with built-in ticking clocks.
1-shot kills of maxed-out soldiers happen quite a lot in vanilla UFO. The Heavy Plasma Gun does 0-230 damage per shot and can fire up to 6 times per turn; maximum armour is the front plate of a Flying Suit at 110 and maximum health is 61.
TFTD is a mixed bag; your armour will stop ordinary alien weapons, but the "armour doesn't help" stuff gets more common as well.
Apocalypse definitely has near-invincible soldiers, though, both at the start and the end (though not the middle). I usually don't lose a single soldier until Devastator Cannons show up, and a shielded/disruptor-armoured/cloaked soldier takes a minimum of three shots to bring down, which have to be from different weapons in a specific order.
EU4 and a few of Paradox's other grand strategy games get weird when examining the effects of AI micromanagement. For the uninitiated: the state of the board is recalculated once per day of the in-game calendar, which will generally be every second or three depending on game speed. Orders given by the players are input between updates, and many are immediately visible on the board to all players. This can occasionally lead to a situation where on March 23rd 1446 I want my armies marching toward Oxford, but then my opponent will see that and flank me by ordering a march towards Sussex, so I see that and change directions to head them off at Sussex, and then they choose to go back to Oxford, so *I* switch back to Oxford, and then they go back to Sussex, and....
In human multiplayer games, the typical way of solving the issue is just to have the game run at a defined speed and let the chips fall where they may, with the last set of orders at the update going through. Against the computer, it doesn't come up excessively often since the AI tends to issue fairly broad directives without too much micromanagement. But when the AI does choose to micro, it *will* beat the human through reaction time or just sheer persistence.
Right now the AI is mediocre at operational-level thinking, but already has an insurmountable edge in the finest-grain tactics. Paradox have taken some steps to reduce the effect of the latter (moves that take multiple days 'lock-in' after 50% of the time elapses), but I'm not sure it's fully soluble without a significant change to the game dynamics.
(Stellaris largely resolves that issue by making combat and a few of the most micro-critical actions all real-time, but has other AI compute time problems so bad they keep breaking the game trying to avoid them. No, I did not care for the recent pop growth changes, how could you tell.)
Diplomacy? https://arxiv.org/pdf/2006.04635.pdf
There's a lot we don't know about how to train an AI to do this sort of multiplayer, game theoretic, theory of mind stuff.
But I'm glad that researchers are out there teaching the machines how to betray.
> showing that a common pattern is good growth until ~1980, poor growth until ~2002, then good growth until ~2012, then poor growth again.
I'm not convinced it's that common, from the graphs. It fits Sub-Saharan Africa. In LATAM and ME/NA there's a blip in ~1980, but it's not obvious that growth after is slower than growth before. And in South Asia, neither ~1980 nor ~2012 seem like anything happens.
Does anybody knows good books with a lot presented and solved analytics cases. Examples of good cases:
1. estimate average cookie lifetime for your site userbase
2. for offline store chains test impact of new interior design
3. extrapolate results of survey for all population (in survey there is x1% people from cities, y1% liberals, ... while in population there is x2% people from cities, y2% liberals, ...)
I'm not sure I see that pattern in the GDP graphs - S Asia does not show a dip in 1980, and LA/Caribbean growth restarts in about 1986 after about a 5-year dip -- but more importantly, the numbers are in US dollars (and current dollars - not adjusted for inflation). Wouldn't the numbers be affected by the value of the US dollar? I might be wrong, but would think we would want to see the numbers in purchasing power parity to see if there is an actual pattern. PS: Eg: See World Bank GNI data here https://data.worldbank.org/indicator/NY.GDP.PCAP.CD. From 1990-2019, China's per capita GDP has gone from 729 to 8242 (11x) in constant PPP int't dollars, but from 982 to 168040 (17x) in constant US dollars (and from 317 to 10216 in current US dollars, which the posted graphs use).
Current dollars *specifically is* adjusted for inflation, and won't get artifacts from the historical USD exchange rate (as the historical USD doesn't enter the picture; the only exchange rate is other-currency-then vs. the constant USD-now).
Does any genetics expert want to volunteer to be my genetics post proofreader and make sure I'm not messing anything up?
I'm not exactly a genetics expert, but I've got a PhD in molecular biology and bred mutant mice, so I might be up to the task. Are you looking for a quick check of the jargon, or a deep peer review style feedback?
I think Scott just wants to make sure he isn't committing any mistakes obvious to PhD-in-biology folks, like you.
Not sure I'm an expert, but I have a masters in biology. I'd be happy to give it a look over as a backup proofreader.
I can also pitch in - though I specialized in microbial genetics.
Sorry, didn't give background. Biochemistry PhD here.
I can! I have a master's in biology and working on a PhD
Professor of genetics, from France. I would be absolutely delighted to participate!
I just read a presumed-Eliezer story called "Kindness to Kin" and it was quite great. https://www.reddit.com/r/HFY/comments/lom9cb/kindness_to_kin/
I liked it to, thanks for linking!
Thank you for sharing that story, it was thoroughly enjoyable. Definitely agree that it’s rather Eliezer-ish, whether from actually being him, or just having a similar style.
Eliezer at some point confirmed in a Facebook post that this was indeed written by him, and this was how I even heard of the story.
Oh, ha, must've been the same for me. I added it to my reading queue without remembering why but now that you say that...
For people who read more fantasy then me:
I get the impression that there've been three main style of mainstream fantasy over the last thirty years or so, though I'm not sure if it's really a culture trend or just what I've ended up reading.
In the early nineties, you get unironic adventure fantasy set in a huge fantasy world (Think classic D&D, Dragon Lance or Wheel of Time). the focus is more on the generic gist of the world than on specific mechanics. Lots of magic and adventures, not too much time sitting around contemplating.
late 90s-2000s seems more the "gritty fantasy" era (think Game of Thrones). Lots of random brutality, occasional graphic sex, more "historical" sense in that there's less old-timey fantasy adventure values (less honour, more realpolitik and murders). Some magic, but usually more as a tool than the main story focus.
2010s-now seems more historical fiction and hard mechanics style (think Brandon Sanderson or Naomi Novik). Feels like the authors have done a lot of historical research to make their worlds historically realistic and have spent a lot of time designing formal magic systems with strict rules (not that earlier authors didn't spend a lot of time on worldbuilding, but earlier worldbuilding tends to feel more brainstormy and less processed - which one I prefer changes by mood, TBH).
Curious how accurate this feels to people and, if accurate, what you'd guess the next trend will be (seems like we're just about due for one).
Urban fantasy/paranormal romance., possibly starting with The War for the Oaks (1987). Starts with the contemporary world., and may spend the whole book there.
I believe Laurel Hamilton's Anita Blake books were the first with public knowledge of vampires, werewolves, etc., in the contemporary world rather than supernatural creatures having to live in hiding.
These books can have various degrees of grittyness, though I don't know of any that are full grim dark.
A fair number of them are in first person from a single point of view.
The thing that disappeared, rather surprisingly, was the picaresque fantasy. One or two adventurers, perhaps with loosely attached ethics, having separate adventures (not much of a story arc) in a large world of supernatural wonders. This was completely supplanted by more complex worldbuilding.
The major authors for picaresque fantasy were Robert E. Howard (Conan) and Fritz Leiber (Fahrd and the Gray Mouser).
Would you also include Jack Vance?
I'm not sure that Vance characters were as likely to be wanderers, but I don't think my memory is very complete on that.
You are right, I don't think Vance characters were that likely to be wanderers, but for me "One or two adventurers, perhaps with loosely attached ethics, having separate adventures (not much of a story arc) in a large world of supernatural wonders" is a very good fit for many of his books, the cycle of the dying earth for example.
It seems fine (although I'd put the old high fantasy as having its heyday earlier) to me, although I feel your definition of fantasy is a bit restrictive (although it does match what bookstores would mostly shelve things as).
I would add in magical realism, which I think started getting popular in the late 90s (? Not sure), as fantasy that was more "literary", usually very soft and concerned with small stories. Often from translated works ala Jorge Luis Borges, but there appears to be a burgeoning amount of native English works (at least from my perspective buying ebooks)
Also the "new weird", which near as I can tell grew out of Lovecraftian horror into its own thing. It seems to have porous boundaries and straddles soft sci-fi at times (as in much of Jeff Vandermeer's work), but books like Perdido Street Station are unquestionably fantasy. Usually very dark in tone, obsessed with apocalyptic/devolving imagery, and again often more "literary" in theme and presentation.
Does the Twilight series count as fantasy and/or magical realism?
Well, in the bookstore it'd probably be shelved as YA or romance, but I could see it counting as magical realism, although from what I've gleaned from what's trickled into the mainstream (having not read the books myself), it doesn't carry many of the hallmarks of the genre. But genres tend to be messy, porous things.
As the genre matures, it'll take on more and more of the good craftsmanship techniques that mainstream literary fiction already uses. Fantasy writers will get less and less slack for mistakes (overlong physical description, bad pacing and stereotyped characters being three major widespread flaws in fantasy). Brandon Sanderson has better craft than Robert Jordan; sorry, he just does. Each generation both learns from and builds on the previous one.
The future trend is that fantasy will begin to use modern and post-modern literary techniques. Vandermeer is leading the charge there (this is most obvious in City of Saints and Madmen), but there are others. Kelly Link is an obvious example. But you might talk about Haruki Murakami, Jonathan Lethem, China Mieville, Mark Danielewski. Stephen King's Hearts In Atlantis or David Mitchell's Utopia Avenue. The walls between fantasy and mainstream will become blurry.
"As the genre matures, it'll take on more and more of the good craftsmanship techniques that mainstream literary fiction already uses. Fantasy writers will get less and less slack for mistakes"
This is very much not how it's working out in reality. Robert E Howard repeated an adjective inelegantly here and there, but he was a far better prose writer than the stodgy brick-trilogy emitters we've been enduring since the mid-'80s. (Most likely, those repetitions were the product of him basically publishing first drafts written very rapidly for pulps, and any amount of allowance for editing would have seen him shift them himself.) Dragonlance wowed a generation by being offensively bland milquetoast, and now the rise of online fandom gives us loud, angry and voluminous apologetics for the most godawful printed-fanfic-level writing imaginable. Sanderson may be a better craftsman than Jordan, but that's only because it's such a low bar to clear. He is in no wise a better craftsman than Dunsany, Eddison, Leiber, Smith.
It makes sense to me that we would disagree about this, given that you seem to be including Eddison and Dunsany in the fantasy genre. I haven't read them, but they seem more like precursors or ancestors. It doesn't seem quite right to compare them to Sanderson; it's like asking him to stand up to Elena Ferrante.
I suspect that, as a genre emerges out of the mainstream literary works that inspired/defined it, the literary quality dips significantly. Which would make sense if fans don't read very much outside their chosen genre. And the pocket paperback glut of the 70s and the current Internet fanfiction flood wouldn't help. But despite that, I think there's going to be a trend toward advancement in any genre. I think you can see it happening in fantasy, horror, sci-fi, mystery and experimental fiction. And I think you can see the quality dip happening right now in YA, which is a much more recent genre.
I hadn't read him before, but I sampled a few pages of Robert E Howard, and I don't think his craft stands up very well next to, say, Jemisin or Rothfuss. He trounces Dragonlance, though.
But these are matters of taste. Maybe I just prefer the contemporary style, and you prefer the older style. I like my account, but I'm not too attached to it.
I can't speak to Sanderson since, while I have a vague feeling I may have read something by him, I can't call it to mind.
But the "gritty fantasy" era in the original comment is really the same genre as Low or Picaresque Fantasy, which goes back to at least the 40s with Fritz Leiber's "Fafhrd and the Grey Mouser"; maybe less overt sex'n'gore than "Game of Thrones" but the same spiced with cynicism and down among the rogues and thieves and fighting men for hire milieu than the High or Heroic Fantasy of Tolkien or William Morris (has anyone else on here read "The Well At The World's End"?)
So I don't think that it's so much "new sub-types arise" (although those do happen), but rather that the times of the real world favour different sub-genres; sometimes we want heroes who are truly heroic, other times the taste is for anti-heroes. Sometimes it's highly coloured sword-and-sorcery or planetary romance, other times it's sliderule physics hard science.
Fafhrd and the Gray Mouser were cynical about people in general and about the gods, but I don't remember whether the stories were cynical about governments.
Conan is sort of an odd example of grittiness. As a king, he's a very good king who protects his people from greedy aristocrats and protects a religious minority from persecution. But most of the stories are before he becomes king, when he's a thief and murderer and looks heroic only because his enemies tend to be even worse (and because his own murders are mostly offstage).
Conan does get a degree of character development over the entire series of stories, and he does have his own ethical code that he sticks to. He's also not a total grimdark hero or rather anti-hero, as he does show unexpected sensitivity at times, e.g. in "The Tower of the Elephant":
"Tears rolled from the sightless eyes, and Conan's gaze strayed to the limbs stretched on the marble couch. And he knew the monster would not rise to attack him. He knew the marks of the rack, and the searing brand of the flame, and tough-souled as he was, he stood aghast at the ruined deformities which his reason told him had once been limbs as comely as his own. And suddenly all fear and repulsion went from him, to be replaced by a great pity. What this monster was, Conan could not know, but the evidences of its sufferings were so terrible and pathetic that a strange aching sadness came over the Cimmerian, he knew not why. He only felt that he was looking upon a cosmic tragedy, and he shrank with shame, as if the guilt of a whole race were laid upon him.
'I am not Yara,' he said. 'I am only a thief. I will not harm you.'"
Very happy to see The Well At The World's End mentioned here! I read it a few years ago. The language and style were tough to get used to but it gave me the most Tolkien-esque feelings of anything that wasn't Tolkien. (To be clear, I mean that as a massive compliment! And it might be worth mentioning that I *didn't* have the same reaction to other works that get cited as pre-Tolkien examples of genre fantasy-- Phantastes, The King of Elfland's Daughter, or The Worm Ouroboros.)
I love "Phantastes" and "Lilith", but MacDonald is his own style and I think you can definitely see the "Scots devotee of German metaphysics" at work there as with Carlyle. I constantly have the feeling "okay, there is something going on here that is going over my head" but that is not a bad feeling.
"The Worm Ouroboros" and all the Zimiavian novels are again sui generis. I think you need a high tolerance for cod-Elizabethan (which luckily enough I have) and it is very much its own moral/ethical system, which is not what Nietzsche would call "slave morality": some people just *are* better than others, they are higher, nobler, stronger, and the values of that world are Beauty, War, Honour, and Love. Peasants gotta be in the background tilling the ground and serving the lords and ladies, but this is not a meritocracy or a Christian worldview. Everybody loves Lord Gro! I love Lord Gro, C.S. Lewis loves Lord Gro, but he's a back-stabbing schemer who switches sides at the drop of a hat, has no problem with literal dishonourable back-stabbing or poisoning or otherwise pragmatically getting rid of a rival, does not seem to have any discernible principles, and the one scrap of justification given is that he always switches sides to the *loser*, not the winner, because he has some kind of inclination towards the underdog. Which means that he may faithfully serve you as long as you are on the way to power, but once you get power, he's as likely to change over to your enemies as to stick around.
And Dunsany again is somebody different. So yeah, pre-Tolkien, it was all over the place; post-Tolkien, there is a definite model for High Fantasy, whether you follow it or try to subvert it (GRRM is subverting Tolkien, whatever else he's trying to do, it's pretty clear he's not trying to subvert Morris or Dunsany or Eddison or Branch Cabell).
I think your eras might be too short. Gritty urban SFF extends back to the 80s in its mainstream, with titles like Neuromancer. I'm also not totally sure what I would say is the main variant right now or over the last five years. I don't find that the Hugo winners align very closely with Sanderson, and neither the Hugo's nor Sanderson align all that well with my personal taste, but I still get lots of new fantasy books to read all the time. Fantasy also, of course, is experiencing the same loss of gatekeepers as everything else, I read plenty of self published authors and am no longer dialed in to publishers like Tor.
It feels like you're overfitting on too few datapoints to me. To me it feels more like the fantasy genre has a shitload of different subgenres and that occasionally an author that specializes in one such genre becomes really popular than that the genre as a whole goes through different phases.
Unironic fantasy adventures: I don't particular like this subgenre and therefore haven't read many modern examples myself, but wikipedia tells me the last Dragonlance book came out in 2008 and the last Riftwar Cycle book in 2013, so hardly early nineties.
Gritty Fantasy: Don't know if there are any good examples of this genre from before the 90's, but it's definitely still going after the zeroes. The Books of the Ancestor trilogy somewhat overlaps with hard fantasy, but the last book in that series came out in 2019, the Tethered Mage came out in 2017 and The Traitor Baru Cormorant is from 2015. Also technically A Song of Ice and Fire is still ongoing.
Hard and Historic Fantasy: I love Brandon Sanderson, but he's hardly the first person who came up with a detailed magic system. Assassins Apprentice came out in 1995 and while it may not have had as detailed a magic system as Mistborn, it does have clearly established rules and mechanics. Plus, I understand that D&D's Vancian magic system is based on the Dying Earth series, the first book of which Wikipedia tells me came out in 1950. As for well researched historical fantasy, see, for instance, the Deverry cycle (first book 1986).
To be clear, I'm not saying different genres are never created and that genres don't become more or less popular over time. It's just that the fantasy genre is really diverse and as it has grown it has become more capable of catering to different audiences.
Re gritty fantasy: depending on your value of gritty, much of Barbara Hambly's stuff is from the eighties and has a lot of death and nastiness. The degree in medieval history is definitely visible. It's hinted at to a large degree, rather than being portrayed outright, but it's there. The Guardians of the Flame series is in the eighties and makes it clear that a real fantasy world is not a nice place - rape and people dying nastily are things that very much happen. Also an example of the "college students visit their RPG world" subgenre.
I agree with you about my extrapolating from too few days points. I also agree about the "shitload of different styles" part, but it does feel like there are fashions (although I'm not sure and it might just be what I ended up reading from each period). E.g. there might have been a dragon lance book in 2008, but few people were still talking about it.
I also kind of disagree about the specific point of assassin's apprentice. You're right in it having a more rigid rule-based magic system in some ways, but I think their approach to magic is still more old-fashioned unironic adventure. E.g. the Skill isn't so much something with specific rules that characters try to take advantage of, it's kind of "magical telepathic connection that's also in some ways a metaphor for the protagonist's ability to connect to other people", and while the wit has a bunch of specific things it can or can't do, its role in the story is more about the characters connections to animals/nature vs civilization. (Compare mistborn or stormlight, where Sanderson intentionally came up with a technical magic system and then tried to think of how people would use it in practice).
(I like both approaches depending on my mood, but it does feel like fundamentally different styles).
I think mainstream fantasy is dependent on what prior knowledge people have.
If the reading public doesn't have a lot of fantasy knowledge, you get a lot of diverse works.
Once Tolkien's concepts became the one common reference that almost all fantasy readers could be expected to have in common, it becomes hard to deviate without going out of your way to ensure that people know you are deviating. If everyone thinks of Gimli when they hear 'dwarf', you need to go out of way to differentiate if your dwarves are not like Tolkien's (aside from the places where Tolkien doesn't venture, such as the classic subject of dwarven women and beards or lack thereof).
The next generations of fantasy literature (the "gritty era" and beyond) I think derive from a need by authors to do something new to sell the story. It's a lot easier to subvert the reader's expectations by changing the story than by changing the object concepts (such as fantasy races). It also helps that repeated use of the unironic adventure fantasy exposed plenty of holes in the concept for people to use as starting points for decontruction, and once those holes are opened, there's no easy way to ignore them.
To give an example of how changing knowledge impacts storytelling, there's an early D&D adventure that takes place in a dungeon that is actually a long crashed and buried space ship. If this is a new idea that the players are not familiar with, it's probably a fun adventure as the players slowly figure out what is going on. However, once the idea gets used enough, it becomes common background knowledge, and the assumption needs to be that the players may recognize what is going on and how to abuse it right off the bat, especially if the players have given more thought to the concept than the GM.
Aside from Pratchett, I've never been an enthusiastic reader of western fantasy. However, it's fascinating to see the parallels between the development of western fantasy as you've outlined it and the development of the Japanese Isekai subgenre ("Isekai is a subgenre of fantasy in which a character is suddenly transported from their world into a new or unfamiliar one"). While the basic Isekai idea has been around for a long time ("A Conneticut Yankee in King Arthur's Court" and "The Wizard of Oz" both fit the definition), it's something that as a subgenre took off relatively recently.
From my understanding, the ones written when it started to be popular enough to be broken out as a subgenre are akin to the unironic adventure fantasy. I suspect that the main point which drove the popularity of the subgenre originally was that the protagonist was a blatant self-insert vessel for the reader, which made it easy to justify the reader's knowledge of what was going on matching up with the protagonist's knowledge; anything the reader needs explained, the protagonist also needs explained.
Once the reading have enough Isekai stories down, you need to differentiate your stories to make them stand out, and hence you get "dark and gritty" as one way to do this. Part of this is that at some point, since your protagonist is a reader proxy, the assumption is that the protagonist should recognize that he is in an Iseaki. Since the unironic adventure fantasy Isekai cast the protagonist as the hero, one way to challenge this is to have him either depicted as or end up supporting the villain (if not an outright villain protagonist).
The other development that has to go along over time is that Isekai authors will introduce story elements involving modern knowledge (visible back to "A Conneticut Yankee in King Arthur's Court"), and once introduced enough that the audience remembers it, they may start asking about it when it doesn't appear if it would be applicable. For example, someone from the modern era should have at least a basic idea of disease theory (what bacteria and germs are, basic sanitary principles to prevent infections, etc); it's easy enough to bypass needing to worry about this with magic healing to ignore disease entirely, but if some sort of disease is going to play a role in the story, then any modern era character is going to be able to at least make some constructive suggestions (and I've seen several fascinating takes on this very subject). Thus, the stories gets harder over time.
I'm still coming up to speed on everything between Tolkien and Sanderson but I don't find this classification very satisfying.
One trend I've noticed, and enjoyed, in recent years is authors leaving the high-fantasy paradigm and getting their structure straight from earlier sources like myth and fairy-tale. Neil Gaiman does this a lot, and so does Novik's Spinning Silver which is my favorite recent fantasy book. Her Temeraire series, on the other hand, I'd class with Game of Thrones as a setting that's more historical-fiction than fantasy: beyond the basic conceit there's not much interest in incorporating magic with the plot, so it's presenting itself as what's likely to happen in more or less "realistic" terms (though in practice still heavily influenced by the author's preconceptions). "Hard" settings like Sanderson's, or the Eragon series, use magic more effectively but they face a similar problem of sometimes straying into territory that's more SF than F. To my mind, the distinction is that in fantasy the worldbuilding serves the plot, while in SF the plot serves the worldbuilding (and yes, this would classify, say, Star Wars as more "fantasy" than many fantasy novels).
Minor correction: ASOIAF is SciFi, not fantasy
https://www.youtube.com/watch?v=aTUbAK1DsOc
We've had some great recent guests on the futurati podcast.
Episode 29 with Jason 'the progress guy' Crawford was really illuminating. He did a good job of pushing back on my criticisms of progress, talked about how he came around to the stagnation hypothesis, and sang the glories of concrete:
https://www.youtube.com/watch?v=ga9ZT7qo8g8
Episode 30 had us interviewing one of the world's foremost experts on innovation in education policy, Andreas Schleicher, and how we can apply those lessons to the U.S.
https://www.youtube.com/watch?v=hCRmuEhxDBk
In Episode 31 we spoke with the unexpectedly funny Brad Templeton, a renowned futurist who told us why we're completely wrong about autonomous vehicles:
https://www.youtube.com/watch?v=TFsU0NUjW9U
Episode 32 had us branching out a little bit by talking to non-futurist Elaine Pofeldt, author of a book on businesses with one employee worth at least $1 million:
https://www.youtube.com/watch?v=YuvLyiSS9tU
I've been running data-analysis/decisionmaking challenges on LW for the last few months, to a small but intensely positive response. I'm posting here in case any SSC readers with a Data Science/Analysis background want to try their hands.
The first one is https://www.lesswrong.com/posts/HsxT2cpPWYzTg9tpY/d-and-d-sci; the answers are at https://www.lesswrong.com/posts/pux6NYtaFdqTwyz94/d-and-d-sci-evaluation-and-ruleset; others can be found from my lesswrong profile.
One thing that changed after 1980 is that current-dollar growth rates started having a low US inflation rate tacked onto them instead of a high US inflation rate. Switch to a constant-dollar view and things look a little different.
Talking of solipsism, as somebody is below, reminds me of the story about the professor of solipsism in Cambridge university. I forget the name. He lived to a fine old age, 95 or so, and never married so he had an apartment to himself on campus. The students and faculty took really good care of him, visiting him every day in his retirement, keeping him company and making sure that any illness was promptly treated.
“After all”, they surmised, “ when he dies it’s curtains for the rest of us”.
I wonder if someone can shed some light on why physicist Richard Feynman relied on older mathematical techniques and didn't trust the newer techniques. My question is prompted by this quote from Stephen Wolfram (who was a physics grad student at Caltech while Feynman was a professor there): "It’s kind of interesting to look at [Feynman's handwritten note with calculations related to a Feynman diagram]. His style was always very much the same. He always just used regular calculus and things. Essentially nineteenth-century mathematics. He never trusted much else. But wherever one could go with that, Feynman could go. Like no one else." I assume that Feynman understood how to use techniques like complex analysis and tensor calculus. Anybody know why Feynman didn't trust them? Wolfram's article is here: https://www.wired.com/2016/07/my-time-with-richard-feynman/
I think you have to filter that assertion through the fact that Wolfram is making it, and if there is anyone who has staked his reputation and world-view on New Math -- and especially math that has to be done on a computer -- it would be he.
Where does Stephen Wolfram suggest that Feynman didn't trust complex analysis or tensor calculus?
He doesn't. I was trying to think of techniques that were developed after the 19th century that might be relevant to physics. It was my speculation.
Complex analysis is squarely 19th century, I think. From context, my guesses about what Wolfram meant are that Feynman didn't like using computers to do calculations, or possibly that Feynman did not appreciate Wolfram's "new kind of science" schtick and its usefulness to physics (i.e. this is just Wolfram's "unappreciated genius" self-narrative).
You might want to bear in mind Feynman invented the most original and fruitful mathematical physics technique to have appeared in the last century or so (his path integral reformation of quantum mechanics).
Feynman mentions in Surely You're Joking, Mr Feynman that he learned calculus from an old textbook his physics teacher gave him, and as a result he approached problems with a "different box of tools." Which got him a great reputation for solving problems because if someone came to him for help with a math problem they'd already tried everything they could think of, so Feynman just had to try something in his toolbox that they didn't know about.
He doesn't mention anything that he didn't trust, but I imagine his approach to solving problems would look unusual to other mathematicians if he commonly uses methods he learned out of an old textbook.
What are the best resources for a balanced, thoughtful perspective on whether an undecided person should have kids? I've seen arguments for both sides - I can make arguments for both sides - but I've never seen a rigorous, evidence-based approach.
Paul Graham has addressed this through his essays. http://paulgraham.com/kids.html
I read that essay as mostly anecdotal, not evidence-based.
Thanks, Ayush. I'd read that one before and thought it was one of Paul's best. He acknowledges his bias in the essay but still offers one of the most thoughtful approaches to the topic.
What kind of rigourous, evidence-based approach would you like?
There's certainly studies looking at (e.g.) self-reported happiness of parents vs non-parents, but the results you get depend heavily on exactly what question you ask and when you ask it. Having kids is a matter of trading off most of the good things in your life, in exchange for other good things.
Thanks, Melvin. Having read some of the research that compares self-reported happiness of parents and non-parents, parents appear to be less happy than those without kids but report higher levels of meaning.
This seems like a mixed bag - as you'd probably expect - yet the social expectation is that most families will want or have kids. That feels heavily skewed for what I perceive as a fairly balanced risk/reward profile.
I guess I'm trying to find something that can thoughtfully connect these two facts and offer a holistic view of whether the benefits are worth the drawbacks with something less subjective than "When you have kids, you'll understand they're the greatest," or "I pity all the families having to corral their kids on a hot summer's day while I eat ice cream and have infinite freedom."
Do you know of anything like that?
I have found meaning far more important than happiness in my life. Before getting married, I was honestly just drifting for lack of motivation. I played a lot of video games and probably would have said I was happy, but looking back I wouldn't take that again at all. I am probably slightly less likely to say I'm happy now, but I much prefer life with a family.
I guess what I'm saying is make sure you're asking the right question, and understand the implications.
You're asking the wrong question. The main argument for having kids is that it's what we're hardwired for in terms of maximizing our personal life meaning. I guess I could try to kludge together some objective measure of meaningness, but that's besides the point.
Having kids is the most subjectively rewarding experience in life. Demanding rigor and evidence is completely missing the point.
Fair enough, Nick. I suppose the most important thing I'm trying to find is a balanced position - because everyone is necessarily on one side of the fence or the other, folks usually pretty tribal on this topic. If rigor and evidence aren't the order of the day, I'd at least like to see a balanced argument. Do you know of any of those out there?
I have a tough time imagining one. Some people don't seem to derive much meaning from kids, or don't pay attention to them and then end up disappointed. There's a lot of room for positivity and growth, but a lot of room for negativity and failure. Your experiences will vary dramatically from everyone else you ask; I sincerely doubt the question is answerable in a way that, after having kids, you would look back on as being useful.
On the meaning side, I can't really overstate it though. Prior to having kids I had occasional moments in my life that touched me deeply. Campfire moments, I think of them as, because a lot of them occurred around fires with friends. Probably a few dozen altogether over my first 33 years.
Since having kids it's a constant, daily and nightly stream of moments that put all those to shame. When I say it's what we're hardwired for I mean that in a way that doesn't make sense until you experience it. There's nothing else to compare it to.
All the rest of the pluses and minuses are things that can be written down on a pros and cons table, and compared, and balanced, and argued about. You're probably already familiar with everything that can go on that list. Financial costs. Teenagers. Grandkids around your bed when you die. Stuff like that. The one piece that I think gets missed is the meaning piece, because it's impossible to convey from people with kids to people without them.
The good news is that, whatever you decide, you'll be happy. In one case it'll be because you tapped a deep well of human experience that you never would have otherwise known, and in the other, it's because you didn't and you won't miss what you never experienced.
As a counterpoint, I haven't found life with kids any more meaningful than life without. It's more fun, interesting, etc. and I wouldn't trade it for the world, but I haven't experienced a perceptible increase in meaningful moments the way you describe.
Maybe an interesting way to rephrase the decision - and make it a bit more tractable - might be ‘how will having kids be for ME’?
That way you have 2 ‘smaller’ but maybe more objective questions to answer: what is my personality/preferences, how do kids affect one’s life typically, and so to what will that ‘add up’?
Like (hey, just thinking out loud :)): try taking the free mbti/personality test at 16personalities.com and then read until the ‘parenting’ section, which might make the kids question a bit more tangible to reason about?
Kids are a heap of fun when they're young. The adolescent years are a trial, however.
I think there's a lot of variation among adolescents about how difficult they are.
It's possible that how they're treated when they're kids has something to do with it. I'd be interested in data on the subject.
From what I hear (and extrapolate from when I was a kid) they're great at first, then terrible, and then once they're 26+ they become great again, if you raised 'em OK.
I believe this, but naturally I would.
(I wasn't "a teenager", have gotten on very well with my parents my entire life, and was raised by people who viewed kids as miniature adults, and treated us as maturely as we acted. Anecdote isn't data, but...)
I am surprised no one has recommended Brian Caplan's "Selfish Reasons to Have More Kids."
His recommendation is up front in the title, but his book is full of references to lit (recollections from skim-reading long ago)
It's an irreversible decision that will potentially destroy any joy you get out of life. So an undecided person should lean towards minimizing risk
I’m considering quitting my job and going to a bootcamp to learn data analysis/ML/AI. Anyone know of a good program in the NYC area? Preferably with an I person option.
Insight is good, if you can get in.
Would love to read a deep dive into the prospects for GLP-1 receptor agonists as weight loss medications. Semaglutide 2.4 mg generated some buzz recently, and now apparently there is Tirzepatide. Has anyone tried semaglutide or liraglutide, or prescribed them to a patient? Lots of questions about side effects, long-term use, cost/insurance coverage, etc. But these trial results seem remarkable. Are we on the verge of having obesity treatments-other than bariatric surgery-that actually work?
Good pillow recommendations?
Some pillows last longer than others, and lifespan correlates reasonably well with price. But if you're asking what is the most comfortable pillow, I'm firmly of the opinion there's no such thing. What's comfortable depends entirely on your preferred sleep position and what feels good to you. I've had all kinds of pillows -- down, form, fiber-filled -- cheap and pricey -- soft and firm -- and my current set-up is the most comfortable I've ever had. It's a fat couch throw pillow, topped with an ancient, deflated-looking pillow that seems to be filled with crumbed foam. For me it's perfect, and I'm prone to neck cricks and general pillow discontent, so that's saying a lot. Just experiment around.
I've had a pretty good experience with Allerease down alternative pillows. They're the right mix of firm and soft for my head.
One caveat is that I've noticed the down alternative pillows don't keep "cool" as well as regular down ones. If you like having a cool pillow against your head or face when you lie down and try to sleep, the down alternative pillows will warm up faster.
Let me be more specific: anyone have a recommendation on a thick but soft pillow that whisks heat away from my head and stays cool without being too firm?
OP, I suggest describing your perfect pillow and seeing if anyone can recommend a model. Cool idea though.
Almost orthogonal to your question but hopefully useful: my wife got some silk pillowcases and they are wonderful. They are a little more work to wash, but we each have two pillows and only the "top" one of each pair is wrapped such, reducing the workload.
Thanks for the replies just wanted to see what recommendations people threw out. It just seems one of those products where there’s loads of options and it’s hard to pick.
Agree that it’s pretty individual but that down/down alternative is tried and true. And apparently I should dress it up in silk ;)
If you're a side sleeper, I recommend Pillowcube.
How good is it? As a side sleeper I'm intrigued, but as a miser I don't like taking risks on new products.
I am extremely happy with it. Best purchase I've made since my Allbirds. I sleep through the night much more often now. After I got one, my wife tried it for one night, and we immediately bought one for her as well. If I thought they were no longer going to be made, I would try to buy enough of them to last for the rest of my life.
Can you comment on whether you think that your Survive vs Thrive theory of Right vs Left politics works for covid responses?
I can come up with several ways that you might respond, but it seems better to ask you yourself.
Scott has briefly addressed that in the past: https://slatestarcodex.com/2020/04/19/open-thread-152/
There, he acknowledges that the response to Covid appears to be a counterexample to his thrive/survive model. However, in the comments on that open thread, several posters point out that since Covid mostly endangers the weak/elderly and isn't an existential threat to humanity, it actually fits in the model that Republicans would be willing to sacrifice the weak and elderly so that the rest of us can continue to thrive, while Liberals are more willing to let the economy take a hit in order to save them.
(Also, it's worth noting that Republicans tend to live in rural areas and are more likely to have jobs that cannot be done from home, so Liberals are both more affected by the pandemic and less affected by the lockdowns.)
The US is 82.5% urbanized, so unless there's something very wonky with that definition you might be overstating your case.
I probably am, but "tend to" is a vague enough wording that it provides me with quite a lot of ambiguity to hide behind.
There's two different definitions of "urban" here. The definition that yields the US as 82.5% urbanized counts basically anything where residences are organized (possibly even everything within an incorporated community as well).
But the definition that lines up well with the political divisions is tripartite, into "urban", "suburban", and "rural". The way I've heard this divide jokingly described is: if you can step outside naked and no one sees you, you're rural; if you can step outside naked and the neighbors call the cops, you're suburban; if you can step outside naked and the neighbors ignore you, you're urban.
Everything "urban" or "suburban" counts as "urbanized" in the former definition. "Urban" areas usually vote 80% for Democrats (though some parts of many cities are as low as 60%, and there's often a few rich neighborhoods that vote for Republicans). "Rural" areas usually vote 80% for Republicans (though there are a few regions of Black or indigenous populations that vote for Democrats). "Suburban" areas can go either way (often depending on population density).
Thank you ! I hadn't seen or had forgotten that.
Possible answers for my own question:
(1) Survive vs Thrive only claims to be maybe 80% accurate and this is one of the times it fails.
There might be evidence for this internationally. Trump and Bolsonaro made the news for having lax covid response and bring right wing. Abe is right wing and had a strong covid response, as did some central European countries. If there isn't a correlation between how right wing a country's government is and how lax their covid response is, then this is probably coincidence. The alternative history where Ben Carson became president instead of Trump is also interesting to consider here.
(2) The Right looked at a 1% mortality rate and thought it was nothing to worry about. The Left looked at a 1% mortality rate and decided it was a crisis. I'm not sure if the people making the decision were numerate enough to think this way. Would people's responses be different if covid had a 5% mortality rate?
(3) The Left's support for and the Right's opposition to government agencies was more important than Survive vs Thrive. This also seems to be what's going on with school choice. Maybe Survive vs Thrive works less well in the short term or for particular issues and better for grand strategy questions or issues that persist for decades.
Has anyone had any experience playing a digital piano keyboard (one of the 88 key models with a weighted key action such as Yamaha, Casio, Roland, Kawai etc make) on top of a regular office desktop with the height adjusted down to 24 inches or so, rather than on a dedicated stand? I'm debating whether to try and refactor my sit/stand desk setup to include my piano-- I compose with piano and computer frequently and it'd be nice to do it with both of them at the same adjustable height. And it seems like the key height should be OK in "sit mode" if I adjust it all the way down to a 24 inch desktop height, but I don't know whether I need to worry about instability of the keyboard on top of the desk, or resonances with the desktop, or some other such complication.
> I compose with piano and computer frequently and it'd be nice to do it with both of them at the same adjustable height.
Do you have your piano literally right next to your computer? If so I can't foresee any harm from making them the same height if you do it right. (As for resonances or whatever, if the piano + supports and computer + supports aren't touching each other, there shouldn't be much of an issue.) Otherwise you should keep them as is because your hands will have plenty of time to adjust to a different height while you move and changing the height of a piano or computer desk is generally hard.
The tax code was much more favorable to you if you started a company in the Postwar Era, versus getting rich off regular income. Same time that the top tax bracket was 91%, the capital gains tax rate was 25% AFTER excluding 50% of the gains from taxation altogether. You'd see Hollywood types try to turn their income into that to reduce taxation, by using stuff like setting yourself up as a corporation selling shares of stock (with said company "leasing" your services) or the Oil Depletion Allowance (IE you exclude something like 28% of the income from investing in Texas oil wells from taxation).
re 3, big oof
Join our growing team
Failed to load postings. Something went wrong
This is exactly why they need you to join their team!
Regarding the growth slowdown, a book that I think does a good job of painting the big picture is 'Fully Grown' by Dietrich Vollrath. According to him: mostly demographics as well as the shift to services where productivity growth is inherently slower. His biggest takeaway is that none of this is inherently a bad thing.
Anyone looking for a 'Priest' to the Rationalist 'Wizard' are encouraged to check out Mr. Tom Murphy at https://dothemath.ucsd.edu/ An Astrophysics professor come.... neomalthusian (?), he is back to blogging in support of his new textbook on economics and the physical world which can be found here. https://escholarship.org/uc/item/9js5291m#article_main
Longtime rationalists will enjoy spotting the motte and bailey arguments, the strawmen, assuming the conclusion, and a host of other illogical and sometimes bizarre reasoning.
On the plus side, his blog does allow comments and he frequently engages his readers, so if you are interested in trying to argue the Wizard side, you might get some traction there.
When publishing predictions, Scott always mentions that some people consider 50% predictions meaningless as they indicate lack of information. Could someone please explain the reasoning behind assuming it is meaningless? Because, suppose I have a playing die, and believe it is biased so I expect it to roll up the number 6 with 50% chance (and have no idea about relative chances for the numbers from 1 to 5). But this clearly means I have some information, and if it is true, I could benefit from betting on it against someone who e.g. thinks the die is fair. So what's the reasoning behind the opposite view?
I think it's based on an idea that everything is either true or false, and so there's a natural two-element partition for every proposition. (I may not be entirely charitable here, because I'm not sympathetic.)
But I think the particular discussion at hand is based on the fact that he evaluates his predictions based on "calibration". Instead of measuring how close each *particular* prediction is to the truth (and perhaps, comparing to someone else's prediction on the same proposition), "calibration" says to group together all your X% predictions and see how close you are to X% true. Scott usually does it by grouping together the X% predictions *and* the negations of the (100-X)% predictions, to get a larger dataset. But when you do this, you automatically group together the 50% predictions and the negations of the 50% predictions, and thus automatically get 50% correct. (Though I think for this one particular bin, he just chooses the 50% predictions and not their negations, which then invites the question of how he chose which one of the two to be the prediction.)
Scott does yes-or-no predictions, so a 50% prediction does no better at helping others find the correct answer than random chance. Even if you have information saying the chance is 50%, you perform no better than someone with no information at all.
This really depends what your priors are. If you have literally no information (and use a uniform distribution as your default prior), then a 50% probability gives you no additional information. But most people have some information already, like a prior that a 6-sided die is at least somewhat fair. In this case, a prediction that it lands on 6 with 50% probability is meaningful.
He also sometimes fakes a probability distribution with predictions, like the chance that covid deaths are above 10/100/1000k. In those cases, 50% acts as another data point.
This isn't the same thing. Scott is predicting binary outcomes with a probability of 50%. The equivalent on a six-side die would be predicting 6 comes up with 1/6 certainty. But note that, in this case, you can still calibrate. If you make this prediction across six different die and they all all come up six, there is a decent chance you're wrong about die in general and need to update that some common manufacturing might exist in all of them making 6 more likely.
The issue Scott runs into is really specific to binary outcomes. Say he makes 4 predictions. He's 50% certain that Covid deaths will be under 500,000, 50% certain that the US murder rate will be under its 30 year low, 50% certain that Biden wins the election, and 50% certain that GDP growth is less than 5%.
How does he calibrate? By this phrasing, he got 100% correct, not 50%, so he's miscalibrated. But phrase two of those the other way (50% certain Covid deaths will be over 500,000 and 50% certain Trump wins) and now he got 50% of them. Perfectly calibrated.
When you can be either perfectly calibrated or perfectly miscalibrated making exactly equivalent predictions simply depending on how you choose to phrase them, it isn't that the predictions are meaningless. It's that the attempt to use them for calibration can't work.
Basically what Kenny said.
Meant to write "over its 30 year low" but that proves the point because again, it's the same prediction!
"Joe Biden will die this year" is a binary choice, but if I were to be 50% confident in it, that would be telling you a lot of information!
Minor complaint, when I post I'll sometimes not be able to continue on ACX, and I'll have to reload it. (Which means I lose my place in the thread stream.) I think this is related to my keeping my mouse over the post button after posting. A little slashed red circle appears. And if I leave it there, it stops working. Does this happen to others? I've got laggy satellite internet.
I've got good cable internet, and it happens to me too. I haven't noticed the slashed red circle.
Saw the red circle just now, but it worked fine.
Thanks. The red circle thing was a red herring. I've tried to make it fail that way without success.
I'm reading the latest Michael Lewis book, *The Premonition*; it's a post-mortem of the US covid response. I've only just started, but so far it's great.
His last book, *The Fifth Risk*, was published in 2018, which among other things was about how the Trump administration had crippled the United States' ability to deal with catastrophic risks. Turns out he was right.
But, as he says in the introduction of this book, it turns out that the rot spreads much deeper; "Trump was a comorbidity".
Right now it's just introducing some of the characters and what they were doing before the pandemic. One of them is a public health official in Santa Barbara, Dr. Dean, and the book goes into some detail about what that is like — and what the CDC is like. One example of many: Once there was a patient, a college student, who had symptoms of meningitis B and was in a state of shock (they had to amputate both his legs to save his life). Dr Dean set out to control the outbreak — and it's astounding just how useless the CDC was throughout all of it. I don't think blockquotes markup works on Substack, sadly, but here:
> She got on the phone with the main guy at the CDC and his silent crowd. The guy strongly disagreed with her doing anything. "What he actually said," recalled Charity, "was, 'That decision is not supported by the data.' I said, "Oh really — there is no data." She outlined a plan she'd created: thin out the dorms by moving some of the students into hotel rooms; shut down the intramural sports teams; and administer a vaccine that had been approved in Europe but that the FDA had not yet signed off on. "The CDC guy said, 'We're not going to do any of that, and if you do that, we're going to put it in writing that it was your decision and we disagreed with it,'" Charity recalled.
> [...] But in the end the campus ignored the CDC and did everything Dr. Dean recommended. From start to finish, [...] "the CDC wasn't pleased with her. The CDC kept saying 'There is no evidence to back it up.' They didn't have any evidence, because there is only one case every four years."
> The root of the CDC's behavior was simple: fear. They didn't want to take any action for which they might later be blamed. "The message they send is, We're better than you and smarter than you, but we're letting you stick your neck out to take the risk," said Charity.
> Charity never would know which of the measures she took had controlled the disease; she knew only that all of them together had. [...]
> Two years after the UCSB meningitis outbreak, the CDC finally published a report on how to deal with a meningitis outbreak on a college campus. On its list of best practices were most of the things Charity had done at UCSB. [...]
> Charity had washed her hands of the CDC. "I banned their officers from my investigations," she said. The CDC did many things. It published learned papers on health crises, after the fact. It managed, very carefully, public perception of itself. But when the shooting started, it leapt into the nearest hole, while others took the fire. "In the end I was like, 'Fuck you,'" said Charity. "I was mad they were such pansies. I was mad that the man behind the curtain ended up being so disappointing."
Seeing as "listen to the experts" is a common refrain, it's important to remember that *credentialed, educated, powerful, and authoritatively titled* is not the same thing as *expertise*. Those men at the CDC were credentialed and titled. Dr. Dean is an expert.
The SNAFU principle (people are unlikely to tell the truth in a high-punishment environment) needs to be better known.
It's strange, people are doing to the CDC what they did to FEMA about 10-15 years ago. Neither agency was ever meant as a first-line responder to emergencies, for the simple reason that (domestic) emergency response is a state and local responsibility, that being the nature of a federal republic. There's nothing in the Constitution that gives the Federal Government the responsibility *or* power to do emergency response (except for external threats, e.g. invasion and war, which the military handles). So it's been a state and local responsibility for ages.
Until Katrina. Then all of a sudden people thought this Federal agency which existed to disburse emergency loans and write down best practises for sharing among agencies (FEMA) should *be* a front-line responder, and were shocked that when asked to be that, they were awful.
Same thing is happening now to the CDC. They were never meant as a first-line responder to pandemic, that has always been understood to be a state and local responsibility, like all other emergency response. The CDC collects data, mostly, does some research, hands out grants for research, and writes down best practices (long after the fact) for sharing. They were never meant as a first-line responder, and that of course is reflected in who they hire and their institutional reflexes (well illustrated by the anecdote).
And yet...people are now acting as if they *should* be a first-line responder, and finding it outrageous that when pressed into the role, they're terrible. It's strange, it's almost like people have forgotten that we live in a federal republic, with some pretty strict restrictions on what the Federal Government *can* do.
There's certainly an argument to be made that for pandemics we *ought* to have a national-level first responder, since disease doesn't respect state boundaries, but, tricky Constitutional questions of legal authority aside, this is not what the CDC is, or ever was.
(1) The CDC should provide accurate information about the control and prevention of diseases. It instead provides inaccurate information.
(2) The CDC sure thinks they should be a first-line responder. Remember when they made a coronavirus test kit, and the FDA banned every test kit (including the WHO's) except theirs — and then the CDC's was defective? That's not unfair societal and political expectations being *imposed upon* the federal government, that's the federal government fucking things up and accidentally killing people.
In the recent book review "The Wizard and the Prophet" I noticed almost no pushback in the comments from people with a more radical environmental perspective, i.e. people who take a non-human centered view of the environment.
Is this a perspective that is just not common in the rationalist community? Because it seems like a very important addition to the conversation. The 'Wizard' community has many successes if you take a human-centered approach, but if you expand your scope to the natural world as a whole I think it is clear we are not innovating our way out of environmental damage.
Examples:
1. Deforestation continues on a massive scale to this day. The rate has declined some in the past couple decades, but rainforests and other habitats are still being destroyed daily at an astounding pace.
2. Evidence of population drops for insects appeared last year. The data was likewise astounding, claiming upward of 1-2% population drops per year, or 30-40% declines since the mid 20th century. There is not enough data to be confident about how this is taking place, whether it is concentrated in certain areas/ecosystems or more widespread, but that a problem exists seems fairly undeniable.
3. Overfishing is still occurring worldwide. Some specific commercial populations have recovered when managed, and some countries are practicing generally responsible fishing, but overall it is still a large issue.
4. Climate Change, which was cited as perhaps the 'Prophet' side's best issue, will likely continue to exacerbate all the issues which we are already creating throughout this century.
5. Plastic and other pollutants in the ocean are having other large effects on marine ecosystems.
I suppose a 'Wizard' point of view could point out that many of these problems do have technical solutions in theory and even in practice in some locations, but I think the 'Prophet' would counter that just because solutions exist does not mean they will be applied, and that they in fact will not be until a large amount of damage is done. An irreversible amount of damage.
I just really didn't like that book review, or it's lens for approaching our interactions with the planet. We are having a profoundly negative effect if you look beyond the human view.
I think you bring up good points, but they may be better discussed in an even numbered thread
As far what is or isn't common in the rationalist community, I think this adversarial collaboration Scott hosted on his old blog is a good sample of how a lot of rationalists approach the problem of valuing life generally.
https://slatestarcodex.com/2019/12/11/acc-is-eating-meat-a-net-harm/
The main takeaway is that the paper takes a long time trying to determine what animals are conscious, and how conscious they might be. One takeaway quote:
"Of particular relevance is the conscious experience of pain and suffering, which we regard as morally undesirable when it occurs in ourselves or others. Most animals have damage sensors, but triggering these may not result in the subjective experience of ‘suffering’ if the animal is not conscious or the stimulus is a constant presence that has been accustomed to.
And this is the absolute crux of this investigation; if animals suffer under current farming standards to the point of preferring non-existence then there is a moral burden on meat eaters to justify eating them."
It seems to me (I'm not a rationalist, I just like to hang out with them) that most rationalists are utilitarians. If they seem human-centric it's likely because humans are the only creatures whose happiness or suffering we are intimately acquainted with. As such, a topic such as overfishing is only really relevant if you can show that it would hurt humans, since as far as we can tell fish likely aren't conscious and might not even feel pain in a way that is meaningful for a utilitarian calculation.
I'm not a utilitarian, so I can't really speak further on the specifics.
For people who like looking at art on Instagram, here's mine: https://www.instagram.com/willa.applegate/ Thanks for checking it out :)
Both the reading and writing experience are remarkably bad for such an apparently well funded website as Substack.
I wonder if we can get any of the Substack people to explain their expectations for what it would be like? Their Help page really isn't; it's geared towards writers (for how to publish on Substack) and the section for readers is all about "how to pay for a subscription".
It doesn't appear to be the case that they anticipated lively debate in the comments sections, so no edit, link, formatting, etc. functions. That's why I'd like to know what the mental model coming over was - did they think of it more as a newspaper online page, where subscribers can leave comments below an article or opinion piece, but nothing more and certainly no kind of conversation (and where the paper or media owners can delete or close down comments they don't like)?
No joke--I've been wanting to visit India and Brazil. Based on the likeliest trends, when will COVID-19 levels decline enough in those countries for me to be able to visit them and not find half of the tourist attractions and restaurants closed?
Had mild success with posting job ads in an open thread before. I know back then this was accepted, not sure if it is anymore (if not, please delete, and my bad). Anyway:
I'm curios if anyone here would like to lead an infrastructure/ops/devops team (starting as 1 man job, will hopefully soon grow into a 2-3 people team if all goes well).
The product is something called mindsdb, fully open source "generic" machine learning. The thing you'd be working on is our cloud offering (so not open source, but running 98% the same code as the open source version)... which is rather important, since it's how we can actually afford to develop a fully open source product long-term.
Fully remote, offers start 80k/year but the cap is much higher (basically low pay for the US and insanely high pay for literally anywhere else in the world except for the trade-eco hubs), work contract is whatever (employee via local company branch, contractor, working via a 3rd party company... we don't care really, unless you're US or UK based, in which case I think it has to be a standard employment contract). 3-5 meetings a week (total 2-4 hours) + slack for other communication, people aren't sticklers for immediate answers but ideally you an be kind-of online around EU and/or US friendly working timezones.
Please ping me if it sounds interesting, it's honestly much harder to find people that can to devops tham I thought :)
Here's a challenge to the "Dark Forest Theory" (https://bigthink.com/scotty-hendricks/the-dark-forest-theory-a-terrifying-explanation-of-why-we-havent-heard-from-aliens-yet)
If the Theory is right, then there is still a way for one alien species to take over the galaxy without attracting attention. It would involve a "two-phased attack" with self-replicating nanomachines. An alien species could, over a long time and with great secrecy, seed every solar system in the galaxy with its own Von Neumann probe, which would contain self-replicating macro- and nano-machines. Once every solar system had a probe, the aliens would send out a signal, and all of the probes would start self-replicating. They wouldn't just make "Gray Goo" copies of themselves--they might make soldiers, weapons, and other advanced technology. Any enemy aliens would be overwhelmed, or at least forced to reveal themselves to fight back.
This could be done without revealing the planet the probes trace their origins to. One day, it would seem like the whole galaxy suddenly came alive at once.
If this theory is right, then we should also not expect to hear back from any alien civilizations, and also we should be even *more* scared about the prospect of there being any.
(Honestly, given the multitude of proposed solutions to the Fermi paradox and our extreme lack of concrete data about it, I'd put very low weight on any hypothesis. The solution I like the best, though, is still https://arxiv.org/abs/1806.02404.)
Cryptocurrency/DeFi/etc has been on my metaphorical radar for quite some time, but I feel hesitant about investing/working in it. Here are some of my worries, in (very roughly) descending order:
1. Coinbase/Binance/etc doesn't allow people under 18 to join and trade (which is think is for legal reasons), and I'm one of them. (And, more generally, the legality of using cryptocurrencies/DeFi/etc.)
2. Even if I don't use a company like that, I might need to install special software in order to do this properly, and I'm afraid of messing up my computer beyond repair in the process.
3. Not everyone accepts cryptocurrency as payment for commerce.
4. I'll have to pay taxes on it, which I don't know how to do properly.
5. There are a lot of criminals involved in this area.
6. Someone recommended (https://www.youtube.com/watch?v=sEtj34VMClU) that I use relatively extreme security measures before even beginning (such as writing passwords in a paper notebook, using 2FA everywhere), and I have no sense of calibration on whether this is too much or not enough.
7. Hackers/scammers/ordinary market fluctuations could wipe out any profits that I make.
8. Mining cryptocurrency takes a lot of energy, which produces a lot of environmental damage. Renewable/zero-emissions energy sources could mitigate a lot of it, but it's not exactly straightforward to find or obtain.
9. I don't understand any economics.
Can you clarify whether these issues are as important as I think and recommend ways I could take care of the important ones? Thanks.