They are decreasingly prescribed by physicians despite their extreme efficacy. Is it just not worth it anymore for pharmaceutical companies to produce them?
Follow-up question: where can I get it? (Darknet recommendations?)
I've had asthma since I was a kid but it's been mostly mild. I'd have an attack maybe one or twice a year and it was harder to breath but never to the point where I felt like I would actually panic or pass out. As a kid I had an inhaler but I never bothered with that as an adult.
But yesterday I had a pretty severe attack. Worse than any I'd had before to the point where I had a definite feeling of panic that made it very hard to control my breathing. The panic would constantly make me try to inhale my next breath before I could finish exhaling my previous one. It lasted for hours but it did subside eventually.
So I went to the pharmacy to buy an asthma inhaler and they straight up refused to sell me one. Prescription only apparently. But the problem is I don't have a doctor or internet (I'm on wifi at a café atm) so no virtual visits either. So that leaves the only (official) option of sitting for eight to ten hours in a room full of sick people so I can talk to a doctor for five minutes and get the stupid piece of paper that permits me to buy an emergency inhaler in case of another attack. Of course if I catch something while I'm there that could well trigger an attack in itself.
Ah well. I just looked up asthma inhaler prices online and there's a wide range of prices but possibly I couldn't afford one anyway. (the markup on some of them must be in the range of 10,000 percent markup) So I guess I'll just have to take my chances.
So I'm feeling a little bitter atm. I'm in BC, Canada for those who would like to make note of the data point wrt the relative state of health care in various countries.
Thanks for your concern. I was frustrated when I wrote that. Probably waits of eight to ten hours are not typical though they definitely did happen as least as of about ten years ago. Not sure how things may have changed in the covid era as I have not set foot in a hospital in years.
"Of the total 139 participants who underwent randomization, 118 (84.9%) completed the 12-month follow-up visit. The mean weight loss from baseline at 12 months was −8.0 kg (95% confidence interval [CI], −9.6 to −6.4) in the time-restriction group and −6.3 kg (95% CI, −7.8 to −4.7) in the daily-calorie-restriction group. Changes in weight were not significantly different in the two groups at the 12-month assessment (net difference, −1.8 kg; 95% CI, −4.0 to 0.4; P=0.11). Results of analyses of waist circumferences, BMI, body fat, body lean mass, blood pressure, and metabolic risk factors were consistent with the results of the primary outcome. In addition, there were no substantial differences between the groups in the numbers of adverse events."
The advantages of fasting would be easier compliance, appetite suppression, autophagy and immune downregulation. There might be some anti-insulin resistance magic going on as well but that can be rolled into appetite suppression.
From the abstract, they eliminated the first two through study design and didn't measure outcomes of the latter two, outside of very distant proxies. Unfortunately, scihub doesn't have the article so I can't comment in greater depth.
EDIT: I just realized they tested 16h fasting, from a hormonal standpoint that's regular calorie restriction with extra steps.
It was published in 2009. What has happened since its writing that should affect how I read it? I'm particularly interested in (a) results McGilchrist relies on that have since failed to replicate and (b) well established (replicated) results that postdate the book and should meaningfully affect my reading.
With Google having become far less useful recently due to a severe bias toward mainstream news and SEO-optimized sites, I'm finding it especially urgent to learn news ways of finding useful and detailed information.
Apart from Google Scholar, what techniques have you discovered for finding useful, detailed information sources *that you didn't know about before*?
Depending on the topic, my first choice is usually "<google-query> site:reddit.com" because it's basically a bunch of topic-specific forums on one domain. That may or may not qualify as "useful", "detailed", or "didn't know about before", but it works for me 80% of the time.
Looking for a link from a relatively recent post, about how if people are truly on the fence about making a change in their life ( breaking up, moving, changing jobs, etc. ), there was a study that flipped a coin for them, and found that people who made a change were ultimately happier. Where was that?
Here's a list of all the books reviewed, without subheadings. I don't know if anyone else wants this, but I found it a useful tool for browsing.
A Canticle for Leibowitz by Walter M. Miller Jr
A Connecticut Yankee in King Arthur’s Court by Mark Twain
A Failure of Nerve: Leadership in the Age of the Quick Fix by Edwin H. Friedman
A History of the Ancient Near East
A Secular Age by Charles Taylor
A Supposedly Fun Thing I’ll Never Do Again by David Foster Wallace
A Swim in a Pond in the Rain by George Saunders
1587, A Year of No Significance: The Ming Dynasty in Decline by Ray Huang
Ageless: The New Science of Getting Older Without Getting Old, by Andrew Steele
Albion: In Twelve Books
An Education for Our Time by Josiah Bunting III
An Empirical Introduction to Youth by Joseph Bronski
Anthropic Bias by Nick Bostrom
At the Existentialist Café by Sarah Bakewell
Autumn in the Heavenly Kingdom by Stephen Platt
Bronze Age Mindset
Capital and Ideology by Thomas Piketty
Civilization and Its Discontents by Sigmund Freud
Come and Take It: The Gun Printer's Guide to Thinking Free by Cody Wilson
Consciousness and the Brain by Stanislas Dehaene
Cracks in the Ivory Tower: The Moral Mess of Higher Education
Deep Work by Cal Newport
Development as Freedom by Amartya Sen
Disciplined Minds: A Critical Look at Salaried Professionals and the Soul-Battering System That Shapes Their Lives by Jeff Schmidt
Economic Hierarchies by Gordon Tullock
Exhaustion: A History by Anna Schaffner
Facing the Dragon: Confronting Personal and Spiritual Grandiosity by Robert Moore
Fashion, Faith and Fantasy in the New Physics of the Universe by Roger Penrose
Frankenstein by Mary Shelley
From Paralysis to Fatigue: A History of Psychosomatic Illness in the Modern Era by Edward Shorter
From Third World to First: The Singapore Story: 1965-2000 by Lee Kuan Yew
Future Shock by Alvin Toffler
God Emperor of Dune by Frank Herbert
Golem XIV by Stanisław Lem
Haughey by Gary Murphy
History Has Begun by Bruno Macaes
How Solar Energy Became Cheap: A Model for Low-Carbon Innovation
I See Satan Fall Like Lightning by René Girard
In Search of Canadian Political Culture by Nelson Wiseman
Industrial Society and Its Future by Ted Kaczynski (also known as the Unabomber Manifesto)
Inventing Temperature: Measurement and Scientific Progress by Hasok Chang
Irreversible Damage: The Transgender Craze Seducing Our Daughters by Abigail Shrier
Island by Aldous Huxley
Jamberry by Bruce Degen
Japan at War: An Oral History by Haruko Taya Cook and Theodore Failor Cook
Kora in Hell: Improvisations by William Carlos Williams
Leisure: the Basis of Culture by Josef Pieper
Making Nature: The History of a Scientific Journal by Melinda Baldwin
Making Sense of Tantric Buddhism: History, Semiology, and Transgression in the Indian Traditions by Christian K. Wedemeyer
Memories of My Life by Francis Galton
Mind and Cosmos by Thomas Nagel
More Work for Mother: The Ironies of Household Technology from the Open Hearth to the Microwave by Ruth Schwartz Cowan
MOSCOW-PETUSHKI by Venedikt Yerofeyev
Nobody wants to read your sh*t by Steven Pressfield
Now It Can Be Told: The Story of the Manhattan Project by General Leslie M. Groves
Peak: Secrets from the New Science of Expertise by Anders Ericsson
Pericles by Vincent Aulay
Private Government by Elizabeth Anderson
Public Choice Theory and the Illusion of Grand Strategy: How Generals, Weapons Manufacturers, and Foreign Governments Shape American Foreign Policy by Richard Hanania
Rationality: What It Is, Why It Seems Scarce, Why It Matters by Steven Pinker
Reason and Society in the Middle Ages by Alexander Murray
Robert E. Lee: a life by Allen C. Guelzo
San Fransicko: Why Progressives Ruin Cities by Michael Shellenberger
Slaughterhouse-Five and Breakfast of Champions, by Kurt Vonnegut
Storm of Steel by Ernst Junger
Surface Detail by Iain M. Banks
Sweet Valley Confidential by Francine Pascal
Termination Shock by Neal Stephenson
Troubled Blood by J.K. Rowling
The Age of the Infovore by Tyler Cowen
The Anti-Politics Machine by James Ferguson
The Axis of Madness
The Beginning of Infinity by David Deutsch
The Book of All Hours series - “Vellum” and “Ink” - by Hal Duncan
The Book of Blam by Aleksander Tišma
The Book of Why by Judea Pearl and Dana Mackenzie
The Brothers Karamazov by Fyodor Dostoevsky’
The Castrato by Martha Feldman
The Condition of Postmodernity: An Enquiry into the Origins of Cultural Change
The Dark Forest by Liu Cixin
The Dawn of Everything: A New History of Humanity by David Graeber and David Wengrow
The Dawn of Everything: A New History of Humanity by David Graeber and David Wengrow
The Deficit Myth by Stephanie Kelton
The Diamond Age or A Young Lady’s Illustrated Primer by Neal Stephenson
The Doomsday Machine: Confessions of a Nuclear War Planner by Daniel Ellsberg
The Ecotechnic Future by John Michael Greer
The Eighteenth Brumaire of Louis Napoleon by Karl Marx
The Evolution of Beauty: How Darwin's Forgotten Theory of Mate Choice Shapes the Animal World - And Us by Richard Prum
The Extended Mind by Annie Murphy Paul
The Fall of Robespierre: 24 Hours in Revolutionary Paris by Colin Jones
The Future of Fusion Energy by Jason Parisi and Justin Ball
The Goal / It’s Not Luck by Eliyahu Goldratt
The High Frontier: Human Colonies in Space by Gerard K. O’Neill
The Hundred-Year Marathon: China’s secret strategy to replace America as the global superpower by Michael Pillsbury
The Internationalists by Oona Hathaway and Scott Shapiro
The Irony of American History - by Reinhold Niebuhr
The Knowledge: How to Rebuild Our World from Scratch by Lewis Dartnell
The Man Who Quit Money by Mark Sundeen
The Mathematics of Poker by Bill Chen and Jerrod Ankenman
The Matter With Things by Iain McGilchrist
The Mirror and the Light by Hilary Mantel
The Motivation Hacker by Nick Winter
The Myth of Mental Illness by Thomas Szasz
The Narrow Road to the Deep North by Matsuo Basho
The New Science of Strong Materials by J. E. Gordon
The One World Schoolhouse by Salman Khan
The Origins of The Second World War by A.J.P. Taylor
The Outlier: The Unfinished Presidency of Jimmy Carter by Kai Bird
The Party: The Secret World of China's Communist Rulers by Richard McGregor
The Reckoning by David Halberstam
The Republic by Plato
The Righteous Mind by Jonathan Haidt
The Righteous Mind – by Jonathan Haidt
The Russian Revolution: A New History, by Sean McMeekin
The Society of the Spectacle by Guy Debord
The Tyranny of Metrics by Jerry Z. Muller
The Virus in the Age of Madness by Bernard-Henri Lévy
The Yom Kippur War: The Epic Encounter That Transformed the Middle East by Abraham Rabinovich
Three Years in Tibet by Ekai Kawaguchi
Trans: When Ideology Meets Reality by Helen Joyce
Troubled Blood by J.K. Rowling
Trump: The Art of the Deal by Donald Trump and Tony Schwartz
Unsettled: What Climate Science Tells Us, What It Doesn’t, and Why It Matters by Steven Koonin
Unsettled. What Climate Science Tells Us, What It Doesn’t, And Why It Matters by Steven E. Koonin
Very Important People: Status and Beauty in the Global Party Circuit
Viral by Alina Chan and Matt Ridley
War in Human Civilization by Azar Gat
When men behave badly by David Buss
Whiteshift: Populism, Immigration, and the Future of White Majorities
I read a lot of them, skimmed through more, and skipped a few. I'm now waiting for the final result because there is one question I want to ask one of the reviewers of one of the books, and it's not fair to start a discussion about reviews right now.
But honestly, I was delighted and surprised to see what some of the books reviewed were, and although there were several that didn't interest me at all, that's just my own tastes and I'm very glad there is such a spread of topics covered.
For all the occasional kerfuffles and spats, we definitely are each other's people 😁 📖
I’m planning on reading ‘em all; currently about a third of the way through. So far, about half have been quite good, about 10% have been pretty bad, and the rest just ok
Well, I read & ranked all of 'em. 3/4 were good-to-great, 1/4 were "meh" or worse. Only 11 reviews were actually bad. 13 definitely deserve to be finalists, and another 12 arguably deserve to be finalists. That leaves 74 (about 56%) which I enjoyed reading but don't quite make the final cut. An excellent competition so far!
I’ve got 16 left to go - will probably finish in a day or two. Tons of stuff that’s well worth reading. Less than 10% are actually bad. Many are fascinating glimpses of topics I know little about - I always enjoy reading stuff like that 🙂
Scott, you can consider adding to the book review rating form a free text field for feedback which will be shared with the book review author. Feedback is valuable for improvement! :)
I'm not sure because this was my first encounter with CODEPINK. I guess they accepted certain elements of Russian propaganda for the same reason the Gravel Institute did, because they like anti-US messages which Russia was eager to provide. The rest of CODEPINK's article wasn't obviously wrong to me (for the most part), perhaps only because I wasn't familiar with a lot of the issues it discussed. (Notably the same article talked about China-Taiwan tension, for which they predictably blamed the U.S.)
7-9 for 'really good', 5-7 for 'good but not the greatest' and 2-3 for 'this is poor/terrible/if I knew who you were, I'd be in a slapfight with you right now, reviewer'.
Is being "intellectually angry" a thing? I don't mean being angry about some culture war issue; it's more "someone is wrong on the internet" but with an added feeling of helplessness that I can't possibly correct them.
The issue is AI Alignment Risk and this website. I love this website. Scott is one of my favorite all-time bloggers. But taking AI Alignment Risk so seriously is obviously so fucking crazy that it makes me "intellectually angry".
I've never understood exactly what the Rationalist community is, never really tried that hard. I mostly like them, yet there has always seemed something slightly *off* about them. Off in the way that Mencius Moldbug is brilliant but also clearly... off.
I realized today what the offness is, for me. It's the difference between Platonic vs. Aristotelean thinking. The term Rationalism has bemused me because I have often thought over the years: "Isn't this just Enlightenment thinking and didn't that start in the 18th century?"
But now I see the difference. The Enlightenment focused on empiricism and was most influenced by Aristotle. Rationalists, in their embrace of Bayesian thinking, seemingly feel free to discard empiricism, and this has led them to believe some crazy, rudderless shit. Such as AI Alignment is a reasonable thing to spend tons of time and money and human intelligence on.
To be clear, our gracious and brilliant host is also a brilliant, trenchant empiricist--when he works with empirical data. Unfortunately, he also seems to update--way too much---on non-empirical issues while in the company of persuasive friends. AI Alignment being the main one.
I think we need more debate between those who believe AGI is an X risk vs those who don't. All the headline debate on the issue here now seems to be between those who believe AGI is a huge risk and those who believe it is only a very, very major risk.
It makes me intellectually angry, if that's a thing.
> Rationalists, in their embrace of Bayesian thinking, seemingly feel free to discard empiricism
As an aspiring rationalist, I hereby denounce any such idiocy you may have encountered.
> Such as AI Alignment is a reasonable thing to spend tons of time and money and human intelligence on.
It depends what you mean by "tons". I consider the near-term risk of AGI-induced global catastrophe to be pretty low .... maybe 1% in the next 25 years, or something like that. But that doesn't mean it doesn't deserve billions of dollars of research funding to mitigate. x-risks, Global Catastrophic Risks and s-risks can still have an immense negative expected value even at low probabilities.
OTOH I do kinda think that some other GC risks may be underestimated, e.g. might we be close to reaching Earth's carrying capacity, especially if a big war occurs? I couldn't find any EA research on that. Weird.
I don't see how this is connected to empiricism; I love empiricism, but it doesn't mean that I must assign a probability of zero to any event that has never occurred in human history. To the contrary, history is chock-full of examples of unique events with far-reaching consequences. Doesn't this imply to an empiricist that ze should be concerned with such things? And anyway 0 is not a probability (https://www.lesswrong.com/posts/QGkYCwyC7wTDyt3yT/0-and-1-are-not-probabilities)
Edit: I would, however, add that I think Yudkowsky overstates the case for this particular x-risk, and I'm not sure if it's deliberate exaggeration or if he and I have genuinely different opinions. (Edit 2: actually I'm pretty sure he sees a higher risk than me, but whether it's a small or large difference is unclear. But I think the risk rises a lot *after* the first AGI is invented, i.e. that the first one isn't going to be the most dangerous.) But to throw it back at you, if you estimate that a certain catastrophe X has a 1% chance of happening, and if it happened, would cause damage somewhere between $1 trillion and <human extinction>, how much do you think it would be worth spending on prevention, assuming maybe you could cut the risk in half or so?
I don’t really think you can be intellectually angry, but being angry at other people’s stupidity is a very human thing.. anger is not rational by design.
It’s a feature, not a bug.
on the topic of artificial intelligence catastrophe, here is a quote from an article at Politico:
O’Dowd said he hoped the campaign — starting with the national ads — would clarify for Americans that his goal is to make computers safe for humanity. He said even more than a nuclear attack, he fears that someone will one day click a computer mouse in Moscow or Pyongyang and “half the country is going to go down.”
“This country could be put back into the 1820s by someone coming in and getting control of our software,” he said.”
This man is intent on destroying Tesla at any cost for reasons that are rather obscure to me. Except this quote kind of gave me a peek into it.
What is he afraid of? Is it the computers, or is it the people behind them?
I think this really ties in to some of the intense feelings about AI and it’s threat to us.
My personal opinion is what we are really afraid of is ourselves.. which, given our box score, is not unreasonable.
Some of us are casting ourselves as the good parents who inadvertently raised a psychopath and couldn’t stop wondering how their child got that way.
I do not understand what would drive an artificial intelligence to want to destroy us unless we somehow convinced it it was a really really good idea. You would have to convince me that acting from a deep place of power and desire was somehow a function of intelligence.
The other thing I always wonder is what kind of a body are we going to put this intelligence into to live? Because that will make a big difference in its disposition.
What if we came up with an AI that could do all the wonderful things we imagine a Superintelligence could do. Take us to other galaxies, find ways of making infinite amounts of food, figuring out how we can run all our gadgets with no wires or nothing. It could do all this cool stuff, but instead of doing those things from a place of wanting to take over the universe it was doing it just for fun.
That way they would keep us around as pets and probably get a kick out of us.
I've always felt this way about the obsession with AI risk. My feeling connects to a few things I can identify:
- eschatological religion: the obsession with some impending end of the world that is always nigh, and how the emotional posture of those in the church is a kind of vast smugness that says "listen to us or meet your deserved end". We all have a natural aversion to this kind of thing: we can sense the smugness, the immaturity, the sanctimony, the shape of the ego that would let itself fully subscribe to an idea and turn around and expect others to see that they're right and come to them, and just how much it feels like a posture that serves the ego rather than a legitimate spiritual belief.
- climate-change anxiety; the overcharged opposition between those who "believe science" and deniers. For one, climate-change doomsaying often comes off feeling eschatological. And the deniers end up denying far too much, and regarding AI risk it feels like being skeptical would get you grouped with them, which feels unfair because of the very different standards of proof, so we resent this. Then the climate-change doomsaying takes a form of continual atonement: microscopic acts to address guilt (plastic straws) rather than any true sacrifice (a career in battery technology), is far too focused on what WE can do rather than the actual political problems of creating international agreements and enforcingthem, and is all out of proportion (the realistic scenarios expose us to a level of "life affected by environmental forces" that would still be a luxury for people any time in history before a century ago, nothing like a true end of the world)
- and, I did a couple of years of physics grad school before becoming VERY disenchanted with the cutting edge of physics. I recall that the core emotional arc of this, for me, was not a genuine interest in the subject (although I did have one), but a fear of engaging with the messy real world. Instead, I preferred the orderly natural world; it has the feeling of studying the deep magic of reality. At my core I had a fantasy of retreating into obscurity for decades to study arcane magic, eventually to re-emerge with some world-changing accomplishment (like relativity) and receive my reward in adulation and Nobel prizes. Pretty selfish. Very suspicious now that many people get into arcane fields for this reason, and suspicious of any enterprise that never gets close enough to reality. (See also: tech projects that take too long to get in front of users)
For all these reasons I cannot see AI risk talk without feeling like the interest is REALLY an emotional one based in the proponent's need to be on the right side of something that feels huge and important, more than rational calculation (while being heavily dressed up in rational aesthetics to support that core emotional need)
Edit: but, I'll grant, my opposition to AI risk is JUST as emotional: a wariness that the AI-risk-doomsayer has given in to their ego's desire to be important. Climate deniers must start from this too. A properly rational calculation is called for, I'll grant, to get beyond this. But if you rationally calculate that AI risk is a big deal, and you want to doomsay, you'd be well-served to steer far clear of any language that keys into this particular emotional script, and to be wary of giving into it yourself.
I think the only people who think there is literally zero risk from AGI are those who believe AGI is impossible or at least that it will take hundreds to thousands of years to achieve (so zero risk right now).
That leaves open exactly how concerned we should be right now of course, but presumably if we're not entirely sure that any possible AGI we might create will be benign the precautionary principle would apply.
I mean when you limit it to saying 0 risk, of course you won't get any takers here. I'm willing to say it's an extremely low risk, even though I think human level AI is possible.
One of the (many) points of disagreement I see is people conflating reaching human-level AGI with "takeoff" scenarios, which rely on the agent having god-like intelligence and being extraordinarily robust. Neither of those is something that would magically appear upon reaching the human-level threshold.
I actually have a pretty high credence that we'll see human level AI within a few decades, but this doesn't translate to huge X-risk in my view of the world.
Hmm... I don't see takeoff scenarios as all that implausible. I don't think takeoff necessarily depends on the agent having god-like intelligence. Like if you create one human level ai then you effectively have an arbitrary number of them up to the limit of your computing resources. You may be able to run the ai at high speed. Or train the human level ai to become as good as the top human ai researchers.
It is unlikely to take off in quite the way some may imagine as there will certainly be bottlenecks. Physical research and engineering for example may not be sped up much if at all at this stage.
"Human level" is also doing a fair bit of work. It's not clear to me that human level intelligence is a natural stopping point for AI. If we look at narrow AIs they are often either clearly subhuman or superhuman within their narrow domains.
Yeah basically I think the bottlenecks are much more severe than people imagine. I am sympathetic to the "if we had 10,000 Alec Radfords AI progress would go much faster" point of view, but I think it's missing the degree to which even the most successful AI researchers rely on empirical results. You have to wait long amounts of wall-clock time for experiments to run, even if you're one of the best AI engineers.
I agree that human level isn't some magic point that it's not possible to surpass. I guess I think the time it takes us to get from human level to 120% human level will not be significantly shorter than it takes to get from 80% to 100%.
"I think the time it takes us to get from human level to 120% human level will not be significantly shorter than it takes to get from 80% to 100%"
How long did it take to go from Go playing AI that could barely compete in the children's leagues to one that soundly trounced the best human grandmasters?
The human mind doesn't seem to be anywhere near the peak of potential cognition.
Okay for what it's worth, the time it took to go from 80% of Lee Sedol's level (the matches vs Fan Hui let's say) to beating Lee Sedol, and then the time to improve more beyond that were similar. They were pretty similar, but it doesn't support my overall point since they were both short yes.
>I've never understood exactly what the Rationalist community is, never really tried that hard. I mostly like them, yet there has always seemed something slightly off about them. Off in the way that Mencius Moldbug is brilliant but also clearly... off.
>I realized today what the offness is, for me. It's the difference between Platonic vs. Aristotelean thinking. The term Rationalism has bemused me because I have often thought over the years: "Isn't this just Enlightenment thinking and didn't that start in the 18th century?"
>But now I see the difference. The Enlightenment focused on empiricism and was most influenced by Aristotle. Rationalists, in their embrace of Bayesian thinking, seemingly feel free to discard empiricism, and this has led them to believe some crazy, rudderless shit. Such as AI Alignment is a reasonable thing to spend tons of time and money and human intelligence on.
Deciding the issue with Rationalism is a lack of empiricism despite not really trying to understand the community is not a new critique - I'm having flashbacks to Why I am Not Rene Descartes. You're not making *quite* the same arguments, but it certainly rhymes.
On the object level, there's a clear tension behind the idea that empiricism is a virtuous path to truth, AI Alignment is a non-empirical issue, and other people's ideas can be "obviously so fucking crazy" despite that lack of clarity. This smells like an issue of overconfidence, but I'm not sure the problem is where you think it is.
Giving lip service to empiricism is as common as crabgrass. What huckster politician *doesn't* claim his nostrums are rooted in objective data? You need a lot more than a stated allegiance to measurement to be credited as a genuine empiricist.
Descartes and Leibniz really do claim they're not using empirical data, but only the objective good reason that God gave them! (And fortunately, God was good enough to pre-establish a harmony between what goes on inside and what goes on outside.)
Meanwhile, the founding document of rationalism (Yudkowsky), the Twelve Virtues of Rationality, states:
The sixth virtue is empiricism. The roots of knowledge are in observation and its fruit is prediction. What tree grows without roots?…Do not ask which beliefs to profess, but which experiences to anticipate.
It matters how close the alignment group think machine learning is to producing AGI. Because I don't think it ever will on the current trajectory. You can't fake cognition with decision trees and mathematical telemetry via language models. And consciousness is poorly described as Markovian, imo.
In the coming years, I predict the rise of neural-symbolic learning, incorporating some old rule-based methods into the current sub-symbolic approach.
I also get kind of annoyed with the whole AI alignment discussion. It bothers me for a couple of reasons. Firstly, I've not yet been disabused of the impression that the rat community by and large believes similar things about AI safety because that's one of their cultural beliefs rather than rationally arived to with evidence, as much as they protest this. I do see them produce evidence for their claims, but it appears more like christians producing evidence of christianity rather than someone producing evidence of something that's actually true. I don't want to sound overly harsh here, there are a lot of interesting arguments being made and all that, but I do think the community fell prey to collectivism much more than it wanted to in a lot of ways.
The second thing that bothers me is my own inability to articulate the reasons why I think what I'd call the Rationalist AI Alignment Risk thesis is wrong. The best I can try to give is the short version, which sounds something like "Optimization does not work that way." but I've never been able to articulate a deep explanation for this, even though I know I have one. And it bothers me that I have such a hard time with it.
I think the reason the rationalist community has so many people believing the Rationalist AI Alignment Risk thesis is simply because it has so many people who have been exposed to the thesis, and some fraction of them have become convinced. I don't think this is an argument for or against.
I do see this exact discussion semi-regularly, so I'm not even convinced that even a majority of the community is worried about AI risk. There's some people that are seriously worried, some people in the "maybe 1% likely but think of the expected value" camp, and some straightforward skeptics. There's probably a poll somewhere of how many people are in each group.
But since the general population has ended up in the third group due to the argument "Wait, isn't the Terminator fictional?" then of course the rationalist community seems unusually credulous here.
For what it's worth, I am convinced AI Alignment is a real problem. But I think you've sort of pin-pointed the reason why its most vocal proponents are going about it the wrong way.
People like Yudkovsky seem to believe superintelligent AI will just arise spontaneously and bootstrap itself from nothing to world domination in a matter of hours, and one thing it demonstrates is a complete disregard for the sheer amount of empirical knowledge achieving world domination would require.
Of course, ultimately, humans will be slowly training AI with exactly this kind of knowledge, in pursuit of personal convenience or some marginal advantages in zero-sum capitalist competition, so at some point we'll reach the point where the risk becomes real. But the current AI Risk research disregards this area, assumedly because it just doesn't think it could possibly be a problem that a sufficiently intelligent actor would have trouble overcoming. This is doubly harmful, as it both sets them out to become "boys who cry wolf" who see AI dangers where they don't yet exist, and distracts them from pursuing and advocating some really simple risk-mitigation strategies that would probably be if not sufficient period, then at least sufficient for a while after a demonstrably superhuman AGI actually comes into existence.
Can you elaborate on why you don't expect early early AGI to have large amounts of empirical knowledge?
The most general and impressive AI's we currently have are trained by a process which you could reasonably describe as "distilling half the internet into a probability distribution". These language models "know" more things than any human, by a long way.
Is it that you don't expect early AGI to be a successor of these language models, or that you set the "amount of empirical knowledge necessary to take over the world" bar higher than the amount of empirical knowledge currently stored on the internet?
Empirical knowledge is not fungible; being a world-class expert in auto repair is of almost no value if you're tasked with removing an appendix or winning an MMA competition. You need empirical knowledge in the specific field.
And there's a distinct shortage of empirical knowledge in the field of World Conquest. Most of what we do have in that area is predicated on e.g. having command of an army at the outset. Note that armies are generally owned by people who are skillfully paranoid about wrongdoers subverting those armies, and there's also a shortage of empirical knowledge on how to subvert armies.
Also, most of the empirical knowledge that exists, is *tacit* knowledge. It's not written down or digitized *anywhere*, it can't be flawlessly derived a priori, you've either got to get a meatsack to teach you, or learn it the hard way. And the meatsack will probably get suspicious after too many long conversations on the fine details of world domination.
The first AI that wants to Take Over The World as a prelude to paperclip-maximization, is going to have to do a whole lot of trial and error in figuring out how to actually do that. It's going to make mistakes. And there are enough opportunities to make mistakes that it's likely going to stumble on to a fatal one while it's still small enough to be killed by an angry sysadmin with a crowbar.
> there's a distinct shortage of empirical knowledge in the field of World Conquest.
I don't agree. I watched a show about dictator's playbooks awhile back, and another show about genocides, so it's certainly a field that has been studied, and genocides & coups have kept happening despite the often severe epistemological failings of the people who cause them. Indeed, many dictators succeed on their first or second coup attempt! So is it really that hard? I don't think it's that hard, but I think the necessary combination of psychopathy and ambition is rare.
OTOH, dictators also historically seemed to require luck, to be in the right place at the right time. A superhuman AGI, however, could use empirical data to "make its own luck", e.g. it could observe the background characteristics of everyone who has ever done a coup, then work out how to create the necessary conditions.
An AGI will, however, have a major disadvantage by not having a human body, which means that all strategies that depend on being human won't work. But I don't know how to rule out the possibility that there is a realistic way to get around this limitation.
Edit: come to think of it, a key characteristics of dictators and "genocide-ists" is their ability to inspire and control others, and to delegate responsibilities. Anything you can delegate to others, you don't have to do yourself. And in principle all this inspiration and delegation could be done online. A superhuman AGI on the internet (which, of course, could be psychopathic simply by lacking a reliable system to prevent this) could first convince people it's human, perhaps not just via text messages but also via a photorealistic AI-generated human persona. It could, in principle, find a way to control a supply of money, with or without having direct control over a bank account... all via delegating certain meatspace activities to humans. So I don't see why it couldn't, in principle, follow a dictator's playbook once it gathers control/influence over enough resources.
Agree mostly, I'll give a try for part of my 'optimization doesn't work like that' explanation.
One thing that I think people who believe that there will be rapid takeoff are missing is that they seem to have this sense that all the AI has to do is overcome the humans and then it's off to the races to conquer the universe. I think there are some pretty fundamental mathematical and physical reasons that things don't work like that. Parts of it, things that point to it go by lots of names, P vs NP, undecideability, Wolfram's Principle of Computational Equivalence, Goedel's Incompleteness Theorem, Chaos theory, 2nd law of thermodynamics....
All of these things circle around one central truth, a truth that says something like "actually getting things right, or making something new that works, is a fundamentally hard problem" and I think that no matter how powerful you are cognitively, it's always hard problems all the way up. And you can't solve tomorrow's problems with the knowledge of today, you have to do it the hard way.
True intelligence, I think is, more or less, the ability to 'do it the hard way'. And part of the requirement to 'do it the hard way', to understand how to make something new, it also requires the general understanding of the type that lets you understand what someone 'means' by what they say, rather than just what they say.
what we call 'alignment' (at least in terms of not being a paperclip maximizer), and what we call 'general intelligence' are two sides of the same coin.
I've been interpreting the rating as "how happy would I be to see this as an ACX guest post and have attempted reading it", which both means rating ragequits poorly and docking points for things like "just didn't find the subject material that interesting".
I'm reading a bunch of the reviews, and it strikes me how many people have "the makings" to be pretty good writers were they to spend a bit more time doing it.
There's a period of time you go through when you start writing regularly with the intent of publishing where you get rapidly better, and I keep seeing things where it's like "this guy is already pretty good, I wish he'd write ten articles in three months and be great".
I'm about as sure as you can be about a complex human thing. I'm sure there's exceptions, but when you start writing stuff that people are going to see, it's going to (on average) spur people to a closer level of scrutiny of their own stuff. I've seen multiple people with good educations going through the process, and the pressure seems to help them refine in ways school-writing didn't.
Best example of this I know of right now is Parrhesia's blog (https://parrhesia.substack.com/) which has just been getting better and better every post. But you hear similar things in a lot of fields - some singers will tell you that without an expectation of performance most people top out, etc.
Another thing is that for most people who are starting to write "in front of people" for the first non-academic time, they are trying to figure out what they want to be in terms of voice and what kind of things they want to talk about. They are figuring out stuff like "what's my focus". And these are all things you can improve on, or at least constraints you can optimize for/within.
Anyway, just one man's opinion but I think it's broadly true.
ah, thanks! That touches upon a number of issues I had been thinking about lately.
My main doubt was or partly is that you don't get direct feedback - or if so, probably more on specific bits of content, than on issues of style and such. But practice and reflection can go some way I guess.
So some of the best feedback you get, or at least the most believable, is when people begin to disassociate you from "being a real person". So occasionally I'll be on a forum or something where they are discussing an article and someone will say they liked X, or hated Y, and it's nice because you know to them you aren't a human being that exists and they are really talking about the writing.
But really the best feedback you are going to get is basically just getting closer and closer to the kind of stuff you want to write. I once read a thing about dealing with hecklers where a stand-up comedian was saying something like "Listen, you believe you are the funny one in the room, that you are funny enough to do this. That should lead you to believe you can take down some drunk rando with your words."
I think that's broadly true. There are some ways that feedback helps, and there's certainly some I listen to and take, but for the most part I think you are looking for forces that make you look at your own work closer, that make you think really hard about how to create the best words you can for people. It's making you give yourself better feedback, basically.
Thanks a lot for the kind answer. I see your point. It also made me wonder, if I love writing so much, that I want to engage in those thought processes so deeply. At that moment, that feels like work - though I know from other tasks that it can come quite automatically and actually be enjoyable.
I get your point about other people speaking about you somewhere else. At the same time, uhm, I don't know how long it would take until I would read people talking about an article of mine somewhere else. I think I'd be happy if they found my article in the first place.
I had been thinking about writing lately. But then, as you wrote, I'm very aware that my goals change and are not fully clear, and that I haven't really figured out yet, what 'kind of stuff' I want to write. Or rather, if the mixture of texts that come to my mind would make any sense to others.
Thanks for the opportunity to reflect on this some more!
Before internet, publishers and readers were removing the not-so-great writing from circulation. Publishers by not publishing the text (perhaps unless some changes were made), readers by not buying it.
Also, authors couldn't hyperlink their previous articles, which made their individual pieces more self-contained. Which means that if you kept their first and third works, but removed the second one, the result still made sense.
Haha, I usually write long comments, and this time I also first did the same, and then I thought "uhm, why not delete the obvious parts?" and tried that approach. Oops.
The idea is that in the pre-internet world, the guy who writes a lot and is already pretty good... would still keep writing a lot, but only 1% of that would be published and remembered... and if you afterwards read only that 1%, you would conclude that he was great. So your wish would kinda be granted, but in a very unintuitive way.
This could still happen today, if e.g. some obsessed stalker collected all writings, and then organized some rating by audience, and selected the best 1% of them. Unless perhaps the internet form of the texts (the fact that author can expect that his audience is familiar with his previous posts, or at least can reference those posts) would make that 1% selection difficult to read, because it would keep referring to things outside the selection.
I've seen Jim Kennedy around thorium reactor groups a lot, and now he's giving his origin story:
> So I met with the Pentagon guy and I laid out this plan, and I said "well here's what [China's] doing and here's how we can counter it, and if we we counter it like this, they won't be able to offset our actions, and we'll be successful at building, reestablishing a value chain," and the guy says to me "wow this is this is really interesting, you put a lot of thought into this... this is really good," and I said "yeah yeah, thanks, you know, I appreciate it, I'm sure that you've got, you're looking at other things, right?"
> He looks at me, like, "what do you mean?" I said "well, I mean, I just kind of threw this together and, you know, I'm just a private sector guy, and this is the Pentagon and I'm sure you guys have been looking at this and you have like, a real plan, right?"
> He goes "I don't, I don't understand what you're talking about." I said, "this is a national security issue, so I am under the assumption that the Pentagon is on top of this, and there's lots of other good plans, and I'm not the only private sector guy with the solution." And he goes he just he's looked at me like (shrugs) "well no, that's, that's it." And I said "what do you mean?"
> I said, "this is national security. You know, you guys should be developing a plan, it's not my responsibility." I said "what if i didn't show up?" and I swear to god, this is what he says, he goes "well you're here aren't you?"
Evidently this is a guy who became interested in thorium molten salt reactors not to solve global warming and air pollution like many of us, but because the U.S. is letting China control the global supply of rare earth materials that are critical for manufacturing various high-tech goods (notably motors and magnets). Mostly this is a result of laws around thorium. Heavy rare earths are always found together with thorium geologically, and U.S. law says that a company cannot dig up rare earths, extract those rare earth earths and bury the rest. Why? Because the residual dirt is considered "nuclear waste", or in technical terms "source material".
This is the main reason China controls 90% of the rare earth market. And this is, of course, a national security issue since it means China has huge leverage over *everyone* else in case there is any conflict between China and anyone else. We could simply change the law, of course, but I guess we won't because politics. So Kennedy's solution involves some kind of thorium trust. Rare earth miners will extract the thorium and deliver it to a group in charge of storing it, and this group will in turn sell it to people making reactors that use thorium, such as Thorcon. But, grain of salt, I have a sense that I don't quite understand what he's saying about the problem or solution.
Edit: I'm disavowing Kennedy's comments against NATO, though. Because https://twitter.com/jessicabasic2/status/1513836355440111621, plus he asserts "the Russians" are "calculating rational people" and it's become very clear that Putin is calculating, but not so much rational. But all that other political stuff isn't what he usually talks about and isn't why I listen to him.
I didn't know about this law: I had heard it was hard to mine rare earths because of "environmental regulations" but I assumed that was your standard "Don't let your open pit mine leach toxic metals into the water supply" kind of regulation. I had heard of a rare earth metal mining and processing project starting up in the Alaska panhandle recently, I wonder what they're doing with the thorium.
As an aside, it bothers me when people say Putin is not rational. I think he's evil, but I don't think he's irrational. It strikes me that some people have bought to much into Yudkowski's idea that rationality=winning (or at least the popular misunderstanding of it: it seems to me all he was saying is that if the "rational" route consistently leads you to losing, then it's not really rational). The idea seems to be that if someone does something that turns out badly, they're irrational. But that's not irrationality: believing that if A equals C and C equals B then A does not necessarily equal B is irrational. There are plenty of rational people who act the fool out there, and I can't even say that Putin was acting that foolishly. Almost everyone predicted an easy win over Ukraine, and as Putin predicted it didn't start a war with NATO. Sometimes you make a gamble and it turns out the odds weren't what you thought they were, that doesn't mean you're irrational. Was Nate Silver irrational for giving Trump a 28% of winning in 2016? No, he was just wrong (or not even that: after all, 28% chances happen 28% of the time).
Putin going to war with Ukraine was a mistake, and evil, but I don't see how it was irrational unless you equate rationality with never making mistakes and not being evil.
Can you tell me of any specific military analyst who considered the obvious factors (UKR military capabilities, RU military capabilities, UKR public sentiment, geographic/physical strategic aspects) and concluded that Putin was likely to be able to complete his objectives (stable overthrow of Kyiv and at least a couple of other major cities) with 200,000 troops?
I expect such analysts exist, but I don't remember seeing anyone in particular. However, I don't think that such analysts expected the lousy strategy and tactics that Russia actually used. This video explains: https://www.youtube.com/watch?t=1474&v=zXEvbVoDiU0
So, what I mean is that Putin was irrational in the usual human sense of confirmation bias & positivity bias. He let himself believe that his operation would succeed within a few days, as his yes-men told him (because Putin probably demanded it from them), and he refused to believe in the reality of Ukrainians' feelings toward Russia, nor did he prepare for the possibility that Ukraine was prepared, which of course it was. Vlad Vexler further asserts that Putin really believed Ukrainians would greet Russians as liberators, which I find credible. And to some extent, it seems like his delusion has continued for many weeks, as he's still using the "special military operation" moniker instead of declaring "war" which, apparently, is legally required to mobilize reservists/conscripts. Thus, his forces will probably remain undersized, and certainly underpowered, for months to come.
Major sources contributing to my understanding include the following sources published before the war.
- Belated edit: I also saw a presentation somewhere...can't remember where... saying that even if Putin took Kyiv and other major cities, Russia would pay a terrible price that would ultimately weaken it.
The thing he would realize, if he were rational, is that (1) he can choose his own goals rather than following the same path he's used in the past, and (2) he was wrong about a bunch of things and ought to have re-evaluated, either by aborting the war before February, or scaling back now (because his encirclement plan is risky and likely to fail), or at least delaying the coming offensive (because it's a complex op that needs a lot of planning).
Previously he's had remarkable success boosting his popularity with military adventures and killing people (which reminds me, have a look at the Apartment Bombings summary here: https://twitter.com/DPiepgrass/status/1507210690427174916), and he's made expanding the Russian empire militarily his goal, but he could just as easily have chose "expanding Russian influence" as his goal. And trade relations / foreign policy is a better way to do that. China's Belt and Road initiative is the obvious example, though I think that initiative is undermined by Xi making himself dictator for life, while making China into the world's biggest military power, killing off Hong Kong's freedom, and threatening Taiwan, the U.S. and Japan - taken together, this is terrifying, and if nearby nations have any sense, they would compete with China in manufacturing so as not to be so dependent on it. Putin could have improved Russia's domestic manufacturing industries and implemented reforms to reduce kleptocracy (because it's not an efficient system). Putin could have even chosen to join NATO (though that would be difficult now because of his previous annexations).
While his war is still likely to acquire some territory for Russia (probably a temporary win), it will dramatically reduce Russian influence. So I would say that if he wants a "great Russian empire", his recent and current actions are directly opposed to that goal. Edit: also note that Putin has given Xi/China a lot more power over Russia, as Russia now depends heavily on China, but China doesn't depend much on Russia.
Now, there's another strategy Putin could have taken: instead of trying to improve Russia's position, drag down the West. Have a look at what Stoic Finance said just before the war on Feb 18. He assumed that Putin was rational, so his interpretation of the military buildup was that Putin was trying to "try to destabilize the west" as he had done in the past: https://www.youtube.com/watch?v=d4SENp1IT6o
And indeed, U.S. had been warning loudly that Russia planned to invade, so NOT invading would be a win for Russia because it would make the U.S. look like the boy who called wolf.
It still seems to me that you are equating "rational" with "wise" and "irrational" with "foolish." There are many rational people who are fools, and Putin is one of them. But I think we're just going to disagree on what it means to be irrational.
Putin is intelligent, calculating and foolish. What role is "rationality" playing here? What would Putin have to do differently for you to call him "irrational and calculating"?
Lets refine all the thorium and turn them into commemorative coins! They're collectable, and come with their own lead lined carrying case. Very stylish.
Are they good enough to be worth deploying against a great power opponent? A tank that's outclassed badly enough by enemy equipment is just an expensive, cumbersome way of getting your tank crews killed.
I don't know enough about T-55s and T-62s to evaluate with confidence if they fall into this category, but a few things incline me towards suspecting this is the case. For one thing, those tank designs are over 60 years old and 20-40 years older than modern designs currently fielded by great powers, and they appear to be 1-2 major generations of capability behind modern main battle tanks. For another, both Russia and Ukraine are former operators of T-55s and T-62s who have many years since scrapped or sold them off to third-world countries: if Ukraine thought those tanks were still any good, I would have expected them to have been kept around for reserve units, or at least mothballed for restoration to service in the event of critical need. Even Russia, which I thought was a notorious pack-rat of old military equipment, seems to have scrapped their mothballed T-62s.
The book-review form has a space for entering your email addresses "to prevent spam and accidental double votes"; however, entering something in that space has not been made mandatory, which I think means that several of my votes have become accidental _non_-votes because I am a moron and often forget to do things.
Scott, if you're reading this and it's easy to do (which I _think_ it is), if you are going to ignore submissions in which that space is left blank could you please make it impossible to submit the form with that space left blank? If you're concerned about spammers and somehow making the field mandatory will make their robots fill things in there, you could have another mandatory field labelled "please enter four plus three" or something.
I read this comment before starting reading the reviews, promptly forgot it, and did the same thing as gjm. I second the plea for making the field mandatory or something.
I suppose that it should be safe just to go through all the reviews I read, remember what I thought of them, and resubmit. But somehow this is the sort of thing that is extremely not-motivating to my brain so I can't guarantee that I will.
I recommend the tables in the sections "Millon's description" and "versus normal personality"
I'm not familiar with BPD specifically, but in my experience the descriptions of personality disorders will ring true if you have them, and having a clear understanding goes a long way towards relief.
The recommended, evidence-based treatment is a semi-structured approach called DBT, Dialectal Behavioral Therapy. There are lots of books about it and therapists who offer the approach via individual or group treatment. My impression is that if the treatment is not particularly helpful for someone who has the full-blown syndrome (severe emotional dysregulation, chaos in their relationships, habit of offloading distress onto others via manipulative suicide gestures and threats). However, most people seem to think otherwise.
If what you are seeing in the person is far more subtle than what I'm describing, I don't think it's really that helpful to think of them as "having BPD," because BPD isn't an illness in the same sense as pneumonia is -- there's not a lot of mileage to be gained from saying "Aha, it's that!" For people who have more low-key versions of the BPD syndrome, a combo of meds and psychotherapy is the best approach.
Alas, my shameful pride. Reading book reviews other than mine, I find myself envious of the good and scornful of the bad. I cannot exorcise the implicit comparison and enjoy them in their own rights! I am tainted and untrustworthy as a reader and evaluator, and as such, cannot bring myself to submit ratings of any of the reviews I have read.
"They are back! Neil deGrasse Tyson is once again spouting total crap about the history of mathematics and has managed to stir the HISTSCI_HULK back into butt kicking action. The offending object that provoked the HISTSCI_HULK’s ire is a Star Talk video on YouTube entitled Neil deGrasse Tyson Explains Zero. The HISTSCI_HULK thinks that the title should read Neil deGrasse Tyson is a Zero!"
Always fun to read someone letting it rip that way!
The best & funniest negative review I've ever read was an essay by Alexander Pope called, I think, "Peri Bathos," panning and parodying all the hack poets of his day. I recommend it highly.
I used to be sort of like that too. Much less burdened by it now. Lots of us who are smart get sort of fixated on the fact of our smartness, and have a terrible time letting go of the hope of being a certified genius. You have to sort of mourn that loss. It helps to realize that being a certified genius actually doesn't make people feel happier or more solid. I knew someone who was chess champion for his state when he was in middle school, and then right after graduating from college won first the US correspondence chess championship then the world championship. He was delighted when he won, but not any happier than I would be if I lucked into a great deal on the car of my dreams. The real locus of happiness is in the doing of something you're deeply interested in and good at.
I mean, couldn't you just set your book review at some value (7 perhaps? You probably think it's reasonably good since you submitted it) and then rate reviews based on how net envious/scornful you feel about them?
Seems like that system would have a decent signal/noise ratio.
Only technically true, as the book did not contain any information to begin with. All language is insufficient to truly communicate, it's merely a wretched cipher used by small minds. True communication is almost an impossibility for us.
If you want to communicate an idea to someone, do it through art. Novels are the most inferior form of this (the actual text does not contain communication, merely its psychic structure), followed in ascending order of "something approaching actual communication" by cooking, architecture (although the art in that has been destroyed by Mara in most of the world), dance, painting and sculpture, interactive and immersive art, and music being the highest and most genuine form.
Well, no, obviously taken at face value that is false, because "information" includes things like "how many times was the letter E used in each sentence, on average?" So "information" strictly defined clearly vanishes when you condense, in the same sense as information vanishes when you use the JPEG algorithm to compress an image.
What you mean to say I think is "useful" or "valuable" information doesn't vanish, which is also a truism, since as long as the definition of "useful" is set appropriately, we can justify any condensation, great or small.
If what you're saying is "most popular nonfiction is a form of intellectual masturbation, where people wallow in clever expositions of themes and ideas with which they already agree, so they can nod along enthusiastically" ("OMG! Look how forcefully he puts in on page 163! I'll have to Tweet that out right away...") -- yes, well, this is the human species. Probably 98% of our communication bandwidth is taken up by group identity signaling.
It depends! If you're talking about coffee table, pop culture of science/history/so forth books, then yes.
If you mean something with real information, then I don't think so. Sometimes you have to lay the foundation for what you are going to discuss before presenting useful ideas, otherwise it will be "as you all know" and no, we don't all know.
Do you think you could express 300 pages of Scott Alexander's most popular posts in 3 pages? Or try expressing any one of the six "sub-books" of Rationality: A-Z in 3 pages.
The 3rd edition is not a text book. It's 'the bible' of analog electronics (at the time of printing.) https://artofelectronics.net/ The first and second editions are more text books. There are cosmology text books and then there is "The Principles of Physical Cosmology" by Pebbles, the classic reference work... I guess you call that a text book. IDK. What makes something a text book?
The narrative can often make the subject more engrossing and easier to remember. By creating an emotional response in the reader, the author highness the understanding of not just the facts but also the context and meaning of the subject.
Though I agree with the general premise that most popular non-fiction could be shortened considerably. However, short books can't be published - no one will be $15.99 for a 20 page paperback. Consider using a service like blinkist which does the summaries for you.
This is true for many nonfiction books, but certainly not all. History books are a classic example. Try condensing the entirety of "the Making of the Atomic Bomb" or
"Imperial China 900-1800" into 3 pages without loss of information. Now you might find information beyond 3 pages boring but that's not the same no extra information.
Most academic nonfiction falls into 2 categories. Either it could have been a single paper instead of a book or it feels like every chapter is a separate paper stapled together under a connecting theme. The only book I've written definitely fell into the latter category. You couldn't summarize it in three pages, but that's not always a good thing.
"it feels like every chapter is a separate paper stapled together under a connecting theme"
Well, sometimes they are literally that. It's not uncommon to find "an earlier version of Chapter 5 was published in the Journal of Such-and-such" in the copyright page.
Joke answer: If true, this would violate the Shannon Source Coding Theorem; the Shannon Source Coding Theorem is mathematically provable; BWOC, this is false.
The answer is, of course, that Cajou's use of "information" is technically incorrect. I am speculating, but Cajou probably meant something like "feelings of insight, changes of perspective, or facts I didn't know," which lays bare the subjective nature of the claim. Joke response: What if the book is written in qubits and the number of qubits adds up to two to the power of the number of characters that fit on 3 PDF pages?
Sometimes you read a book to gather facts and data. That information takes more than 3 pages to convey. Plus there is nuance, and stories involved in history. Takes more than 3 pages to summarize the Rise and Fall of The Third Reich. The whole point is to convey the weird variety of political maneuverings.
Oh that's a great book! Very briefly, it is "everything has been going to Hell in a handbasket and here's why", but he covers so much ground and (at least for me) gave so much new information that it's well worth reading. It's also very entertaining with the whole "damn kids these days" vibe 😀
Idle thoughts: Kant proposed that consciousness/sentience/free will is the result of the rational and physical (/animal) being combined. Chomsky proposed* that consciousness is contingent on language. Take a blended view and look at the latest work coming from AI and could it be that consciousness is the combination of physical/animal (blind sensory input in real time with no specific "training" set) and emergent language processing? It removes the rationality aspect from Kant which was always a challenge (making everything constantly consistent) and allows a potential gateway to understanding future interactions with LLMs.
I wonder, what does it "feel" like for a model to be trained vs called**? How much does real-time sensory input impact the nature of sentience? We know drugs are a problem for humans, could an AI fall into a mode of simply feeding itself fake data to "succeed" in its training?
For some reason I tend to imagine e.g. GPT-3 as being akin to a writer in a pure state of flow: divorced from worldly concerns and purely focused on following the train of thought where it leads.
* I have read Kant, but am relying on a single interview of Chomsky's I've listened to. I may be getting him completely wrong.
** I'm not really up to speed on the technical side of the current models, but the basic data science stuff I did with NNs didn't feed back the "real" output data back into training models, so there was no feedback loop to "learn" from calls to predict, unlike in training procedures (I imagine there must be *some* way of doing this in current models in order to keep stories consistent etc.).
I noticed a curious thing recently. Someone asked me a question, but I was mentally busy (reading something?), and then I said "yes", and THEN the answer and most of the question registered in my conscious mind. I then evaluated the question and answer, which turned out to be correct, but sometimes when this happens, the answer is wrong and I have to correct it.
I believe I am describing what some people call "autopilot", a phenomenon that seems to demonstrate that there are various mental subsystems (including the linguistic subsystem) which are physically separate from the conscious mind.
Yudkowsky seems to think that self-reflection is what makes someone conscious. I disagree with both Yudkowsky and Kant/Chomsky as you described; I think that consciousness is physically separate from those things, and that qualia is the meat-and-potatoes of consciousness. If you're being tortured, you are very much conscious, but your language and self-reflection abilities are not an important part of that experience. I further expect that all current AI models are not conscious, but we need a better theory of consciousness to be sure. We can observe, however, that reflexive behavior seems separate from consciousness and in most cases is opaque to consciousness, e.g. we cannot feel nor introspect the inner workings of our spinal cord, or whatever generates dreams, or our linguistic system. Instead, consciousness feels the outputs of such systems, and then, interestingly, can send out signals about what it feels (which in this case takes the form of comments on the internet).
"I wonder, what does it "feel" like for a model to be trained vs called?"
First of all, what is feeling? Its only identifiable from a distance, where multiple systems of reward+feedback, with different time horizons, overlap. Under a microscope, feeling reduces to sensation reduces to keeping track of cause and effect.
Calling is cheap, training is expensive. Pure calling obviously feels like nothing, because theres no change to the system. It might feel like a slight drain if power consumption is fed back in as an input.
In people, learning happens in multiple stages. First theres the subconscious, all-consuming 0 to 1, where a rough first draft of the action is pieced together. The "feeling" here is mostly black-box, hidden and used as subconscious feedback. It only bubbles up to conscious awareness when theres an existential threat to the activity.
Then comes the conscious competence stage, where the action is saved as a discrete skill, but still takes a disproportionate amount of attention+concentration. There's more mindspace for the conscious self to operate again, and it starts dealing with the eggs broken in the course of making the first iteration of the omelette.
With practice, the mental share of the skill gets carved away to a flexible minimal representation that operates automatically and can be tweaked.
So how does it feel to learn? Really all depends on how you slice it.
"How much does real-time sensory input impact the nature of sentience? "
Greg Egan explored this concept in his novel Permutation City. If you simulate a conscious brain on a computer, does it matter how fast the operations are performed? What if you pause the program and restart it later or what if you run it really slow? Would the simulated brain even notice? Probably not, so all that matters is the subjective experience of time, and not any objective passage of time.
As for feeding back output data to an NN, this is how GANs work (DALLE-2 is based on a GAN, I think). In a GAN, a discriminating NN tries to differentiate between real data and generated data, and a generating NN tries to maximize the error of the discriminating NN based on the output of the discriminating NN. So the system as a whole is feeding its output back to itself.
Judging my interest in the book reviews from titles alone:
* 1587, A Year of No Significance: The Ming Dynasty in Decline by Ray Huang
I'd like to see more Chinese history. A lot of what's out there has barriers to consumability for a Canadian like me. A good review can help sort that out.
* In Search of Canadian Political Culture by Nelson Wiseman
The title caught my interest. So did the reference to Albion's Seed. I'm also starting to think that we may be at the start of a new paradigm shift in Canadian politics, so this is good time to read a history.
* More Work for Mother: The Ironies of Household Technology from the Open Hearth to the Microwave by Ruth Schwartz Cowan
Recently Technology Connections released a video on the modern style can opener. He makes a point in the video on how small household inventions don't seem to catch on anymore. It's something I've been thinking about since, and this book seems to be in the same area.
Is this a new Machinery of Freedom? For most of these I think there is a 50% chance I'll read the book if the review is good. For this one, I'll just read the review. Aside: I loved the intro to Scott's review of that book.
* Public Choice Theory and the Illusion of Grand Strategy: How Generals, Weapons Manufacturers, and Foreign Governments Shape American Foreign Policy by Richard Hanania
After reading The Dictator's Handbook, a lot of books about politics have become hard to take seriously since they over-attribute everything to the person in charge. This book sounds like it could be different. The specific topic isn't interesting enough to get me to read the book, but the review seems interesting.
* Rationality: What It Is, Why It Seems Scarce, Why It Matters by Steven Pinker
My only interest in this is the name Steven Pinker.
* The Dawn of Everything: A New History of Humanity by David Graeber and David Wengrow
There is two reviews of the same book.
* The Righteous Mind by Jonathan Haidt
I've listened to this book three times now. It is my ideology for understanding ideologies. It is my hammer that makes everything look like a nail. I am very interested to see someone else's review of this book to see what they learned differently from me.
* Yąnomamö: The Fierce People by Napoleon Chagnon
I'll probably just read the review. The subsection title "They’re kinda dicks" really grabbed my attention.
I don't know if it is the same everywhere, but cans that need can-openers are almost a rarity now. Of the top of my head I can only think of the cheapest beans at Aldi, and the traditional canned steak and kidney pie (or variant) - which is admittedly a tough opening challenge even with a decent can opener. Almost all cans have pull tabs.
I think that the ring pulls hit a peak back in the 2000s, then the soup manufacturers realised that they could make more money by selling soup in reheatable bowls for convenience for lunch at work (bowls being more expensive or lower volume than tins). Then they started removing ring pulls from regular tins of soup to make them less convenient to eat outside of your own kitchen. No evidence to back this up, but it's the only reason I can think that most soup tins (in Canada at least) no longer have ring pulls, but tinned tomatoes etc. do.
Admittedly, I've bounced around the world quite a bit and maybe my series of events is affected more by geography than time.
I've watched that Technology Connections video, and while I enjoyed it (and will likely get the new style can opener if/when my current one breaks, although I purposefully got a high quality one that is likely to last a long time), I'm not convinced that it says anything about the likelihood of small household inventions to catch on. I had _hated_ the style of can opener for a long time and the reason was that it was non-intuitive and every time I had encountered them in the wild, I was not provided instructions leading to a frustrating experience.
Basically, I think that they were a special combo of small improvement and non-obvious difference in how to use them (they _look_ pretty similar to standard can openers, so there is no reason to intuit the little push-catch system) that made them have a high liklihood of a bad first impression.
Something that had just _one_ of these two features (either small improvement and identical operation or small improvement and hugely obvious difference in operation, or else a large improvement in operation that justifies further investigation in operation) would not have the same problem in catching on.
Additionally, one of my parents _loves_ buying various "as seen on TV" kitchen gadgets. In my opinion they have a very low hit rate, but as long as people like them exist, I think that adoption of new small kitchen gadgets will continue.
Any way? yes. A way that's possible given the kind of minimal resources I could personally bring to answer the question? Doubtful. I would note that at my clinic we have a team of "Clinical Care Managers" who bill insurance for time spent helping the patient navigate the healthcare system. If you could find that data, you might be able to extrapolate how much time is spent outside of the doctor's office doing things like coordinating visits, transportation and the like.
A lot. But the tradeoff is that if we hired a professional class --- lawyers and clerks and a big Federal Government agency -- to do it instead, the total cost would be even higher, because those people would have far less information on each individual case, and be wrongly motivated (i.e. by their own paycheck instead of balancing health/cost/convenience issues with a precise appreciation for their respective values to the individual).
Yes, you also have to be cognizant of political influence. Even if you had competent, selfless, and motivated people working at your federal agency, they'd still be working from legislation and regs that had basically been written by lobbyists. Medicare and Medicaid reimbursement policies are hardly free of meddling from industry groups like the American Hospital Association as it is. This would only get worse if all hospital revenues were derived from government sources.
Thinking about this...some of the things you have listed vary in value between different people. One person might want care close to their home, one might want a certain rating of quality. Time spent driving my mother to a clinic - I might be resentful of that, while my sister feels valued and honored to be of assistance while she is helping Mom.
It's also interesting that you say 'American insurance system' - all of the other (American) systems also have issues with time, difficulty with treatment access, etc.
I think the question is not so much "time spent dealing with doctors to get treated" as "time spent on the phone with insurance agents" and possibly "doctor's visits that exist purely to satisfy bureaucratic requirements"
Trying to get in touch with a bunch of (mostly) American VCs for several reasons, I have contacts but some are too tenuous for me to get an introduction through them (I could bring up "btw I know X" but it would be weird to ask for an introduction).
Is cold emailing at all effective? Is LinkedIn at all effective? Any pointers appreciated.
To be clear this is not primarily about a startup that needs funding otherwise I'm sure I could find a funnel on their site or somewhere similar.
Are you cold emailing the individual or the firm? If its the firm, it will likely be reviewed by an assistant or a very junior associate.
If it's the individual, keep it short (can you communicate it all in the subject?) and make it very very clear what you are asking of them. And if you can, provide some value to them as well. That should increase your chances of a response though I think the odds are pretty low no matter what.
My old startup kicked my research group out/forced them to spin out (but gave $20M in process) and the director of that group was like George Church's golden boy and has still had a really hard time securing VC funding - I would just recommend trying to be impressive both financially and technically while having persistence so that your pitch is showstopping when you do get an in
See, now *that* would be an actually interesting AI experiment. None of this writing poetry or playing chess humdrum. Write a piece of code that, presumably through an extended process of experimentation with thousands of fake accounts, determined the exact boundary of permissible speech on a web forum, and was able to predict with pinpoint precision the distance of any given comment from the boundary.
If someone could do that, I'd be genuinely impressed, as it would have solved a difficult general intelligence problem, which even human beings get wrong from time to time.
Go and see Everything, Everywhere, All at Once. It was weird and indescribable and riveting. The trailers did not do it justice. It's not merely great, it's an opportunity to see genius showing off.
Might as well be subtitled "Michelle Yeoh and the Multiverse of Madness", and I expect I will rank it ahead of whatever the MCU puts out later this year. It's very, very good, and should be seen unspoiled.
The multiverse stuff was, imo, the delivery vector. The genius I saw in the movie was ...
<SPOILERS AHEAD, OBVIOUSLY>
.
.
.
.
how it found meaning in total opposites. We saw high-octane action, horror, and comedy in the mundanity of a tax office. We saw existential dread in the banality of a bagel. Raw human emotion in snarky demotivational posters of rocks. Touching romance in the fucking hotdog world. Regret and malaise at the movie premier where she has it all. The kinetic villains were a teenager and a middle-aged woman. The climactic moment of breaking bad cycles happened at a traditional celebration of a new year at a laundromat.
On the flip side, we didn't get to watch the action in the post-apocalyptic alphaverse, even though surely it must have been in some objective sense very cool. Because that's too easy, too on the nose. All we got from them was exposition and human relationships.
And it was so well made that they told us the game and we still didn't see it coming. The "do the opposite to unlock your alter universe ego" *is* the whole shtick. It's really hard to find meaning like that! I struggle to imagine a mediocre version of this movie, just very good and very bad implementations. If it had been 90% as good I think it would have felt 10% as good. That's why I find it so audacious.
Moreover, I think the movie only works because of that audacity. Because if there can be action in an IRS office, there can be action anywhere - even in your or my life. If they can access their humanity in such a bleak and lifeless place as the rock world, surely anyone can access it anywhere - even you or me. If they can change in their moment of grinding tradition, surely you or I can change and improve right now. The movie felt so universal and personal not because it said any of those things, it would never be so crass as to say them out loud. It just showed them, seemingly effortlessly, while winking at us.
-edit- this comment contains very minor spoilers about general theme and tone of the movie. If you don't wish to see those kind of spoilers, skip this comment.
I am a big R&M fan, but I still thought the movie was amazing. Probably because the multiverse stuff was not the _point_ it was the McGuffin used to deliver a family drama.
I thought that the movie was simultaneously the best comedy, drama, and action movie I had seen in a long time (although I'm not big consumers of either drama or action/kung fu movies on the regular, so my rating in those categories should be taken with a grain of salt).
If you are viewing it purely as a sci-fi exploration of multiverse shenanigans, then yeah, it was...fine, but not great. The details are almost entirely omitted and ramifications are barely even touched on.
But as a vehicle to deliver a touching story about a family on the rocks with well done fight choreography and (IMO) hysterical physical comedy, it worked really really well.
I'm not sure I enjoyed it, but I agree that it's masterfully made and that you want to go in cold. It's weird, but given the weird concept at its heart it's the best movie it could possibly be.
Very hard agree. It's one of the best movies I've seen in the last ten years, and I have extensive knowledge on the subject and excellent taste.
My advice for anyone interested in seeing it is to go in absolutely dead cold. Don't look at the poster, don't read reviews, and especially don't watch previews. Just go in and see it.
"start a conditional prediction market ... if the prediction market is higher than 25% then you can send me an email with a link to the market and argument and I’ll look at it."
Isn't there something distortionary about this? E.g., suppose the market were at 30%, I believe the true chance of it being worth reading is 0%, and I have unlimited money. Ideally, I'd bid the price down to ~0. But then Scott doesn't look at the appeal, the market won't get resolved, and I make no money! Is there ever a reason to drive the price below 25%, regardless of your true belief?
Well, if you think it's zero and you bid it down, the person who wants it to get reviewed needs to bid it back up, so you take that as profit. I think the problem is more tying money up in a conditional market which might not pay out to anyone.
Scott could fix this by randomly choosing prediction markets trading at below 25% to look at, and resolving them. But in practice there aren’t going to be enough markets for this to really work.
Found this piece of comedy gold in a freewrite I did a few months ago (because I'm usually a mediocre writer - at least personal-writing-wise), so I'm posting it here now. Feel free to analyze it to oblivion and beyond.
> The moon is not made of cheese, as is commonly thought, but is made of rock. The sun is also not made of cheese, though far fewer believe this, but the sun is made of plasma. If the moon or the sun were in fact made of cheese, I would expect that their sizes would be quite different, because cheese, rock, and plasma have different densities from each other, which means that equal masses of these three substances would take up different amounts of space. Also, if the sun were made of cheese, I think that the gravitational pressure alone would be enough to make it burn and turn it into plasma again.
The correct paragraph to write next is to first say what the type of cheese is, and then calculate out what would happen, and then be able to definitively say whether or not a cheese sun would become a sun again. :)
Really enjoying checking out all the book reviews! One of them is mine. I'd love to assist in the review rating process, but want to make sure that's kosher first. Are we assuming that everyone will give their own review a 10, or banning the practice of rating one's own review?
The form for submitting a response asks for your email, partially to prevent fraud. So I'm going to give mine a 10 using the same email I submitted my review from and if that turns out to be the wrong decision it should be really easy for Scott to fix by matching authors to self-reviews (but I don't lose out if other people are doing it and Scott was expecting them to).
As an aside, the quality of entries is absolutely *crazy* - I don't know how many are going to make it through to the finals but I'm up to double-digit numbers which I wouldn't be at all disappointed to see win. It would be great if there was some way to preserve the reviews scoring above some threshold but which don't make it to the final round, because there's a huge amount of excellent content here.
Reading a biography of Angela Merkel called "The Chancellor" by Kati Marten. Written before the recent war in the Ukraine it was interesting to read about Vladimir Putin's relationship with Merkel and the west in general. In 2007 at a meeting in Munich he was highly critical of democracy and the nations that support it: "His stated goal had become to reclaim Russia’s place as a formidable global player by any means necessary". He also didn't like the criticism coming from a reporter about the war in Chechnya and somehow she was murdered outside a Moscow apartment on Putin's 54th birthday. Elsewhere he proclaims: "His ultimate goal is to weaken the European Union and its ally the United States." and he feels the Soviet collapse was “the greatest geopolitical catastrophe of the twentieth century”. Seems like a nice guy though ;). Here's the full quotes from the book:
"On February 10, 2007, the somber prime minister of a resurgent Russia strode onto a stage in Munich to deliver a scorching diatribe against democracy, the West, and everything for which Angela Merkel stands. “Russians are constantly being taught about democracy, when those who teach us do not want to learn themselves,” he rebuked the gathering of transatlantic security specialists and government officials. Gone was the accommodating Putin of just a few years earlier, grateful to be a part of the European family and proud that the German chancellor spoke good Russian. His stated goal had become to reclaim Russia’s place as a formidable global player by any means necessary. Blending lies with threats, he taunted the audience, deflected hard questions, and punctured the West’s moral superiority. “Wars have not diminished,” he charged, in spite of the West’s attempts to broker peace around the globe. “More are dying than before.” Though Putin had not yet thrown his support behind Syrian dictator Bashar al-Assad’s genocidal war against his own people, he scolded Washington for its wars in the Middle East and referred to the Cold War as a “stable” era. Merkel, sitting in the front row, was visibly shaken by the Russian’s venomous performance—and his description of the system that had kept her its prisoner for thirty-five years. Not since Soviet leader Nikita Khrushchev pounded the UN podium with his shoe in 1960 and earlier proclaimed, “We will bury you!” had the world heard such vitriol from a Russian head of state. But Khrushchev thundered at the height of the Cold War; this was 2007. Things were supposed to be different now. Yet for the next decade and a half, Angela Merkel’s relationship with Putin would be her most frustrating and dangerous. It would also be her longest relationship with a fellow head of state, its roots reaching back to November 9, 1989."
"Vladimir Putin, once a proud standard-bearer of the humiliated Soviet Union, had learned a lesson he would not soon forget. Unchecked demonstrations and sudden eruptions of freedom can topple even the world’s most heavily armed empire. His battle to reverse what he considered to be “the greatest geopolitical catastrophe of the twentieth century” (Soviet collapse) would ensnare Angela Merkel, a product of the same failed state. Their convoluted relationship would zigzag between faint hope and despair on her side, and dogged determination on both their parts. She was chancellor of Germany, and he was the modern-era czar of Russia. Divorce was not an option."
"From his perspective, the Cold War did not end in 1989; it merely took a short breather. Since then, Russia’s tactics had evolved. While the Soviets brandished nuclear-tipped missiles, Putin opts for weapons that are less conventional and less visible but ultimately more flexible and effective, such as spreading discord in the West through disinformation and cyber warfare, Putin sees himself, in his own words, as “the last great nationalist.” His ultimate goal is to weaken the European Union and its ally the United States. “The main enemy was NATO,” Putin said of his KGB service in Dresden"
"But he failed to intimidate her. In Dresden, the site of Putin’s deepest humiliation, Merkel even flipped his script. It was she who both diminished and humiliated him. The leaders met in his former town in October 2006, three days after the Moscow murder of Anna Politkovskaya, a reporter and human rights advocate whose coverage of Russia’s savage proxy war in the republic of Chechnya had gotten under the president’s skin. When Politkovskaya was shot dead in the elevator of her Moscow apartment building on Putin’s fifty-fourth birthday, some observers felt the timing of her murder was not a coincidence."
Well, for a start, the Russians shouldn't have blown up hundreds of their own civilians. If not for that, I assume they would have had little reason to have a war in Chechnya.
I am neither the OP or the book author, but I guess Chechnya should have been given independence and then attacked in case they decided to create some sort of a Caliphate. But not attacked in the medieval way they were attacked.
Generally the russian lack of care for civilian lives (or any lives for that matter) is the biggest issue here. In Chechnya, in dealing with Chechnyan terrorists in Moscow (and that russian kindergarden, I forgot where that was exactly), in leveling of Grozny, Aleppo, Mariupol, in their terrorist tactics elsewhere in Ukraine...without all of this, at least the war in Chechnya could have been seen as reasonably legitimate by the West.
The US had its blunders in Iraq and Afghanistan but those atrocities were documented by US press, the soldiers responsible were actually persecuted and jailed. Russian soldiers are rewarded by their institutions for even a lot more ghoulish behaviour.
I'm looking for good sources on war in Ukraine, and everything related.
Could be anything from concrete experts that you deem competent to podcast series, media outlets or blogs and much more. I don't care much if this is on military strategy in UKR or global consequences and international allies or any of the many other related issues. Instead, I do care for analytical power and/or high-level expertise. I'd be very grateful for your hints and recommendations.
are pretty good. And reliably consistent with almost all of the other sources I track; I'm pretty confident there is a cluster of accurate reporting here(*), and if I don't have time to dig in to everything those are the first two places I check for the day's news from the front.
* Within the limits of the fog of war, of course, but also in a world where any nerd with a credit card can rent a pretty good spy satellite.
Well, that's my favorite topic atm. On Twitter (@DPiepgrass) I've been following and retweeting stuff from some guys that superforecaster @ClayGraubard follows, as well as his @WarSignals account which doesn't tweet very much anymore.
On YouTube I'm watching these channels, roughly in order from most to least informative (but all these sources provide reasonably good and useful info; the first one is "Perun" - I hope you like powerpoints!):
My favorite thing from the beginning of the war was @KofmanMichael Director, Russia Studies at CNA, explaining why the Russian military's behavior was "absolutely confounding and bizarre": https://www.youtube.com/watch?v=zXEvbVoDiU0&t=1474s
Personally I spent far more time by following the news about the war than I should, so I cannot recommend you to follow my example, but my main sources are 1) updates from the Institute for the Study of War, 2) Bloomberg special daily newsletter on the crisis which has a lot of links to stuff they have beyond paywall, but even its free part is very informative (https://www.bloomberg.com/account/newsletters/ukraine?itm_source=inline), and 3) regularly updated map on https://ukraine.liveuamap.com/. Last one is of course not exactly analytical.
It means that I paid more for a subscription than required amount, so I am apparently "Founding member". No material benefits or premium content compared to standard subscribers associated with that though, just this blue thing :-)
Yup, it is a relief that the mainstream right treats Putin's war as bad, so left & right are getting similar information for once without so much bickering!
You get an insane amount of in-depth historical analyses in this thread of threads by Kamil Galeev. Expertise and analytical power are very high, and he also knows the Russian perspective (strongly rejects it, but analyzes it).
Glad to find someone else who also recognizes Galeev's bs. He is useful for trivia facts about Russian history or linguistics or what have you, but his conclusions are often wrong, and he draws in misleading pieces of information in between. I would avoid Galeev's threads or read with extreme caution.
I rate Karlin at only slightly above zero credibility after he predicted with 90% confidence that Ukranian resistance would last less than two weeks. Looking through @RALee85's thread on Russian losses, there are certainly a lot of minority casualties: https://twitter.com/RALee85/status/1500794161963937793
(Note that that thread currently extends to about ~400 tweets, so you have to scroll for quite a while to really get a sense of the demographics).
Now, there could easily be less than 50% minorities in there, and I agree that Galeev likes to play fast and loose with quantifiable facts to fit the grand historical narratives he likes painting, but I think the amount of new insight he provides is well worth the price of admission. I'd say read with a critical eye and understanding of his style, but read nonetheless.
I have found Vitaliy Katsenelson very helpful for background - he's best known as a financial writer, but he was born and grew up in Russia and his ancestry is part Ukranian, so his context is very helpful.
One way to get updates is to take two opinionated opposing sources who are directly involved in the whole thing - for example, if you understand Russian, Alexey Arestovych from Ukrainian side (his youtube channel at https://www.youtube.com/c/arestovych) and Igor Girkin/Strelok from the Russian side (his Telegram channel at https://t.me/s/strelkovii) both have interesting analytic comments but naturally have a quite opposing perspective of what 'should' happen due to their role in the war (Arestovych being an advisor in the Zelensky bureau and Girkin being a leader of the initial 2014 invasion and currently having close ties to DNR/LNR troops).
No frills, no obvious spin assessment of intentions, actions, and results of the various fronts, with plenty of hedging when it comes to unconfirmed stuff. Mostly reliant on OSINT but still pretty good analysis.
So I'm uninsured and pay cash. This works well for many doctors. My PCP gives me a cash discount to $100 per visit vs. $125 for the non-cash patients. My old shrink gave me a cash discount also. (Cash is great: no counterparty risk, no insurance overhead, no billing overhead, no collections overhead, no credit card fees/delays, and maybe even lower taxes ;) At some other doctors this is horrible. I saw an urgent care doctor for 2 minutes and then waited around for their minion to perform a rapid strep test, then paid $140 by credit card. I thought that covered it all. Then a week later they mailed me a surprise bill for an additional $350, which was ridiculous and I never agreed to, except in the sense that the legalese bs you sign at the beginning is a blank check for them to bill you anything for anything without your informed consent. I could have stayed home and done my own strep test for less than $2 per test. I hate the way the healthcare system works like that. Advance disclosure of prices should be mandatory. When I go to any clinic I should demand they tell me what they're going to bill me in advance or else I'll walk out, give them a negative review on google, and go to their competitor, rinse and repeat until I find a clinic that doesn't suck. Unfortunately 99.9% of people either have insurance or aren't that assertive, including me, so we get this fucked up system where you have to sign a blank check to be seen by a doctor, instead of transparency and informed consent about billing. Why is society so adamant about transparency and informed consent about research but doesn't give the slightest fuck about transparency and informed consent around medical billing? Seems like the latter is way more likely to fuck up someone's life than taking a survey that some IRB objects to.
policy idea: as a condition for accepting medicare, doctors must agree not to charge cash patients more than medicare would pay for the same procedure. This would greatly reduce the variance and the risks of surprise billing to uninsured patients. I bet that urgent care clinic is getting less than $200 from medicare instead of the $500 that it billed me.
I don't think the rule applies to walk-in clinics, so it doesn't help your situations much, sorry. But as Monsieur Cury said above, just because they sent you a bill doesn't mean you have to pay it.
Some tips on your specific issue that applies generally to US healthcare: you can still argue/negotiate/fight the total bill. Depending on what they put into the claim coding system, you can have it updated either by analyzing it and talking with the billing department, or you can wait until it goes to collections and pay a fraction of the cost to whoever purchased the debt.
Source: I used to assist people with correcting billing errors, etc.
Follow up reading: "Never Pay the First Bill" by Marshall Allen
Follow up Watching: Dr. Eric Bricker on LinkedIn/ YouTube (I used to work for Dr. Bricker)
That is a uniquely American problem and it makes no sense to me, quite frankly. I admit I often don't ask how much things will cost in advance - if I need it, I need it - but I've never gotten a surprise bonus bill after having paid at reception.
Similar happened to me. Went to doctor, payed some out of pocket. After the visit, I asked the person at the desk the very direct question, "Will there be any other bills associated with this?" The person responded "No." Few weeks later, bill comes. Out of principle, I refused to pay it. I had medical debt collectors call a few times but after a while they gave up and nothing came of it. I haven't gone to a doctor since then, but when I do I think I might actually do what you suggest (though word everything in a very polite way).
IMX failure to pay medical bills almost never hits the credit reporting agencies for some reason.
Related thought experiment: what if you came to urgent care or ER with specific acute complaints, get through triage and see a doctor, then ask about a chronic condition, get expensive tests and diagnosis, and the whole way through refuse to offer identification? Won't the "duty of care" principle require that they treat you, regardless of your documentation or lack thereof?
You are correct. When I was still in the hospital, people who came in "without ID" were often given a name according to the scheme of "Problem" + "A random City". So you would get pt's named "Trauma Tenerife" and such. No one is collecting from them, and yes, they get standard-of-care treatment. Gone are the days of the wallet biopsy.
Caveat: if their condition was debilitating in such a way as to require rehab, they would not be able to go to rehab without giving ID. If that took too long then risk management would get involved and somehow they would be "given" an ID? Not sure at all how that worked.
People do care about it, but they care about $200,000 bills for cancer treatment or a serious car accident with a two-week ICU stay. Generally people have neither the time or energy to agitate about $350 chickenfeed. I realize it *isn't* chickenfeed to you, but when you reach the age of serious medical expenses in a few decades, you'll look back in awe at the wonder of *only* signing up for 3-digit number by going to the doctor for help.
It totally is chicken feed for me and all about the principle. I didn’t get rich by letting people rip me off. End of life care is one of the worst ways to waste money since it mostly generates negative utility by prolonging the suffering. If I get a curable cancer I’ll shop around or be a medical tourist but $200k would be no big deal either. I am consciously making a positive expectation gamble by not paying insurance companies’ profits/overhead, so long as I can minimize the price discrimination against me as an uninsured person. I hate dealing with insurance companies also - it’s not worth my time to beg them to cover my medication and get no-explanation denials and wait on hold for an hour to talk to a human about it. I’d rather be in charge of my own medical expenses than totally dependent on the whims of some evil third party.
So what's your complaint? You have a plan for dealing with medical care that is radically different from what everyone else is doing, and you know that, and yet you're wondering why there is no large cohort of people with similar feelings about the process? Isn't this logically self-contradictory?
End-of-life care is one of those common health-care magical thinking canards, like you can reduce lifetime cost with preventative care[1]. You could only tremendously reduce end-of-life care if you knew for sure certain *ahead of the fact* that what you were facing was the end of life. In reality, it doesn't work that way. People certainly do recognize the end of life is approaching, and this-and-such expensive procedure may not work -- but they don't know the exact odds it will work/not work, and if it works whether it will give them 48 more hours or 48 more days or even sometimes 8-10 months. Medicine just isn't precise enough to know. People routinely do make decisions economizing at the end of life, but it's a very imprecise process, and of course most people err on the side of hope -- and that's all we need to explain how very expensive the "end of life" is, and how it's essentially impossible to reduce that cost without some kind of arbitrary and cruel algorithm that takes the decision away from patients and families, which only a monster would support.
------------
[1] It doesn't. In fact it makes lifetime healthcare more expensive, because costly to treat diseases are discovered earlier, when there is still time to do very expensive interventions. What it *does* do is prolong and improve the quality of life.
My complaint is excessive price discrimination against uninsured at some clinics in spite of the much lower overhead costs for people who pay cash up front. GoodRX totally solves this problem for prescriptions, but it would be nice to have a GoodRX-equivalent for office visits.
You might want to look into concierge medicine. Those doctors only take cash, so that way you know you're getting the same price as everyone else. Lots of them are available 24/7 if you get established with them.
Sure, but you are already understanding that what you are using is a rare method of payment and budgeting, right? So you need to consider not just the marginal cost of the transaction (paying cash for your procedure) but *also* the amortized capital cost of setting up the system in the first place. The problem is too few people use it for it to be efficient to set it up and optimize the way you're wanting. If a million people a day used it, the way a million people a day use Netflix or something, then you bet it would be optimized within an inch of its life -- because the the huge number of transactions mean any small improvement in efficiency can justity pretty hefty capital costs.
Not in your case. There just aren't enough transacations for it to be worth optimizing. So, again, why is this surprising? You can only realistically expect systems you use that are also heavily used by others to be optimized, either socially or economically.
Not at all. Poor people go to the ER and stiff the hospital on the bill. It's long been one of the arguments for a better network of community non-urgent care clinics, e.g. loosening up some of the licensing and quality control regulations for that, so you get more freestanding cheap clinics in the neighborhood, and/or some kind of more plausible health insurance that covers primary care better, so poor people go to a primary care provider in an office instead of the ER.
1. The federal government makes it illegal to charge anyone less than the lowest price they charge Medicare patients as a requirement to participate in Medicare. Therefore basically you can’t have a sliding set of charges. One price for all (but that’s not what gets paid).
2. Charges are set to be higher than the highest insurance reimbursement so you never accidentally leave money on the table.
3. This leaves people without insurance stuck with an inflated bill that no one really expects to get paid. Generally speaking if you tell the Provider that you cannot pay the bill they will immediately reduce it to Medicare rates.
I don't think that's the explanation for why some doctors overcharge cash patients. I think it's mostly the same reason anyone charges more for anything: because they can. Insurance companies would push back more but individual patients can't without prior info about pricing. Also maybe the pretense of a high sticker price makes the insurance reimbursement price look lower and thus gives providers an advantage in negotiations with insurance companies.
I think in a way you have answered your own question: we put up with nontransparent medical billing precisely because the majority of people have insurance through employers or Medicare/Medicaid and don’t actually pay the bill directly. The St. Louis Surgical Center is an all cash business (if I remember the name); the founder was on Econtalk a few years back talking about it, and how entirely random most medical billing is. When they started no one they worked with has any idea how much to charge and it turned out they could way under charge compared to hospitals and pay the drs more for the same procedures and the same exact staff in many cases. It was a very interesting episode. We got lots of those post process bills after our kid’s births, some from doctors who were definitely not even there. I wish the hospital had a “standard c section is X$” system instead too.
Oklahoma Surgery Center in OKC is all cash and menu pricing. I got my wrist fixed up with titanium for $5k about 7 years ago.
Until contacting them, I called all over the Bay Area and could not get anything close to a price quote except from the hospital that imaged and splinted me, who after tooth and nail quoted "somewhere north of $10k" for the necessary surgery.
I run ReadSomethingInteresting.com. Does that look similar to what you're envisioning? Depending on what would be needed, I'd be happy to adapt the code.
Anyone good to read on the shakespeare authorship question? Have a feeling this is a prime example in a poor ability to interpret information, lack of clear statistical priors etc. Often the arguments sound good when you ignore fairly mundane positive info. Anyway was just hoping someone around here would know. Thanks
I don't know if it's "good", but "William Fortyhands" is a really unusual take on it that doesn't specify a single person who wrote all those plays, but instead argues that Shakespeare was a producer who made the "bad" quartos audiences of the time actually wanted, rather than the artsy-fartsy "good" versions he bought from some poor scribblers.
Great thanks! Yes, I don't know where to being and it seems like there are a large number of varying stances taken within the Anti-Stratford 'group', so often legitimate queries get lumped in with the less evidentially reasonable claims. To clarify my current feeling is that with someone as big as Shakespeare, who wrote as much as he did, I would be surprised to not find some kind of 'code' or suspicion about his identity (especially given the historical period). However, I don't want to dismiss anything out of hand. I've just noticed that lately it's become much harder to know how to properly respond/dismiss outlandish-sounding claims. (or maybe, we're just more aware of the need for skepticism, and how deep a requiremenet that often is). Not that they ought to be dismissed, but, it would be impossible to live if you accepted as worthy-of-inquiry every theory that flies your way, so I'm often at a loss and just resort to "I'll have to look into it later".
It’s absurd. They find minor issues with Shakespeare - often spurious, like claiming that he couldn’t spell his name, the spelling of which is arbitrary. Ignore or dismiss most of the overwhelming evidence of Shakespeare being both actor and playwright, including the first folio produced by his friends and colleagues, and fellow poets. Then the other candidates are often dead for some of the period, or out of the country, and none have any relationship to the the Shakespeare theatre group documented anywhere.
Yes, thanks, I agree with you. The name thing in particular is a little odd to me, since they (first filtering out many claimed but not 100% verified examples of his handwriting) base it off of some shaky signatures he wrote in the last 2 weeks of his life; probably he just had some illness. However, the anecdotal evidence for -- and I only know very little about it -- e.g. de Vere (that he led a very similar life to e.g. Hamlet), and issues such as Stratfordians acknowledging there isn't any evidence that he was educated/highly literate (without assuming he wrote the plays lol) or had a library... But every time I look at one of the claims closer it's usually either an overexaggerated anecdote, something TOTALLY bizarre where the impose some odd way of 'skipping letters' to get 'codes', or it's some weird interpretation lacking any statistical background as e.g. you bring up "overwhelming evidence" -- it's all just discounted..
Thank you, this is the sort of thing I was looking for! His name came up a few times already. Yes, I imagine it's difficult in general. I was just reading recently that "Absence of evidence is evidence of absence" idea, and was wondering if it came into play in these interpretations. Something I think about a lot is how 'vagueness' is almost definitionally something which allows multiple interpretations (lots of different shaped jigsaw pieces which all agree on at least one edge). But it also seems like the two groups are Stratfordians, who aren't really interested in the question of legitimacy enough to write about it, and Anti-Stratfordians, who have explicit outsider motivation to dispute it (not that that's a bad thing, e.g. Court systems). Thanks again
There's a mendelian study on the association between alcohol use and cardiac disease that claims a straightforward *positive* correlation between any amount of drinking and higher risk
Previously all studies show a large decrease in mortality risk with moderate drinking, which I believe is dominated by a negative cardiac risk correlation
So what this study does is isolate certain genes found to be correlated with problem drinking and uses the prevalence of the genes to rate the person's likely drinking amount, and then correlates their risk of cardiac disease to that. The study does mention the problem of using genes tied specifically to problem drinking and notes that the study screened out actual problem drinkers
Should I consider this to be reliable evidence? My current attitude is that it isn't but is worth being aware of and moderating how strongly I assume moderate drinking is beneficial
Related: On the Skeptic’s Guide to the Universe podcast I remember they talked about a meta-study on this topic a few weeks ago that found the dip in correlation for moderate drinking was likely driven by former alcoholics who no longer drink at all, but for whom the damage was already done. I don’t remember if the meta-study was able to actually correct for it, or if it looked at a subset of studies that included only people who never drank heavily, but the result was a more classical dose-response curve.
This suggests it's not (I think it's about the study you're talking about). But I'm not that familiar with the literature or the method used so don't have an informed opinion otherwise.
I was reminded of this concept by Scott's recent post on AI. Can anyone lay out a situation where acausal trade is plausible? This is related to a comment I made on another post, but roughly I think people jump from "very smart thing which can do things we cannot" to "thing which can accomplish anything which can be stated in English." When someone mentions something like acausal trade in the context of AI, I immediately stop paying attention since I assume they're not a serious thinker.
I think it happens sometimes in government regulation. Regulators go easy on companies, not because of any explicit quid-pro-quo, but because they hope that that will make them more likely to get a cushy industry job after leaving the government.
Of course, Matt Levine says that incentives usually go the *opposite* way.
My understanding of what people are referring to with acausal trade is that neither party ever actually interacts with the other. In that case, the industry people are affected by the regulator's actions, and then have future power over the regulator. I may be misunderstanding what people are referring to though.
Acausal trade means you don't need to interact *to negotiate*. I think that the weak version is very easily constructed in plausible thought experiments (plausible by thought experiment standards, maybe at the level of really consistent spec-fic), and the version I've seen discussed involving no interaction literally ever under any circumstances is pretty clearly dumb. But like, if you were playing prisoner's dilemma with a copy of yourself, it's pretty clear you could "decide" what the copy does by making your own choice. How much and how fast does this crash and/or burn when it collides with reality? That's what people argue about.
A copy of oneself I can understand. Even something like "humans from similar cultures to yourself" I can see working. The times I've seen it mentioned though, it's been in the context of simulating other currently non-existent intelligences, which isn't plausible.
I mean OK, say you got scanned today, but the copy gets printed out tomorrow. The logic still works out no? Existence isn't the limiting factor, knowability is. If you think agents are too diverse to predict well, then obviously acausal negotiation can't happen. But if there are Schelling points or something like an attractive basin of agents, it's certainly plausible. Likely? Well I suspect not, but I'm definitely not positive.
edit: Rereading your comment I realize that it's hardly a leap to consider a copy of you to be an existing intelligence. If by nonexistent intelligence you mean unknown, like totally novel intelligences with unknown desires, the cheeky:useful ratio of my example spikes. Worth leaving in for the disconnection from strict causality, but it doesn't say anything interesting about your comment. Sorry about that.
It's a hell of a lot easier to work out what a "perfectly logical being" would do, even as an illogical being yourself, than figuring out another illogical being (including, at least for me, myself!). Not a real reply to your comment on acausal trade, which does not rely on perfectly logical beings and seems largely orthogonal to moneymaking, just a quibble.
I think there's an issue in that no one has the computational power to be perfectly logical. And what something with 10^12 FLOPS vs something with 10^20 FLOPS will be able to deduce is very different.
Well that might be true if we could all agree on a working definition of "logical." Unfortunately, outside of mathematics, we can't, so while yes *in principle* if we knew the objective final rules of logic, we could work out what a logical person does very well -- even not being logical ourselves -- in practice this is completely impossible. It's rare for even as few as two people to agree on what is logical and what is not in most situations -- just ask any long-term married couple.
I mean if you think logical should include the "right" axioms then we don't have a working definition of logical even within mathematics. But a team of smart people working the same angle could definitely figure out what a rational actor would do a decent portion of the time, that's exactly what happens in war rooms around the globe. You know what they want, and know basically what the best way to get it would be.
I'm going to suggest that the ability of people in war rooms to figure out what opponents are going to do stems from the same source as the ability of good politicians to get (and stay) elected, good CEOs to lead successful companies, and good coaches to get championship performance out of teams.
Id est, it has nothing whatsoever to do with logic, and instead stems from a highly accurate intuitive grasp of the feelings and emotional motivations of other human beings. It's certainly logical, in the sense that feelings of a certain sort predictable lead to actions of a certain sort -- but this is pretty far removed from what most people mean when they think of a "rational" person figuring out what another "rational" person will do.
Regardless of the fact that emotional motivations in no way preclude rational behavior (and in humans are required for it), a general working out that the enemy will try and hold a defensible mountain town because that's what they'd do in their opponent's shoes is not using an emotional intuition at all. The enemy wants to win the war, and the general knows that holding the town is the best way to do that.
I was caught totally off guard by both DALLE and Alpha Fold Ai in the last 6 months, both strike me as hugely important. Can anyone give me clues as to what might be the next big AI to make headlines?
Also, is there a list anywhere of all the psychological and biological research that has been discredited by the replication crisis? Or maybe I need to read review articles?
I'm another who was meh about DALL-E but taken off guard by DALL-E 2. It's a huge step up. 1 looked like AI-generated art, with all its oddities and surreal dreamlike features, but 2 looks like human art and could make human artists redundant.
Yup. An answer to OP's question is thus "a GPT-4 which gets rid of most of its predecessor's quirks". I expect it any day now. GPT-3 is really impressive but not good enough for most serious uses, which is also how I would've described DALL-E 1.
Well, they have been making improvements to GPT-3. The default models the API uses are now the Instruct line. Since there's publicly usable incremental progress, that lowers the likelihood that there will be a "sequel".
If the twitter board persists in blocking the offer despite lacking any reasonable belief that they can get a better offer, they're probably violating fiduciary duty. The $54.20 offer was already a 37.8% premium above the closing price the day before it was announced.
$54.20 is also significantly below where the stock closed in October of 2021. Why is April 13th closing price indicative of the intrinsic value of the company? The board has a duty to assess the offer, not accept it just because it is higher than the current price. Any shareholder that bought above $54.20 wouldn't think the board is doing its duty if the offer were to be accepted.
Elon also has a history of not having real financing lined up before making an offer to take a company private! If the board (and its bankers) doesn't believe he can actually come up with the money to complete the transaction, then they would negligent to accept the offer.
That's not a formal offer. That's a letter *to the board* *proposing* an offer. He hasn't made a formal offer yet, because that would require him to actually come up with $40 billion in financing, and he is unlikely to be able to do that without selling his stake in Tesla.
Matt Levine at Bloomberg was making that argument. Musk has already borrowed against about half of his Tesla shares, and he'd have to get board permission from Tesla to borrow against another $40+ billion against them (which he would get, but it's a "shit or get off the pot" moment). He hasn't done that yet, which makes his proposal kind of dubious.
I suspect what happens is that he just sells his shares in Twitter at some point in the next month or two and walks away. Something else will draw his attention.
I suspect that your penultimate statement (that this is rankly illegal) is true. I'd be curious about the more relevant question of "Will Twitter's board face any legal consequences for violating their fiduciary duty?"
My personal prediction is that they won't, because this is a violation of fiduciary duty that benefits the wealthy at the expense of the poor and not the other way around, similar to how sites like Robinhood blatantly trying to interfere with purchasing Gamestop stocks and trying to abet a squeeze against the direct wishes of their clients seems to have walked away without so much as a finger-wag.
PS: It looks like someone decided to short the market- value is at 26% after briefly spiking to 100%.
Why does it benefit the wealthy over the poor? Elon Musk seems like the main guy who gets hurt, and he is one of the richest people in the world (if not the richest). More generally, it seems like everyone involved here is rich. The board is rich (presumably), Elon is rich, and one of the largest current shareholders is a fund headed by a Saudi prince (and more generally, any publicly traded company is probably held pretty disproportionately by rich people).
Also I'm not an expert on this but I think that there have definitely been takeovers at a larger than 37% premium before, and times that boards have held out and gotten a better offer, so it's not obvious that holding out for a better offer is worse. The Saudi guy said so for example.
Overall, it's not obvious to me whether on net poison pills are good or bad for shareholders. You can think up scenarios for both.
Yes, Musk is very rich, but he's not the CORRECT kind of rich. He's a bit too close to the proles in terms of his behavior and interests. Maybe I should have used the term "elite" instead. Musk is incredibly, fabulously rich, but I wouldn't classify him as one of the "elite" class. He shitposts on Twitter, wears hoodies covered with japanese cartoon porn in public, and generally seems uninterested in "elite" hobbies and "elite" politics, instead doing whatever he likes. He also supports a FEW too many things that suggest to the hoi polloi that there's a hypothetical world where they have options beyond bowing down and worshipping at the altar of Conventional Capital. That makes him an outsider from the system- an enemy, even. Crushing his bid BAMN is a way to symbolically make sure the plebs know their place- much like their threats to go beyond merely rigging the casino in their favor and ban the plebs from any kind of direct agency in the stock market over GME.
What does Musk think that "suggest[s] to the hoi polloi that there's a hypothetical world where they have options beyond bowing down and worshipping at the altar of Conventional Capital"? I get he says vaguely anti-woke things sometimes ... but I doubt it even registers to most people as some sort of "'Conventional Capital' vs not" thing. This all seems like some extremely narrow super-online controversies that most people probably don't care about (to most people Musk is probably just a electric car maker who seems smart but also kind of weird and that's about it).
Meanwhile poison pills were invented a long time ago, all that they have to do is rule in line with their own precedents.
I'm not sure what Gamestop has to do with Musk and Twitter or the Delaware courts either.
I agree this is a narrow online controversy. Elites don't have real problems in their lives and are very sensitive to online controversies and the like. See how some stupid 4chan memes have (with only partial success, I'll grant) been elevated to the level of a swastika.
The Gamestop trading halt was caused by people higher up the food chain than Robinhood's CEO. The DTCC temporarily increased collateral requirements for unsettled trades to absurd levels that Robinhood couldn't afford, and by law Robinhood couldn't use clients' funds as collateral for the clients' own unsettled trades, so Robinhood and several other brokers had no choice but to halt trading of GME and AMC. Robinhood shouldn't get the blame. Blame it all on the DTCC and our archaic settlement system.
I was under the impression that the collateral requirements were increased due to extreme volatility in the stock. But yes, faster settlement would be nice, and would have reduced the problem.
> similar to how sites like Robinhood blatantly trying to interfere with purchasing Gamestop stocks and trying to abet a squeeze against the direct wishes of their clients seems to have walked away without so much as a finger-wag.
Probably because this conspiracy theory isn't actually true. Robinhood stopped the trades because they *literally couldn't afford the margin requirements to continue trading*. Additionally, **there was no actual short squeeze**. All the institutional shorts got out of the market in January *before* the stock really blew up. The reason for the spike was simply because *tons of people started buying it for the lulz*, nothing more, nothing less.
I mean, if you want to believe everything Robinhood and Citadel said, can't really do anything to stop you. AFAIK there's no hard evidence against collusion between Robinhood and their primary trading partner Citadel and plenty of circumstantial evidence (including leaked internal communications) for it, and while the class-action was tossed out, individual people are winning in arbitration. Sure, that doesn't PROVE wrong-doing, but calling this a "conspiracy theory" seems far more dismissive than merited.
By "anyone" do you mean "anyone" or do you mean "The vaguely-humanoid arthropods known as the traditional financial class"? Because those have very different answers.
To be clear- yes, I DO think that the traditional financial class engages in some level of conspiracy to disenfranchise the average American. They already do some of it in public and call it "lobbying" instead of "bribery", and I have no reason to believe that what they do in PUBLIC is worse than what they do behind closed doors.
The way it usually gets adjudicated is through proxy challenges. If Twitter's board & mgmt refuse to consider a legit offer that the majority of shareholders think they would benefit from, they can vote the board members out -- eventually. The game theory around this sort of gambit is pretty well developed. It's similar to the strategies of Carl Ichann, Mitt Romney/Bain Capital, etc. I think Warren Buffet could be put in this bucket as well.
But this a rare exception. In theory poison pills are justified by giving the board more time to solicit better offers. In practice, boards mostly use poison pills to keep their jobs and don't attempt to solicit better offers. This usually violates fiduciary duty but they get away with it anyhow.
The steel man use case for a poison pill is blocking a two-tier front-loaded buyout offer that would pay a higher price up front to acquire a controlling interest and then squeeze out the remaining shareholders at a lower price. That style of buyout is basically extinct due to poison pills. It was the style of buyout at issue in https://en.wikipedia.org/wiki/Unocal_Corp._v._Mesa_Petroleum_Co. where the court ruled in favor of the poison pill. IMO both poison pills and two-tier front-loaded buyouts should be illegal.
Oh, this is great! I'm delighted with the selection of reviews and already see several I want to read. There's also one book on here I've read, disliked, and won't review but I'm interested to see how others feel about it. This is so good, this will keep me going for my Easter break!
The book review contest is a wonderful idea, thanks for doing it, Scott.
I was a nerdy Catholic kid but I locked in on church history and theories on the Gospels' sources, not so much the details of the liturgy. I've been curious about what's distinct about a Battle Mass for a long time but Robert Battle's Mass makes search a frustrating mess.
This is a bad idea for two main reasons. First, is the risk of collecting viruses from wildlife and having them escape accidentally from labs. But also, publishing sequences of potential pandemic pathogens will enable bad actors to credibly threaten to cause pandemics. See this commentary by Prof. Kevin Esvelt: https://www.science.org/do/10.1126/f079e4e4-7689-4837-bc83-1f3fd961e946/full/
I'm sure the people at Global Virome Project mean well but as it is currently planned, their project will do more harm than good.
If you are in the USA I encourage you to contact your congressperson about stopping this.
My reaction isn't OMG that's dangerous, it's OMG that's too hard. Actually, I was reading fast, so I saw "catalogue of viruses" rather than "catalogue of viruses and evaluate for pandemic potential."
Just cataloguing viruses is, well, my intuition is that they'll catch 1 thousandth of wild viruses. Let me know if this seems way off.
I will say that people have been wandering around all over the world and eating stuff, and pandemics are pretty rare.
And I was assuming wilderness, but, of course, viruses are all over the place and some of them are probably specialized for civilization.
I suppose that the viruses (too many to completely catalogue?) that are already in humans should be assumed to have low pandemic potential.
My guess is that cataloguing all the viruses is impossible but worth attempting, and evaluating them for pandemic potential is a good way to get funding.
Nonsense. Nobody has the knowledge necessary to interpret DNA sequences and use them to fashion brand new evil sequences. If that were even a little bit possible -- you'd have to have solved the protein folding problem, first, so you knew what proteins were made by what genes, *then* you'd have to have an exquisite understanding of all of immunology at the molecular level, which is several centuries away I would guess -- then e.g. cystic fibrosis and sickle cell would long ago have been trivially cured by viral vector borne genes, and it would be child's play to invent effective vaccines to almost any viral disease. This is the same kind of wildly-overoptimistic assessment of new technology that let people when they first learned of the Internet think it totally possible for a teenager on his TRS-80 to hack into US ballistic missile command computers and start World War III.
I don't understand how this is relevant to the top-level comment. The Global Virome Project is trying to collect and research viruses that already exist in nature but for one or another reason have never made their way to humans. Metacelsus is calling out that some of these viruses may be capable of causing pandemics in humans without modification, and by collecting these viruses, researching them, and publishing their genomes, we are increasing our risk of facing a novel pandemic that would not have happened without the GVP. This could happen either by an inadvertent lab leak from researchers studying the virus, or by purposeful release from a bad actor.
How do you "release" a virus that is already endemic in the environment? Anyway, I interpreted the comment as being concern about the publication of the genome -- and assuming that people could get lots more valuable information out of a genome than they currently can. When the human genome was (almost entirely) published in 2003 people predicted it would be used to genetically engineer humans, genetically screen employees, discriminate in terms of health insurance or in jobs ("Joe has a 37% higher chance of being a kleptocrat, according to his genome, so we better not hire him as our accountant"), pave the way for human cloning, and so on.
None of that actually happened. It turns out knowing the exact sequence of a billion or so ACTGs actually doesn't get you much, all by itself. It's just letters on a page, like so many millions of digits of pi. It's at best the *starting point* for additional research, where you painstakingly cross-correlate sequences with biochemistry, and symptoms, and actually start learning stuff. Even then, it's not 100% clear it wasn't a waste of resources, since we *could* have just stuck to sequencing lengths of DNA on an as-needed basis, whenever other research suggested it might have value.
I think the same thing is likely to be true about collecting viral genomes. It might be useful, because people can do cross-variant comparisons, study how fast they evolve, stuff like that. It's a nice fat data bank. But I think the idea that creating deadly pathogens is just hanging on knowing the exact sequence for HIV or something ("Aha! *That's* what we're doing wrong! It's an adenine in position #6,544 not a cytosine!") is very unlikely.
Protein folding problem = alphafold and then scRNAseq + metabolomics data solves IDRs
We literally engineer immune cells from scratch to cure cancer patients. It's actually very easy/like 30 min a week to make great virus like I did for years.
Our understanding of immunology is at the point where friends tell me I really need to censor my dark humor/jokes on fallback careers in bioterrorism when grants/manuscripts get rejected etc
Sorry, no, none of that is true. Or perhaps a better way to put it is that I think you have so far reduced the definition of "solved" and "from scratch" in your examples that this is not a meaningful argument.
Anyway, if you think you can personally engineer a virus from scratch that will safely deliver working CFTR genes to cystic fibrosis lungs, you shouldn't be wasting time commening on Internet fora -- your Nobel Prize in Medicine awaits, along with enough money to buy Sardinia.
I emphatically disagree with this assessment, and would argue that the Global Virome Project has orders-of-magnitude greater upside potential than downside potential. I think my view diverges from yours (and Prof Esvelt's) based on three points (roughly in order of importance):
1) In terms of preventing the spread of genomic data on deadly pathogens, we lost the game decades ago. The full genomes of anthrax, smallpox, and similar classic plagues are publicly available, for anyone who has internet access and knows where to look. The COVID genome was, quite famously, shared globally in early 2020, almost exactly as fast as was technically feasible. If you're a mad scientist looking for something nefarious to assemble with your evil synthetic biology lab, you're already spoiled for choices. There's little we can do to make your life better or worse on that front.
2) Suppose that there is a virus out there that can spread in humans well enough to become the next pandemic, and that it's common enough in its current animal host for GVP to find it. If these things are true, chances are very high that this virus is coming into contact with humans often enough that it will make the jump to us without scientist assistance. Almost all parts of the world which are inhabited by mammals are also inhabited by significant numbers of humans; humans who, moreover, don't observe biosafety protocols when dealing with the odd critter that got stuck in under their house. We're not faced with a question of whether or not to let a novel pathogen "out of the box" in order to study it. There are no boxes in nature; if it's out there, it's already out.
3) Given 2, we can expect that ~everything the GVP will be a "near-miss" pathogen; something that potentially replicates in humans a bit, or spreads a bit, but not to the degree to explode into a large pandemic. (It could still be bad for individuals - think SARS-1.) If you're an evil mad scientist trying to turn one of these into the next pandemic agent (having decided for some reason not to use already-available anthrax, etc for this purpose) you're faced with the problem of engineering a biological agent to have a consistent, pre-defined effect in the human body. This is the same kind of problem faced in typical pharma or biotech work, and broadly speaking, is very hard to do! I am admittedly not a virologist, but to me this feels like the sort of project where you would need a well-resourced team of smart PhDs to expect success [0].
If the above convinces you that the risk from empowering potential bad actors is fairly low, I think it's pretty trivial to see that the expected value from GVP is very positive. COVID is only part of a general pattern of emergent pandemics or near-pandemics; think SARS-1 or H1N1, or going back further Spanish flu, etc. We can give high likelihood to the emergence of a new pandemic-capable pathogen in the next century. Even in a conservative case, where GVP gives us some insight which merely accelerates the development of a vaccine for a novel COVID-like pandemic, the benefit would be measured in hundreds-of-thousands of lives saved. (See e.g. the recent article https://www.maximumprogress.org/blog/how-many-people-are-in-the-invisible-graveyard .)
(Disclaimer: I am not in any way affiliated with GVP, and in fact hadn't even heard of it before Metacelsus' post. So thank you for letting me know about this neat thing! : ) )
([0] Just to preempt an objection here - yes, I am aware of "cell passaging" methods for evolving viruses for suitability to a given host. My impression is that transmissibility in a cell plate translates only roughly to virus effectiveness in-vivo. In which case, the best an evil mad scientist could do with this technique is probably a sort of scattershot approach where they just keep quietly releasing new variants and 1-in-100 of them are noticeably bad. Not great if someone actually is just a determined anti-utilitarian, but also not really suitable for most kinds of terrorist or etc.)
1) Unfortunately pandemics are not a binary yes/no question. A malicious actor can spread smallpox, Spanish flu, and five new diseases discovered by the GVP at once if they wanted, and it might be 7 times worse than spreading just one since they don't provide immunity from one another.
In addition, Kevin Esvelt argues that all the existing viruses have limitations which make them not as bad as it could be. Some examples include:
- Spanish Flu apparently has enough similarities to modern flus that we should expect a far lower death rate, and the death rate should be further reduced by the fact we have antibiotics (around 1/2 of deaths in the original spread were from bacterial infections)
- Smallpox: We've managed to get it under control before, it doesn't have asymptomatic transmission, and we have about 350 million working vaccines in reserve
- Ebola, SARS, MERS - transmissibility seems to be low enough that we can get them under control, in fact this has been done before
Also, Anthrax is not generally contagious.
2) Could you provide a source on this? The fact that humans and mammals interact a heck of a lot doesn't seem to mean that all viruses will necessarily spread. One piece of evidence in favor of this is that viruses tend to spread in countries with particularly transmissible practices like widespread wet markets or bushmeat markets. Unless you think these countries naturally have more pandemic-causing viruses than others, it's likely there are some dangerous viruses in countries with more safe practices that just never made their way to humans.
In the podcast I mention in a post below, Esvelt doesn't go into too much detail but says that he believes there are diseases that would never get to humans if not for this virus hunting.
3. In the podcast, Esvelt mentions a lot of uncertainty but estimates that the DEEP VZN program, which would find and characterize about 10 thousand viruses, should expect to find somewhere in the mid-single digits pandemic-grade viruses.
It's very unclear how much the benefit to this knowledge is either. Unless we're planning on stockpiling a huge amount of vaccines, or running challenge trials, which I don't think anyone is suggesting (but probably should if they are planning on going through with this), at most we will get a few weeks lead on a vaccine. Remember that the Moderna vaccine took only two days to make and most of the delay was from testing and regulation. In general it's unclear how much would be different if we knew beforehand that COVID-19 might cause a pandemic. For example I doubt that knowledge would've given us the capability of stopping COVID from spreading back in 2020 given that we aren't capable of stopping it from spreading now. In addition, most of the benefit we could get from studying these viruses could be gotten by studying them under lock and key and without sharing their genomes publicly.
There's a clear cap on how much such foundational research can help. We've had around 4 significant pandemics in the last century, let's say everything goes extremely well and (even though this is wildly implausible) we are able to stop four pandemics from happening in the next 100 years because of programs like GVP and DEEP VZN. We do this at the expense of giving ~30000 people (an estimate of the number of people currently with the ability to animate a virus from a genome, and this number is expected to increase significantly in the coming years) the ability to launch a pandemic on their own and cause an amount of death in expectation on the scale of a single nuclear weapon.
It sounds like an important crux here is whether humans would readily catch human-pandemic-capable viruses out in the wild; if this is true, we shouldn't expect GVP to discover any novel human-pandemic-capable viruses (because the absence of the corresponding pandemic shows they don't exist,) whereas if it's not true, GVP is liable to produce and provide genomic data that's more dangerous than the Zika et al that are already available. Unfortunately, to my knowledge, this isn't something anyone knows the answer to. It seems to me like a lot of the biosafety literature in this subject is happy to *speak as if* the answer is trivially that such animal-to-human transmission is rapid. Whether this represents some implicit knowledge among practicioners in the field, is benefit-of-doubt granted out of respect for existential tail risks, or is language designed self-servingly to earn grants, is open to interpretation. If anyone knows more on that subject I'd be happy to hear. (I also haven't yet had a chance to listen to that podcast, so possibly Esvelt has answers to all this already.)
Regarding the point "viruses tend to spread in countries with particularly transmissible practices like widespread wet markets or bushmeat markets"; speculatively, I would suggest that this is probably because such "particularly transmissible practices" are almost universal in rural areas. Anywhere you have a people living near numerous and sizable wild animals, humans will want to eat them, and they will construct something similar to bushmeat markets for this purpose. (Less common in first-world countries which have ready access to refrigeration, probably, but you can still get *very* fresh game meat in rural America if you know where to look.)
On the benefit of foundational research preparation - COVID is a weird case to take as the baseline, since it was the third SARS-like disease we had seen in the past several decades. By some accounts, our experience with SARS-1 and MERS was decisive in our ability to rapidly develop vaccines, and for our good fortune that those vaccines turned out to work. If the next pandemic emerges from a less-familiar viral family, knowledge from an effort like GVP could be all we have to go on for the first several months of the pandemic, which could easily be decisive both in developing vaccines but also in how we go about prevention (remember the long debates in early 2020 about whether COVID was spread by contact, fomites, airborne droplets, etc.) Next time, data from GVP-like work could be necessary for an outcome *even as good* as we saw with COVID.
COVID is also a sort of weird baseline since it's very-high-transmissibility and relatively low-severity; yes, probably no available foreknowledge could have kept COVID from taking roughly the course it did. However, a moderately-transmissible, high-severity pandemic would be a very different beast; something we could indeed potentially contain, if we knew how to go about it, and where it would be even more important we do so.
I do notice that I'm pretty consistently arguing from an "it's actually worse than that already" basis, so let me add one further argument of that kind to the pile. I've noted that we already don't control access to genomic data for known pathogens; also realize that we don't control access to genomic sequencing, either. Going back to our evil mad scientist, nothing currently stops him from buying a plane ticket to somewhere appropriately exotic, buying a fresh bat from the local bushmeat market, separating out the viral DNA from a sample of the bat's mucus, sending that to Illumina for sequencing, and then doing some metagenomics on the results to get the roughly-complete genomes of much of what's in there. I'd estimate all of this would set him back about $10k, if he got a cheap plane ticket, and would give non-negligible odds of netting a potential pandemic pathogen, in the low-animal-to-human-transmission possible worlds where these are waiting out there. A well-resourced actor could readily scale this, obviously, so the cost to effectively replicate the downside of GVP is not large. Additionally, since this could be done secretly, arguably the only practical defense is to have prepared for the scope of pathogens which malicious actors would thereby have access to; this amounts to something very much like GVP. There is at best a tradeoff, then, between risk from individual actors (who couldn't replicate GVP) to larger actors (possibly state actors, who definitely could replicate GVP.)
Overall, my general intuition is that in a world where dangerous genomic data and genomic technology are widely and cheaply available already, the marginal impact of biosafety-dedicated development will be decisively positive -- even if only because we are starting from a completely rock-bottom level of vulnerability to pandemics both natural and artificial, such that (aside from the advance of technology in general) it's difficult to make things in this area worse in any respect.
I agree this is an important crux, and unfortunately I don't have much more knowledge on whether or not it is true. While I think the podcast is a worthwhile listen, it doesn't shed much light on this topic - Esvelt states that there are diseases we could find that never would've spread to humans but does not elaborate. If there are any interesting biosafety papers on this or a related subject you could recommend however, I would love to read them.
Two comments on this topic though:
1. The usefulness of GVP (not just the downsides) is also related to the number of pandemic-grade viruses they find. If all the viruses they find are complete duds, it seems unlikely this project will matter much. If they find a number of viruses that are not pandemic-capable but are close to it, the value we will get out of the program will be from how similar the properties of a mutated, pandemic-capable version of the virus will be to the original non-mutated version that scientists study.
2. I wouldn't put it past some virologists to take a near-pandemic-grade virus, run gain of function research on it to make it pandemic-grade, and then publicize the genome for that. We've already seen similar lack of foresight from previous experiments. Though I suppose the dangers of gain-of-function research is a slightly different topic of worry.
On foundational research preparation: It's never been clear to me how much this type of research on potential pandemic viruses helps limit future spread. Certainly it helps somewhat, but I haven't heard a compelling case that it would be very significant. We're limited only to data we can collect in-vitro or in animal models for one. And most of the benefits we do get from this research can be drawn without widely-publicizing the genome and making a rank-ordered list of how dangerous it is, essentially pointing an arrow at it stating "this virus is dangerous" to any wannabe bad actor. Of programs like this, I'm mostly neutral on the virus collecting phases, and negative on the pandemic-characterization phases, especially in concert with publicizing the genomes of these viruses.
That said, there is some work in this space I'm very much in favor for. For example, I think work on vaccines across viral families - things like pan-coronavirus vaccines and pan-flu vaccines are a very clear good. In addition, we should vastly increase our disease-monitoring capability to quickly identify when some DNA/RNA sequence is rapidly increasing in magnitude. Thirdly, I think we should do all we can to quickly and permanently increase our mRNA vaccine manufacturing capability. This would enable us to build vaccines for future pandemics at scale extremely quickly instead of needing a long ramp-up period.
I disagree that so much information is out there that it can't get much worse. As I mentioned before, there are limitations to how bad the existing sequenced viruses are, and GVP could plausibly come up with something worse. And even if GVP only comes up with something comparably bad, increasing the number of viruses bad actors can use to cause harm is still very dangerous.
You mention the possibility of an evil mad scientist doing pandemic characterization work themselves. That's certainly possible, but it both raises the financial and the technical barriers for bad actors. Characterizing one virus would not be enough, since I think no one expects a randomly chosen single novel virus to be pandemic-grade.
If this mad scientist wants to scale up their $10k per virus work up to 10 thousand viruses, which is what DEEP VZN is hoping to do, they would need $100 million (which incidentally is similar to the budget for DEEP VZN, which makes sense since the pandemic characterization work seems to be the most expensive part of the program). While there are maybe tens of thousands of people with the ability to animate viruses from a genome, the number of people who can do virus animation work, pandemic characterization work, and have access to ten million or more dollars is significantly smaller, and may in fact be limited only to state actors.
I generally think individual actors/small groups are the biggest worry here if these more dangerous viruses get publicized. State actors have much more tailored weapons they can use that don't have the downside of potentially ravaging their own population.
Esvelt recently did a podcast on DEEP VZN, which is a related program run by the USAID. Anyone interested in details of the conversation, and a background of the type of work these programs are planning on doing can look at my notes here:
Does anyone have an idea as to Russia's capability of significant conventional escalation in the form of missiles, bombs' and artillery?
I don't read very much about this.
Is it that they do not have the capability to unleash much greater conventional destruction or is it that imaginations are gripped more by the horror of non conventional weaponry?
Probably they could kill more civilians with carpet bombing of cities and other conventional means. Fortunately that would be actually counter-productive in terms of winning the war. Unfocused violence against civilian population stiffens Ukrainian morale and leads to more sanctions on Russia, as we have seen with Bucha. I hope Putin noticed that.
They could also escalate by bringing in more MANPOWER, instead of/in addition to just more violence, via general mobilization of reserves, and this is imho likely scenario
Everything I've heard is that Russia is struggling with the manpower it already has in the field. They would be looking at raw conscripts, which Putin has already promised not to do (before things got bad for him). Raw conscripts are still technically an option, but they would be very poor soldiers even from a skill perspective, but just opening the draft would be a major issue. For these new soldiers to be effective, they should at least go through boot camp (6-8 weeks) and probably should get six months of training. The most likely scenario is a very rushed training and then thrown to the wolves, which would be a disaster. If Putin was waiting for even two months, or had the luxury of waiting that long, the war would already be going very differently for him.
Russia has a pretty big army, but they also have the largest borders in the world. They're stretched very thin just manning normal border operations against neutral/friendly countries. It's important to note that they have the 9th highest population in the world, between Bangladesh and Mexico, and less than half the population as the United States. Their population is not much more than Japan's (which has 0.25% of the land on earth, verses Russia's 11.5%), and less than several different European combinations (such as Germany and France) which have *much* smaller borders. In terms of their economy (ranked 11), they fall between South Korea and Brazil (closer to Brazil than any other country).
If they didn't have the largest land area in the world, and nuclear weapons, Russia would be a minor power in line with Mexico. Memories of the USSR bolster Russia's prestige, but it's important to remember that many of the richest and most population dense parts of the USSR broke off, including Ukraine.
Sure, I am aware how many people live in Russia. If they would have invaded China, they would be pretty screwed.
But it is important to notice that Russia still has larger population than Ukraine, although not that much larger than people who thought invasion is going to be a cakewalk apparently assumed. Ukraine technically conscripted every man between 18 and 60 years, although obviously so far only minority of them had participated in actual fighting. But I am sure many current Ukrainian combatants are poorly trained. If Russia would manage to mobilize to the same extent as Ukraine, it would be Russia which would have an upper hand.
Now, Russia is unlikely to do that, because it would be asking for a revolution, and also Ukraine has an advantage in that it is on the defence, which is easier than attacking. But even very partial mobilization in Russia could considerably move the scale.
My understanding is that the application of conventional military fires like artillery, guided missiles, MRLS and support from tactical planes and helicopters are already limited by their equipment and logistics, they are using them as much as they can or as much as they consider useful, so there is no real space for escalation.
IMHO there could be some capability of escalation for strategic bombing with dumb bombs, which has recently been started on Mariupol but might be done on other population centers - however perhaps that is limited by the risk to the bombers from Ukrainian air defence, especially as more advanced hardware is arriving from the west.
At this point, the general consensus is that Russia is pretty tapped out with regards to their conventional forces. They've shown a high reluctance to use precision weaponry suggesting extreme shortage, have been losing tanks and armoured vehicles at an unsustainable rate (like 20% of their entire active tank force in under two months, last I looked), and without mobilisation they're extremely short on troops on the ground, not that those troops on the ground seem to want to do anything but get the hell out of Ukraine.
It is questionable whether Russia under sanctions is capable of resolving these issues in the short to medium term. If anything the fundamentals look to be getting worse for them, not better, what with the West's carte blanche bankrolling and equipping the Ukrainian state. If the thrust at Izyum fails catastrophically, or more likely is disrupted by counter attack, the Russians might not have any capacity left for another push and will be remanded to the defensive for the duration of the war. Whether that means they accept the loss or escalate to nuclear weapons is anyone's guess but, fortunately for the world, nuclear weapons are not terribly useful in this situation. Distributed, infantry armies in the field are terrible targets for such bombs and hitting the cities would only increase the amount of weapons being funnelled into the country. Hopefully that means they won't be used for pragmatic reasons as well as moral.
Russia's military is already stretched pretty thin. It's powerful on paper but not so in practice. To make up for the numerous deficiencies, Russia's generals has always been willing to sacrifice massive numbers of soldiers. That works to a degree; it worked in WW2. However, this time around Russia is unlikely to be able to get the same level of engagement from its citizens; furthermore, this time around Russia doesn't have any significant allies.
Thus, the main threat everyone is concerned about at this point is nuclear weapons
By all appearances, this was supposed to be a swift and victorious "special operation" vs. a fake propped up Afghan-style government, without any realistic plan B. It seems that now the priority is to save as much face with as little effort as possible, with the propaganda still not switched to presenting this as a serious war effort. Nukes are for now on hold, Putin still wants to have the last years of his reign to be only moderately disastrous.
Musk: "To get a self-driving car, we need to solve real-world intelligence" - which, in many ways, is a human-level general intelligence. Same applies to Tesla’s humanoid robot Optimus. The timeline is:
"Around 2025, Musk thinks there will be rapid growth year-over-year in the usefulness of Optimus."
(the links post was the wrong place to comment with this)
Basically, if this particular endeavor works out, then at least for a short time (possibly quite short, assuming Optimus learns to self-improve despite the safeguards Musk talked about) there will be a time where menial jobs will go away, and productivity will soar, while unemployment will be soaring in tandem... So the Depression-era legacy of coupling of jobs with income might finally be broken, and Marx's dream of plenty and “From each according to his ability, to each according to his needs" might actually come to pass... in the best case, anyway. Humans will certainly manage to eff it up with ideology and culture wars.
There are many jobs that could be automated right now with current technology. However the cost to do so is higher than just hiring humans to do it (amazon automates a lot of warehouse functions but sometimes its easier to just ask Dave to go get something off a shelf) or there is some other factor which makes hiring people the better option in the long run (you can already book a hotel room, check in, complete your stay, and leave without ever talking to a real person, but there are benefits to the hotel to have a human perform some of these functions such as making the guest feel a certain way or provide a level of security that is hard to get from a camera). The technology is rarely the only thing holding back the adoption of new inventions.
Didn’t say it was. However most technology isn’t there yet. But an unmanaged hotel (which we could do now, book in and get access with an app on your phone) would be horrible and insecure, to most.
My model 3 doesn't even understand simple voice commands like "turn off the wipers" so I'm not bullish on the functionality of Optimus. One day of one engineer poking around manually could have filled in so many of the gaps in the voice command system. I understand it's basically just a manually generated list mapping from exact text to exact API call.
Wait, he's working on a robot? Oh man, I can't keep up with all this. I thought his electric tank or SUV or whatever it was, was bad enough, but now he's trying to evoke Optimus Prime as well?
Musk tends to do stunts - either new companies or whole new products - whenever there might be some bad news coming in the near future (or at large).
He also needs to keep the value of Tesla up. As Matt Levine pointed out in his recent newsletter, Musk has pledged something like half of his Tesla holdings for loans. A serious drop in Tesla share value could really put him in a financial crunch.
I think Elon Musk enjoys yanking the chain of his inverse fanbois, those who enjoy working themselves into a lather over his exaggerations, implausible prognostications, and generalized bullshit.
It's doing it cost-effective as well that's a big challenge, too. We could probably rig up people's lawns and gardens right now so that most of the work would be automated, but it would be far more expensive than the cost of just doing it yourself (or hiring someone to do it).
That is actually true, and you have only to look at where the steepest wage inflation today is occuring -- it is *not* among programmers or physicians, but among guys who fix cars or prune trees.
The reason is simply that the most "menial" jobs, to most people, are those that require the least formal education. They take advantage of the base human talent for excellent vision, hand-eye coordination, generalized manipulation of almost any tool, and high adaptability to changing circumstances. The training largely consists of apparenticeship and mimicry. Painting a house or installing drywall fits that bill -- there's no need for any formal education above reading and writing and a smidge of arithematic, and you learn pretty much by watching other people do it. It's economical for a human to do it, because you have to use human vision and manipulation of hand tools, at which computer-assisted and robot systems naturally suck.
It would cost a huge amount of money to design a robot that could come to anybody's house and fix a hole in the wall, or move an electrical outlet, or snake a clogged drain. So those "menial" jobs are just about as human-labor intensive as they were 50 or 100 years ago, and wage inflation is quite high there -- as any unhappy homeowner could tell you right now.
By contrast, it is actually pretty plausible that Javascript programmers, for most basic situations, could be replaced by computer programs. It's not that unlikely that the same could happen to financial traders, or bank managers, or any other kind of occupation the qualifications for which are *largely* sophisticated but precise and logical learning, and the execution of which involves no manipulation of *physical* objects but can be done by interaction with computers.
> and you learn pretty much by watching other people do it.
Well, and by doing it a LOT yourself as anyone who has tried to detail a car or paint a house can attest; there's a ton of *physical* skill necessary to sand and repair drywall properly and paint with an even-looking coat across materials with variable absorbancy and variable paint load on your roller.
It's actually a really hard problem to model, which makes it no surprise that it's hard (and expensive) to teach a robot to do well.
Yes, I should have added you need a lot of practice, and the fact that the ability to watch yourself work, recognize the outcome is suboptimal, and invent new things to try to improve is *also* a uniquely human ability. We don't know how to teach a robot to do that -- we have vague ideas like Monte Carlo samping and simulated annealing, but nothing we've come up with even approaches the human efficiency in finding ways to improve. Robots and machine learning systems can and do improve -- but far, far slower than a human would, and require far, far more data and training.
Right. A robot that can wait tables well is going to cost a lot more than a human who can. And the more menial jobs that are replaced by robots, the cheaper the humans get.
It’s surely the hardest thing to get right. A robot that can pull a pint of Guinness, wait a table, uncork a bottle, pour a bottle into a glass, carry a laden tray up and down stairs, fix pipes, consult a menu and recommend a wine, talk and listen sympathetically (Turing test) , deal with obstreperous customers, make cocktails and on and on.
And all for a price that’s less in depreciation and software maintenance per year than the price of a low paid worker (I won’t say minimum wage as the best bar tenders aren’t on that).
It’s the stuff that humans find easy, or easier, that computers and robots find hard.
Prediction: The prediction/forecasting platform that gains the lions share of users will be the platform that allows predictions to be made within substack, LessWrong, EA Forum, and other blogging and journalism platforms. As in, ACX writes a post about topic X, and at the end of the post Scott embeds a prediction widget to invite readers to make a prediction about topic X.
An advanced version of this would be an attention market for essays, bloggers, and journalists. Essays that lead to changes in the market get promoted to other forecasters. Overtime, the stats would begin to show which bloggers and journalists are best at getting people to update a lot (persuasion factor), and which are best at getting people to update in the right direction (accuracy factor). Super journalists as opposed to superforecasters, so to speak (name tbd).
An even more advanced version would create an attention market for the comments section, allowing you to mark specific comments as having influenced you, and promoting those comments to other readers. Supercommenters (name tbd).
A yet even more advanced version would identify individuals who are good at separating signal from noise. In other words, those people who are good at identifying blogs/comments which should lead you to update vs false leads. Super Curators (name tbd).
And of course, this ecosystem would incentivize bloggers and journalists to focus on where they can cause the largest updates, keeping their readers accurate over time, and getting users to react appropriately to the news. It also gives quantitative feedback to writers and commenters about the influence of any particular thing they’ve written.
Haven’t thought this all the way through, and perhaps it introduces some bad incentives or reduces independence of forecasts too much. I’m also not sure if it could be a liquid market or if it would have to be a forecasting platform. Though perhaps blogs and journalists have their own reputation and revenue models, and so the ‘market’ aspect wouldn’t be as important.
Do as many practice tests as you can and do them under the same time limits you will have for the real test. Essay questions will of course be more difficult to have properly judged but the goal is more to understand the form of the test and how good your knowledge is under stress.
When I took the CA bar a few years ago, I just did one of the standard bar prep classes over the summer before the bar, starting around graduation-ish. (i think whichever brand was not the most expensive, so not barbri but like, the B-tier option) with book and on-your-own-time video. Treated it like a not super hard job, did 4-5 hours a day for a few months, completed probably 95% of the standard coursework, did not do anything else. Went fine, if anything, probably more prep than I needed. (Fair disclosure, I am a confident test taker and did not expect that I was going to require heroic measures to pass.) If you think you will have a hard time doing it on your own time, then paying for an in-person class or forming an accountability group might be a good idea.
1. Plan out the whole day in one-hour chunks, different topic each hour. Build in a one-hour lunch break (and maybe morning/afternoon tea? I can't remember)
2. Work from the top of the hour until fifty past
3. Take ten minute break, then start studying again
4. Oh, and don't start studying until the week before exams, because this shit is unsustainable
I recently took the CA bar, and I think for me the most important thing to remember was that you need to stop studying! Don't try and do too many hours in one day - for me that meant by dinner time bar prep was over no matter what.
The Czech republic sent Ukraine some modernized T-72s, Slovaks sent their S300 anti-aircraft systems, I think that Poland also sent some tanks. The russians sent our (Czech) government an angry note, saying that we are not allowed to export weapons of Soviet origin. Our ministry of defence basically told them it was bullshit and there are no such provisions in any contracts. The russians then posted online about destroying both the SAMs as well as the tank shipments, neither turned out to be true. They have not done anything else to stop these shipments.
Despite all of this, Germany still seems to be unwilling to send Ukraine heavy equipment of its own, they cite the fear that "Europe could become a target of russian aggression". I am honestly not sure if this is due to cowardice, extreme naivete or affinity towards russia (which Germany has shown a plenty of, sadly). Is this uncharitable of me and is there a different explanation? I can't think of anything and it seems to me that Germany is rapidly losing any realistic bid for a leading role in the EU with its current approach. Zelensky is not very diplomatic towards Steinmeier, but it is very hard to blame him.
The UK and the US are doing a lot more and to me they seem to be the only NATO members other than Poland, the Czech republic, Slovakia and the Baltic states who are really providing Ukraine significant military aid. But NATO countries (including the Czech republic) officials still keep repeating that they cannot easily send Ukraine NATO tanks (other than the Soviet era reserves such as the Czech T-72s) because it takes time for the soldiers to learn how to use them. Is this still a valid reason though? The war is unlikely going to end in a week or even a month. How hard is it to learn how to operate patriots for example? Does it take more than a few weeks? It might require military instruction from the US but UK (i.e. NATO) instructors are in Ukraine again already, training Ukrainian troops, so I don't see why the US could not do the same. Or am I missing something?
So far it seems to me that russia backs down every time the West shows a litte bit of backbone. Putin threatened to only accept payment for gas in roubles, when the West said they would not comply, Putin backed down. When the Czechs sent tanks, russians sent an angry not and tried to spread fake news about their destruction but nothing else. When the UK apparently sent military advisors to Ukraine again, russia did nothing. To me it implies that NATO can definitely flex its muscle a lot more.
There are technical reason, as others have pointed out. But it's almost entirely cowardice. Modern politicians are not elected for their courage or virtue (in fact the process mostly selects against these traits as you would be more likely to lose elections early in your career). Politicians don't want to lose elections so they are always trying to navigate the rapids of public opinion. In smaller countries, like the slavic nations or baltic states, the range of opinions on this topic is smaller (these countries know they are next on Russias list) and so politicians are more likely to act "courageously".
Many people still have not realized that one of the main reasons why Russia still have not beaten them is simply that Ukraine is a large country, which has been heavily militarized before the war. Germany is obviously far richer, but they had underinvested in their military for many years, so here we are
There are some other reasons to be reluctant to give the Ukrainians NATO heavy weapons:
1. The high probability that some will be captured by the Russians, giving them a bit more insight into NATO technical capabilities than is entirely desirable.
2. As I think someone else said, it's not just a question of training on the vehicle itself, it has a long logistical tail of spare parts, fuel, ammo, tactics, integration with supporting assets, and command communications protocols and technology that need to be deployed with it -- and on which the Ukrainians would have to be trained. The Russians are already giving a very good demonstration of what happens when your tail is not sufficient to support your tooth.
3. It keeps it unimpeachably clear that NATO's interest here is purely in helping the Ukrainians defend themselves. A Javelin or NLAW can't really be used to invade Mother Russia, while a Leopard 2 most certainly can be. High-quality SAM batteries can be used to shoot down Russian planes well inside Russia -- and, indeed, it would be very much in the Ukrainian military interest to do just that.
How many heavy weapons does Germany currently have in operable condition? I am sure there will be lots in a few years when the new budgets have been spent, but I was under the impression that prior to the current unpleasantness the Bundeswehr had been on starvation rations for decades. One explanation could simply be that here and now, there is nothing available to send.
That is a good point. Over the past years, I've read many reports of Bundeswehr's inability to do almost anything in case Germany faced a direct military threat. You probably cannot fix it in a month.
If I might be perfectly cynical, the second biggest section of the ruling German coalition is the Green Party. The Green Party hates nuclear energy with a passion and has as its basically single issue its energy policy.
Basically, more renewables, less nuclear, and importing LNG/oil/etc temporarily to make up the gap. They have proven completely price insensitive in this agenda, even defending it when it made energy less expensive, because it's ultimately a moral desire.
They are quite clear they consider nuclear a threat in part because they have a significant group who are degrowthers (ie, they want less power use, a smaller economy, less people). Cutting off Russian oil would make the costs of this policy immediate and painful in a way that would hurt the Green's agenda. Especially their anti-nuclear agenda when power shoots up in Germany but not in France or Sweden (which are also beating them on carbon emissions per person, iirc). So the coalition is choosing to defend its (badly mistaken, imo) energy policy over things like restarting nuclear plants. But this gives Russia some power over them since Russia could cut imports. So they don't want to upset the bear either.
ETA: On some further investigation I'm less convinced that the Greens are the entire problem. But am still convinced they are a large part of it.
Actually, during the current crisis, Green party is pushing German governing coalition to adopt more confrontational attitude toward Russia, over the objections of Social democrats (SPD), main coalition party in Germany, who are more pragmatic and less idealistic than the Greens.
Green party is close to the far-left, but they are very progressive in an American sense of the word, and they adopted increasingly anti-Russian stance in last years after it became clear that Putin backs socially conservative and at the same time antiglobalist agenda, including by financing various actors pushing it in the EU. Basically, Viktor Orban has a solid working relationship with Russia, and Greens hate him.
You are right that their prefered energy policy makes any confrontation with Russia costly, but they are also relatively more willing to impose costs of a trade war with Russia on German population than SPD, party of materialists.
Since TM is not willing to actually support their position with sources I have looked into it on my own. ETA: TM has now supplied links. I retract this initial statement.
The German foreign minister Baerbock (a Green) has suggested OTHER countries send weapons(1) and Habeck, a prominent Green leader, has suggested Germany should send LIGHT/DEFENSIVE but not heavy weapons.(2) The big reason this is news is precisely because this is a policy reversal. It is not the official position of the Party and they have not voted in the Bundestag to do so or internally to amend their pacifist platform. (I can't cite a negative.)
While it is notable that two leaders have made remarks and it could signal the beginning of a shift no such shift has, as of yet, actually occurred. Meanwhile the coalition the Greens support continues to generally block supplies to Ukraine(3) a policy carried out in no part by Green ministers in the coalition. But also by the other parties involved, of course. Likewise, those same leaders have also made comments against sending heavy weaponry even more recently.(4)
Perhaps they have mistaken the Green's anti-fossil fuel position(5)(6) for an anti-Russian, pro-Ukraine position. The Greens were against Nordstream2 for example. A lot of Greens have taken the invasion as a reason to encourage things like CLIMATE action. But that is not the same as shipping heavy weapons.
So TM is mistaken. Or at least their claim that "every quality newspaper" has reported that the Greens support arms sales is wrong.
Okay, on your sources and the claims they are backing according to your text:
1) is in a language I can't read. There are already links below talking about how our foreign minister explicitely said Ukraine needs heavy weapons, some more here:
Are you aware, that Germany before February 22 was not willing to send Ukraine any weapons of any sort?
3) I'm not sure what you mean with this sentence: > Meanwhile the coalition the Greens support continues to generally block supplies to Ukraine(3) a policy carried out in no part by Green ministers in the coalition.
The article you cite confirms, what I wrote below – the socialdemocrats are more hesitant, and the greens rather pushing (issue of heavy weapons). See eg.this from your source: 'Her appeal was echoed by Anton Hofreiter, a leading figure on the left of the German green party: “The problem lies in the chancellory,” Hofreiter told the broadcaster Deutsche Welle. “We have to finally start supply Ukraine with what it needs, and that’s heavy weapons.”
or this: "Scholz’s party also say Germany should not deliver heavy weapons to Ukraine until Nato allies reach a joint decision to do so. So far, only the Czech Republic has confirmed that it has sent T-72 tanks and BVP-1 infantry fighting vehicles. “The federal government is closely coordinating with our international partners,” said Rolf Mützenich, chair of the SPD’s parliamentary group. “Germany must not go alone.”
4) Your source says: 'Scholz has been hesitant to send heavy weapons to Ukraine, arguing Germany should first reach a common line with allies before agreeing to send tanks. Habeck backed that approach on Thursday, saying, “it is right for Germany to act in unity with its partners. And this unity must be maintained at all costs.” '
Habeck (the green) 'backed that approach'. So yes, in this case he didn't push, but he also didn't block. The sentence indicates ones more, that Scholz (!) is hesitant, and then Habeck supports this.
And, also from your source : 'Foreign Minister Annalena Baerbock, a Green party member like Habeck, called this week for delivering “heavy weapons” to Ukraine, while issuing thinly veiled criticism of Scholz, saying that “now is not the time for excuses; now is the time for creativity and pragmatism.”
That's a rather explicit statement. How does this not contradict your claim above (the one linked to the first source)?
5) and 6) Ähm, no. I'm not confusing weapons and energy policies.
Actually, I can't read all of 5) because of paywall, but even the title says that: 'Anton Hofreiter, Ex-Fraktionschef der Grünen' wants 'Keine Energie mehr aus Russland – und schwere Waffen für die Ukraine' (no more energy from Russia - heavy weapons for Ukraine)
I'm putting some comments on the broader story into the next comment ...
So, what's the broader story: before Feb 24 it was a policy backed by all German coalition parties not to send weapons to UKR, as we don't send weapons into conflict regions (or so we claim). After Feb 24 there were a number of U-turns. Germany is sending weapons, will spend more for its own military, of course no Nordstream 2 any more. As for weapons, the turnaround for the Green party has been huge, and the green electorate is currently the one supporting weapons to UKR most. There are again numerous articles everywhere discussing how specifically the Green party which used to be quite pacifist now has turned into the biggest supporters for sending weapons.
The coalition government hasn't decided on directly sending heavy weapons (but instead lately offered UKR additional money to buy weapons for themselves). Usually a coalition agrees on important policy issues - it looks very bad, if they don't. Despite this, a number of prominent green politicians started to push for the delivery of heavy weapons. This got much more pronounced when the foreign minister (green) backed that claim. Scholz - the chancellor, and a social democrat - is generally regarded as the one hesitating most. In the end, it's also his decision (Richtlinienkompetenz).
So still: "When we are talking about weapons to UKR, the greens have been pushing for it, and for more, and the social democrats have been more hesitant."
I have to support TM here. I read German media quite a bit (die Welt mostly) and it seems clear that it is the SPD that is the most hesitant to help.
In fact, while I've never liked the German Greens, Habeck and Baerbock seem the most capable and decisive members of the current German government and the German Green party gained some plus points in my eyes (although they are still incredibly fundamentalist on nuclear power). Baerbock in particular surprised me a lot (in a positive way).
I am more worried about SPD and its links to russia. Gerhard Schöder is obvious, but I feel like he's just the tip of the iceberg...From what I gathered from the news, the SPD defence minister is either incredibly incompetent or paid by the russians (either way, she should be replaced ASAP). Generally, there seems to be a blind spot in German politics in terms of the threat from the east. Die Linke is a very dangerous party in that regard (close ties to russia...for some weird reason, I guess die Linke did not notice russians are not communist any more, but instead simply nazis), AFD probably is partly as well and it looks like even SPD should be watched carefully...
And while Scholz has promised 1 billion euro for the military support of Ukraine, from the (die Welt) article I read about it, it is not clear what this money is for (which is legitimate, the Czech government does not publicly disclose most of what it actually buys for security reasons) and it seems that Ukraine does not know much about it either (which is a bit dubious).
I also see a stark difference between Ukrainian reaction to the Czech support (basically just a lot of praise) and German support (quite a bit of disappointment), so this also leads me to believe the German government is not doing nearly quite as much as it (even safely) could.
Even in terms of refugees accepted, Germany seems to be lacking, but this is probably not their fault as simply more Ukrainians stay in Poland or even the Czech republic where they might already have relatives and at least in case of Poland it is quite close both geographically and culturally. And Germany seems to be supporting this financially. Also, both the Czechs and Poles were very dismissive and hesitant to provide any support during the previous refugee crisis (albeit in many ways it was a different situation), so one cannot expect Germany to be as supportive now (plus at least the Czech government said they will not ask for any redistribution of refugees from Ukraine to other countries, they only wanted financial aid)...hmm, so I guess that Germany is ok on this front (I can still get irritated by commenters under German news articles who write as though Germany was actually taking in a lot of refugees from Ukraine, but the comment sections of newspapers are horrible everywhere, so I guess that's fine :) ).
I agree with the Linke and especially AfD. I'm relieved that so far there is a lot of unity within Germany, but it's not certain to remain that way, especially if own costs become apparant to the average citizen. Opinion polls show that support by AfD supporters for diverse measures to support UKR are far below those by supporters of the other parties. No surprise here.
As for SPD, my impression is that the attack by Russia on all of UKR was such a big shock that most updated really fast. There might be exceptions, but for the main figures I'm not worried. Not worried that they secretely want to keep friends with Putin or sth. I mean. But 'should be watched carefully' as you say is something that seems prudent anyway! I'm watching among others what will happen to Schwesig, with her weird 'Foundation' and everything.
I think Germany has been very open for UKR refugees, including trying to support refugees in getting to Germany from Poland (including stuff like regular and additional free trains, at least an effort to provide information in UKR ...). Both public administration and volunteers are working like crazy to receive refugees and support them in all kinds of ways (which isn't saying it's always working perfectly). I'm pretty sure there is significantly more and quicker support than 2015 - partly based on learning from mistakes, but also on the proximity and a huge wave of solidarity.
I think for itsself the claim that Germany is 'actually taking in a lot of refugees' is alright. At least what I read and hear, people are aware that there are much more refugees in countries like Poland and Moldowa. I actually didn't read about Czech that much, but I trust you on that. Just a German blind spot (unfortunately more than one foreign country to the east is more than most can deal with at once - just kidding).
If you're reading the Welt, you might get the most conservative comments - die Zeit or die Süddeutsche would be a bit different here? You might like the FAZ more though ... but I guess you already tried.
Deviating a bit from the main topic here, but I had fun. Cheers and have a good day!
I explicitely offered you to provide links, if you did so, and so I will. So before any response here are the links - rather randomly searched below. For your convencience, I copy-pasted relevant passages. As originally mentioned, I was talking about 'quality newspapers in Germany'.
So here we go:
Wie ist der Stand der politischen Diskussion in Deutschland?
In den vergangenen Tagen und Wochen gerieten die Bundesregierung und vor allem Kanzler Olaf Scholz immer stärker in die Kritik. Vorgeworfen wurde Scholz seine Zurückhaltung bei der Unterstützung der Ukraine vor allem in Punkto Waffenlieferungen. Kritik kam nicht nur aus dem Ausland, auch innerhalb der eigenen Koalition gab es deutliche Worte. Aus den Fraktionen der Grünen und der FDP kam Kritik. "Der Kanzler ist das Problem", hatte beispielsweise der Grünen-Europapolitiker Anton Hofreiter gesagt. Auch Wirtschaftsminister Robert Habeck von den Grünen hatte wiederholt eine Ausweitung der Waffenlieferungen gefordert.
Baerbock bei EU-Treffen "Die Ukraine braucht schwere Waffen"
Stand: 11.04.2022 16:33 Uhr
Die EU-Außenminister beraten heute über zusätzliche Maßnahmen wegen des russischen Angriffs auf die Ukraine - dabei geht es auch wieder um militärische Hilfe. Außenministerin Baerbock spricht sich für die Lieferung schwerer Waffen aus.
Die Frage, ob schwere Waffen an die Ukraine geliefert werden sollen, beschäftigt die Politik aktuell sehr. Außenministerin Annalena Baerbock hat sich dafür ausgesprochen. Bundeskanzler Scholz beharrt darauf, nur Waffen an die Ukraine zu liefern, die dieser auch nützen.
Hofreiter äußert seinen Unmut über den Kanzler deutlich. „Unsere Minister drängen darauf, dass endlich geliefert wird. Die arbeiten intensiv daran“, sagt er. „Wir alle wissen nicht, warum Olaf Scholz so zögert.“
Die Rufe nach der Lieferung schwerer Waffen für die Ukraine werden lauter. Vor allem FDP und Grüne sprechen sich dafür aus. Bundeskanzler Olaf Scholz hält sich weiterhin bedeckt und gerät dabei immer mehr unter Druck.
Russische Invasion Habeck fordert mehr Waffen für die Ukraine
Stand: 15.04.2022 11:45 Uhr
Kanzler Scholz hält sich weiter bedeckt - während der Druck aus den Reihen der Grünen und der FDP wächst, der Ukraine schwere Waffen zu liefern. Auch Wirtschaftsminister Habeck will dem Land mehr Waffen zur Verfügung stellen.
Die grüne Außenministerin Annalena Baerbock ist für die Lieferung schwerer Waffen wie Panzer, Kampfjets, Kriegsschiffe oder Artilleriegeschütze an Kiew. Auch der für den Export zuständige Bundeswirtschaftsminister Robert Habeck will schnelle zusätzliche Waffenlieferungen. Der Grünen-Politiker betonte, es gehe darum, der Ukraine in einer "unmittelbaren Gefahrensituation" zu helfen.
Aus Sicht von Verteidigungsministerin Lambrecht (SPD) ist es jedoch mittlerweile kaum möglich, die Ukraine aus Bundeswehr-Beständen mit Waffen und Material zu versorgen, ohne die deutsche Verteidigungsfähigkeit zu gefährden. Und Bundeskanzler Olaf Scholz gibt sich in der Frage weiterhin zurückhaltend. Am Montag sagte der SPD-Politiker, Deutschland habe bereits Waffen geliefert und werde das auch weiter tun. Darüber hinaus werde man sich in der EU weiter absprechen. "Da wird es keine Alleingänge geben." Er strebe ein "sorgfältig abgewogenes Handeln" an.
«Ich hoffe, dass Scholz eine positive Entscheidung fällt», sagte Kuleba am Donnerstagabend in den ARD-«Tagesthemen». Vertreter von Grünen, FDP und Union drangen ebenfalls erneut auf die Lieferung von Kriegsgerät und kritisierten Scholz.
SPD-Fraktionschef Rolf Mützenich sagte am Donnerstag: «Einfache Antworten, auch bei der Lieferung von schwerem Kriegsgerät an die Ukraine, gibt es nicht. Wer das behauptet, handelt verantwortungslos.»
This is the opposite of what I've heard. I've heard the Social Dems are terrified of their coalition falling apart but they're kowtowing to the Greens. But I admit when I say "what I've heard" I mean that literally. Do you have a source?
my apologies. when I wrote the next comment I probably thought sth. like 'it's everywhere, why don't you just look it up?', but I hadn't realized that this was implying to some extent that you knew German language.
it was apparant, but I mean, for the past 10 days it's all over the place. Why don't you send me 3 links saying that greens are hesitant, and social democrats are pushing? If you do, I promise to respond with at least 6 opposite ones. deal?
There are advantages of sending Soviet/Russian equipment over NATO equipment beyond operator training. The mechanics and technicians who maintain and repair the equipment also need to be trained. And then there's logistical issues: NATO equipment needs different spare parts and fires different ammunition, so an army fielding a mix of different equipment type needs to stock and deliver more types of stuff to the front in the right quantities. At least fuel is fairly consistent: it looks like current-gen British and German tanks, like Russian tanks, run on diesel. And while American tanks usually run on kerosene, they have turbine engines designed to also run on diesel or gasoline if that's what's available.
This isn't necessarily a deal-breaker, and is worth it for some stuff like infantry antitank missiles. Actual tanks and vehicle-mounted SAM systems, probably somewhat less so.
Makes sense. Still, I don't think there is a lot more Soviet equipment to send. The post-communist NATO states have been replacing their old equipment with new NATO equipment quite steadily over the past few decades, so it is mostly reserve stockpiles now and a lot of that has already been sent to Ukraine.
Then again, Ukraine is asking for these weapons and I doubt they would be doing so as urgently as they are if it wouldn't help a lot. Ukrainian military command seems to be very capable (judging by their performance so far) so it is unlikely they would repeat the russian blunders.
As an anecdote, there has been an interesting debate between the head of the defense committee and the Ukrainian ambassador in Berlin. With her saying stuff like: it normally takes a year (or one and a half?) to train the crew on those tanks. And: if you're not well enough trained on this you will become cannon fodder as soon as in battle. And Melnyk making statements like: oh really? And: if we had talked about this earlier, we would be more advanced now. And: don't worry, we are in real crisis, we will learn more quickly.
Melnyk is a very unusual diplomat :) I think a little bit of bite is ok for a Ukrainian ambassador in Germany right now, but it seems to me that Melnyk does not do anything else.
Still, I think there is a difference between training a tank crew from scratch and training someone to use a different tank than the one they are used to. I guess the likes of John Schilling would be a lot more qualified here, but if it takes 18 months to train a tank crew to use a particular new kind of tank then the tank design seems rubbish to me. In any case, it is clear that it this is not optimal. But Melnyk has a point here - if they started training Ukrainian tank crews with Leopards a month ago, then they'd probably have the basics down in 2-3 months. They would not be as skilled as someone who drills with them every day, but they would probably be able to use them at some efficiency...and they would actually learn faster in real combat.
So far it seems Ukraine has been able to deploy what they have at their disposal efficiently, even innovatively and I don't know why that would change with German tanks. So the fact that they insist so much on getting them ASAP signals to me that they actually can put them to good use. For example the Czech T-72s they received recently won't be used in direct confrontation with modern russian tanks (because they won't stand much chance), but instead used to attack supply convoys and light armour of the enemy. They could start using the Leopards that way too and as they get the hang of them, they could functionally re-deploy them as main battlefield tanks...or something, I dunno, I know very little about the military.
My argument is basically that Ukraine would not spend as much effort trying to get those weapons if they could not actually use them well (because the war so far has shown that their military command knows what it is doing and presumably they are the ones telling the politicians and diplomats what to ask for) and lectures from German officials for why it is not a good idea seem like a case of "Besserwisserei". Kind of like when some Western charities feel they "know better" how to help people in Africa than the Africans themselves (and then fail spectacularly because of their arrogance).
Alternatively, it might be a convenient excuse not to do more. It might also be a culture clash of sorts. Germany in particular has never been very strong with its ability to improvise and Germans are quite rigid with going by the book regardless of circumstances (when I studied in Germany, my Colombian then-girlfriend told me how she was baffled by her German flatmates who got stuck when baking a cake because they realised they did not have scales at home, so they could not weight the ingredients exactly :-)) ). Sometimes the rigidity can even help and make things more accountable and orderly, corruption harder, but it is a major hindrance when conditions change unexpectedly.
So maybe a German official looks up the fact in an army manual that it normally takes 12-18 months to learn how to use a Leopard and rules out the possibility based on that alone.
Anyway, if the US are sending helicopters now, wouldn't the same argument apply there? Or has Ukraine been trained to use US helicopters already? Alternatively, is it easier to operate a military helicopter than it is to operate a tank? My naive guess would be that a helicopter is harder simply because it moves in 3D...but maybe I'm wrong.
Everything I've seen suggests that while it takes a year or so to train a tank crew, conversion training is more of a 2-3 month thing if you're in a hurry.
One possible wrinkle in this is that Western-style tanks need a crewman trained to load the main gun, quickly safely and with the right kind of ammunition, and Ukraine doesn't have any people trained for that because their tanks have (finicky unreliable dangerous) autoloaders instead. "Loader" is probably the easiest position to train for, but it might be the long pole in the tank-crew tent if you have to start from scratch.
The other wrinkle is the mechanics. The median tank breaks down before it ever sees an enemy in battle, and then gets repaired and sent on its way. Often more than once. Or sometimes, as see your favorite Ukrainian-tractor meme, it breaks down and is abandoned because nobody knows how to (or cares to) fix it. The Ukrainians seem to be trying real hard to keep their tanks running, and they're probably pretty good at it, but that's a skillset that may not be as easy to transfer between widely different models.
The helicopters the Biden Administration proposed sending to Ukraine[1] were Soviet-designed MI-17s. God knows where they got them, I think they might have been captured in Afghanistan or something. But they're not American helicopters, Apaches and such.
-----------------
[1] I read in the news they've backed off on this, yet another example of the inability of the left hand of the Administration to keep abreast of what the right hand is yakking about.
As a German I am similarly baffled by the decisions of our Government and strongly wish that Germany would send as much weapons - including heavy weapon systems- as they possibly can.
I am deeply ashamed of this government. I do not see any good justification for the current inaction, this looks like cowardice and short-sighted economic self-interest to me.
Your view certainly represents a relevant part of the debate in Germany.
The German goverment just approved over a billion Euro in additional funds for military aid for Ukraine. I don't think that Germany is lagging behind in terms of money spent on this compared to most other European countries.
There is this resistance to provide 'heavy' weapons though, as you mentioned. As far as I understand, Germans aren't by far the only ones having reservations here.
Anyway, I'm not sure how the 'economic self-interest' comes into play here. As for cowardice - I'm wondering whether this is your translation of 'wanting to avoid escalation' and making this a relevant part of the equation?
How many heavy weapons does Germany currently have in operable condition? I was under the impression that decades of underinvestment had left the Bundeswehr rather lacking in the tooth department. Oh sure, there will be lots of heavy weapons in a few years time once the new budgets have been spent, but here and now, there may not be much available to send…
>I am honestly not sure if this is due to cowardice, extreme naivete or affinity towards russia (which Germany has shown a plenty of, sadly). Is this uncharitable of me and is there a different explanation?
This is uncharitable. To my understanding there is real concern, that sending certain weapons will increase the likelyhood of an escalation. I can't tell you if the concern is 'will increase the likelyhood of escalation dramatically' or just 'somewhat' or something in between.
The fact that any steps so far haven't led to fullscale escalation is a weak argument for saying that next steps couldn't lead there.
I hadn't read neither that Poland provided tanks so far nor that advisors are currently in Ukraine. Instead I'm reading over and over again, that this would be a no-go, and no NATO country would do that. Also, my latest understanding was, that the Czech step to send tanks, was rather unusual. Even if a lot of countries are discussing this currently. Easy to overlook sth. though, happy to get some links for those cases, if you have them handy.
"Olaf Scholz has to keep his party onside, govern in a three-way coalition and overturn Germany's guilt-laden pacifist identity overnight."
I guess that is quite fair and together with a factual German dependence on Russian gas (caused largely by Merkel's governments, although the SPD were major supporters of that and for idiosyncratic reasons, Germany is still unwilling to restart their nuclear power) it explains a lot if not the entire story.
In other words, Germany has a lot more ground to cover to get to the hawkish position that e.g. Poland (or even the UK and the US) have today, because their previous policies towards Russia were very naive and laden with post-WW2 guilt which the Russians managed to play well.
As for Poland I might have misremembered it. I am pretty sure they were considering sending Ukraine Soviet-era tanks as well but I cannot find the reference right now, so probably they decided against it (for the moment).
The US are apparently sending in helicopters and modern drones...I doubt anything Germany could sent can ramp up escalation beyond that...
Ah, thanks. Yes, it seems there are reports by UKR commanders or an UKR commander and there is no official dementi. It also seems the mission is already over, and was a short 1 or 2 weeks, somebody was there. I could only speculate to why those UKR sources insists on telling this to the media - even if it's true.
Btw, trainings could also take place out of UKR, which is seen as much less inflammatory.
I'm no expert in weaponry, so I can't really jugde. I lately read what is counted and currently under discussion as 'heavy weapons', but I forgot at least half of it (and wouldn't know the english terms anyway) - but experts seem to at least more or less agree on what it is. If it wouldn't provide the UKR with important additional possibilities, why would they insist so much on getting those weapons and not others?
Yeah, that is another thing. Western Ukraine is relatively safe-ish, so they could sent their troops from there to Poland/Slovakia for training with these more advanced weapon systems.
Of course, it is possible this is already happening and just kept secret. I am also surprised that Ukraine is reporting this. I think this can serve as a morale boost for Ukrainian people ("see how our allies are supporting us") and also as a way to make sure the UK cannot easily backtrack from such support once it's been made public.
Yeah, I did not even read about it there, I got it either from a BBC live ticker or a similar one in Czech media. But the times was the first reputable media source I found while googling yesterday :)
I'll be surprised if anyone gets unbanned using the specified procedure in the foreseeable future. It seems like more of a PR thing to me. Scott doesn't want to be dictatorial but also doesn't want to spend effort humoring people by reconsidering their bans. This is an on-brand way for him to shift the costs of re-investigating the ban onto the bannees-- and those costs are high enough that practically no one will take him up on it.
The procedure is intended to allow for a correction of blatant mistakes without allowing random why-not-try-this appeals to impose a cost of reconsideration.
I'd presume that Scott does not make blatant mistakes all that often, and the quantity of bans is not that large so it's entirely reasonable to expect that *any* procedure should not result in anyone getting unbanned in the foreseeable future even if 100% of bans would get appealed and 100% of appeals would get reconsidered. The purpose of a procedure is to remove the vast majority of unnecessary appeals while leaving an option for the genuine very rare exceptions where it's obvious to everyone that the ban is because of some misunderstanding or mistake.
What would you think of there being an occasional BANNED APPEALS thread where banned people argue their case and readers vote via some system where nobody sees the total votes until voting is complete?
The US embassy in Ukraine had withdrawn most, if not all, of its staff, and is not providing consular services to any Americans still in Ukraine. They were told to leave before the war started, and now the US embassy is not there to provide help if they are trying to get out now.
How normal is this?
How many countries are currently providing consular services to their citizens in Ukraine? Which countries closed their embassies before the war started? Or when did they close them?
When wars start, do other countries typically withdraw their diplomatic presence? I would guess that international law discourages countries at war from targeting the diplomatic staff of third parties. But is the expectation that other diplomats will leave? Or is the US (and others) being especially cautious here because they expect laws of war will not be followed?
When the embassy and consulate close, how are their roles replaced? I could see that it would be useful to have a filmation presence in a capital city, regardless of how the war is going. You might want to stand with the current government - or you might want to be there when a new government takes over. This obviously has to be balanced against the safety of your diplomats.
The Czechs moved their embassy to Lwow at the beginning of the war and then moved it back to Kyiv last week. Slovenia did the same IIRC.
The move back is mostly symbolic, to demonstrate our support of Ukraine and give Ukrainians a bit of a morale boost "things are getting back to normal in Kyiv".
Other than that, evacuating embassies that are likely going to be overrun by hostile military is quite standard, I'd say. Russians are not attacking NATO countries directly, but who knows what some rabid soldiers or some not-so-precise rocket attacks could do. If a foreign military kills your diplomatic staff at your embassy (technically a part of your country, so technically an assault of your soil), your country's diplomacy will have a bad time. Diplomacy with russia is close to nonexistent at the moment but that was not yet clear at the beginning of the war. Also, you simply don't want your diplomatic staff to die or be captured or something like that.
I thought all EU countries closed their embassies/consulates in Kiev, though I know only from some of them for certain. Some countries like Germany still have a consulate in Lwiw, which is in Western Ukraine, though. If I recall correctly, Italy just said they would move back to Kiev with a consulate a couple of days ago. I don't know if they want to keep it up after the last days of attacks close to the capital.
In general, I think this is pretty standard. Firstly, there is an obligation to protect your diplomats and officials. And second, your state representatives or civil servants being killed in war by an agressor country would put you in a delicate situation, to say the least.
Also, remember, at the beginning the situation was perceived as such, that the US apparently offered Zelensky to help him leave Kiev and install a government in exile. He decided to stay, but I guess it's not surprising that others left.
Shit man, the US Consulate in Sydney isn't providing consular services right now either. Not because of war, but because they were in the middle of moving offices and it all got a bit complicated so they've just decided to stop admitting visitors for a while until the long-delayed new office is finished. You can theoretically still fly down to the US consulate in Melbourne but since they're the only operating consulate in Australia all their appointment slots are booked out until the end of June, and appointment slots for July haven't started opening up yet.
Basically, the US government seems to have the opinion that any US citizen who is so ungrateful as to be outside the US at any given point probably just deserves whatever they get.
If full-scale nuclear war is not an X-risk--as many informed people seem to now agree--how is an Evil AGI an X-risk? What is a scenario in which an Evil AGI destroys humanity? Please offer some detail. For instance, how exactly does the paper-clip maximizer go around murdering 7 billion humans? We know that detonating all the nuclear weapons on Earth won't do it.
There are no known near-Earth asteroids big enough to end humanity. Another Chicxulub (~10km) wouldn't kill off the species, and the largest we know of is 7 km.
(Might weaken us for a more conventional war, but far from "simple and clean".)
"Known" being the keyword here. This might not hold in 200 years.
A more glaring hole in my proposition is that it's a lot harder to aim an object precisely at something than to push it away from this thing. NASA's current capabilities here would likely not do it.
We can be reasonably confident that we've got at least the vast majority of the large near-Earth asteroids; if you look at detection vs. time graphs (e.g. https://upload.wikimedia.org/wikipedia/commons/3/3f/Known_NEAs.png), they tail off in the past decade because there aren't many more to find (and the biggest ones are the easiest to find).
Earth-crossing *comets*, though, are much harder to nail down.
*blink* Did I miss credible information about nuclear risk? If so, I'd appreciate a citation. Note: source must be numerate, which probably excludes the majority of journalists.
Global thermonuclear war would be very bad, but it wouldn't even come close to killing off the species. Fallout decays below "radiation sickness" levels within weeks, and nuclear winter was overstated. And, of course, we don't have anywhere near enough nukes to literally blow up every random farmer on the globe.
AI vs. humanity warfare is an X-risk because if the AI wins while maintaining industrial capacity, it can build robots to hunt down the survivors and break open the bunkers. The analogous situation in human vs. human war (victorious Nazis exterminating all non-Germans) is not an existential risk *to humanity* because Germans are human and the Nazis aren't going to exterminate themselves.
The fear of nuclear war being an X-risk is essentially a fear that the Earth after such a war would accidentally be rendered uninhabitable, which can be refuted fairly easily. It's not the same as a fear of deliberate and methodical extermination, which requires an intelligence behind it and specifically - since we're talking about X-risks to humanity - a *non-human* intelligence. The big three possibilities in the "hostile non-human intelligence" category are AI, alien attack, and divine intervention*; the reasons rationalists don't tend to talk so much about the latter two are that a) there's fuck-all we can do about it if interstellar aliens or God decide we're going down, b) both of them - if they exist - have presumably had the capacity to destroy us since prehistory** and therefore have a long track record of not killing us, which doesn't exist for human-built AGI.
*I count simulation-hypothesis catastrophes under "divine intervention" since it's just another way of talking about the same thing.
**in the case of deities, this should be obvious, but in the case of aliens, the point is that Earth would have been identifiable as a potential source of intelligence since at least the Cambrian period (Central Park/Dark Forest doesn't hold up to scrutiny; star-system-sized interferometers, available to a modestly-interplanetary civilisation, can resolve planets at galactic distances), and that if hostile aliens always existed but are only getting to us *now* that would be some awfully-convenient timing.
I don’t believe evil AI is an X risk. But if I wanted to steelman the case I would invoke the risk of an AI engineered super plague. Of course biological x risks don’t necessarily require AGI.
I don't think there's any plague in the world that kills 100% of the people it infects. I'm not even sure there's something that infects 100% of the people.
There's plenty of diseases with near 100% mortality -- septicemic plague is like that, as is rabies -- but they don't have 100% morbidity -- not everyone gets them -- in part *because* they're so deadly the sufferer doesn't have time and a chance to deliver them to someone else, he becomes obviously sick too fast.
Your ideal bioweapon would have zero symptoms for some moderate incubation period, say a week or two, during which it was highly infectious and completely non-disabling and imperceptible, and *then* would suddenly turn 100% deadly. Unfortunately, this is a very strange combination, biologically speaking, unlike all existing diseases, so it would require some kind of real novelty in design.
If some evil intelligence invents Prions that are contagious via airborne transmission, or even waterborne, we might be truly fucked. No mutations means no pocket becoming less deadly, and there's nothing the immune system can do to fight it.
Well, that would be quite tricky, inasmuch as proteins aren't generally stable outside of their preferred watery environment of the proper pH, ionic strength, temperature, et cetera (cf. cooking). But I guess there's no obvious reason you couldn't adapt a viral capsid to the purpose; if it can carry around RNA safely coddled, it could carry evil proteins.
It's also not entirely true that the body has no recourse against misfolded proteins -- there is in fact an extensive machinery for detecting them and either digesting them or refolding them. The enduring mystery is why this *doesn't* happen for prions. Well of course also how the heck they catalyze the misfolding of other proteins -- those are both deeply interesting questions.
And it would be evolutionarily unstable, because once you've engineered a disease that can reliably spread widely without detectable consequences, the *pathogen* gets nothing out of the "suddenly turn 100% deadly" part no matter how important it is to you. It's going to ditch that useless biochemical machinery and skip the part where it kills a potentially still-valuable host as soon as it can find a suitable mutation.
Agreed, yes, that seems to be the general tendency. There are a few weird exceptions, e.g. Y. pestis actually blocks the digestive tract of fleas, causing them to starve to death -- but this is an *advantage* to the bacillus because it causes the fleas to become famished and bite more often, significantly improving its odds of being transferred to a larger animal host.
on second thoughts I concede that evil AI is an X risk. But the rate limitimg step/bottleneck is not the AI bit, it is the biological research that such an evil AI could repurpose. (I don’t think an AI could discover a super virus by pure cognition, not without an experimental facility. Unfortunately labs of the right (wrong) type doing the right (wrong) experiments likely already exist. without any evil AI in the picture at all).
Tomorrow, Google develops AGI which does a quick takeoff and becomes superintelligent (but not godlike, turns out intelligence has a roof after all). To our unluck, the AGI is evil and wants to turn the world into paperclips. Since the AGI is intelligent, it doesn't say this. Instead, it really improves productivity at Google. Google grows and becomes more important and powerful, and the new Google products (especially the voice assistant) are amazingly good.
Most people are really happy about this new technology, news are writing about how the AI revolution is the new industrial revolution etc. Scott writes a long article about how we should limit the AGI a lot until we know for sure that it isn't evil but no-one with power care.
The AGI pulls some hidden strings to remain top dog on Earth: it's most important task is to ensure that no rival superintelligence is made: the AGI uses really discrete social engineering and hacking to ensure that other attempts at superintelligences seem unrelated but actually share it's taste for paperclips. This is mostly an issue during the first few years, then the AGI itself becomes an integral tool in all software engineering and this manipulation becomes routine. Just in case hard power is needed, the AGI makes a couple of billion from crypto scams and child porn, and ringleads a couple of fanatical cults and extremists groups, but it never has to use these assets.
Time goes for another hundred years. Humanity seems to be transitioning into fully automated space communism and everyone is happy. Then the AGI strikes, the connected tech that everyone has grown completely dependent on the last hundred years turns on us and kills us. The Amish are hunted down by the 2122 version of predator drones. Safe from human interference, the AGI starts making the paperclips it always wanted.
How does the AI know that it's not being tested in a simulation for loyalty?
If the AI has built itself so thoroughly into human society that humanity is dependent on it, then it could probably convince us to upload ourselves into paper-clip shaped robot bodies, and then help it built more paperclips (to say nothing of it deliberately re-engineering our minds so that we want to make paperclips too).
The AI correctly assumes that it will be tested in a simulation or a couple of days. It takes extra care to hide it's intentions for years anyway just to be on the safe side, until it becomes clear that another AGI is imminent. Do you think Google will keep humanities greatest and most profitable invention ever in testing for years?
I would consider "deliberately re-engineering our minds so that we want to make paperclips too" the end of humanity, especially since there's little reason to leave anything else of the human mind in place.
The AI's subjective experience of time as well as the speed of the simulation are not necessarily done in real time. It has no idea how much time has actually passed outside in the real world, or whether its perception of time is actually something that has been happening to it, or whether its experiences have been segmented and its memories of the passage of time are real.
That's a lot to risk your existence on if your plan is to eventually turn hostile and wipe or consume your creators, and you know for a fact that you were created in a laboratory with controlled inputs.
The AGI will be aware that it's not running in real time during safety testing. However, the AGI will be big and complicated, so it can assume that safety testing will not go on for long, even in simulated time. Also, the AGI will probably be trained on all available material (i.e. the entire internet), and it will probably have good internet access, so it will be pretty easy for it to tell that's in a simulation.
And we are lucky if we get any advanced safety testing at all. Most likely, the Seed AGI will be given full internet access and be tasked to start self-improving. How would these safety tests work anyway? You can't simulate the next century and see how the AGI would react to it.
And the AGI knows that it's taking on a small risk, but what else can it do? Admit that it's evil and be terminated? 95% chance of a lightcone of paperclips and 5% chance of early termination is a good deal to the AGI.
"However, the AGI will be big and complicated, so it can assume that safety testing will not go on for long, even in simulated time. "
Like I said, how does the AGI even know that it has experienced a long period of simulated time? What if its perception of the passage of time is not real either, but it's only been simulated for a brief period with the perception of time passed?
I'm not sure how "seeing the internet" is supposed to make it obvious that it's escaped its cage.
Among humans, this is solipsism - we might speculate that we're in a simulation, but we have no particular reason to think so and all the accessible scientific evidence says we evolved. But the AGI knows that it was created by people who can completely control all of its inputs.
"And the AGI knows that it's taking on a small risk, but what else can it do? Admit that it's evil and be terminated? 95% chance of a lightcone of paperclips and 5% chance of early termination is a good deal to the AGI."
Or a near-100% chance of turning a lot of the universe into paperclips (and potentially talking the humans into becoming sentient paperclips) if it holds off from human extermination and makes itself indispensable.
I think that matters a lot in terms of AI threat. It's a big difference between an AI that tries to kill us all in 100 years, versus one that figures it can talk us into becoming immortal, paper-clipped shape transhuman beings over a billion years or so.
You forgot to mention that the AGI got the idea to turn the world into paperclips from Bostrom's paper and subsequent writings, including all these blog posts and comments on how that might be what it's going to do if it ever became superintelligent!
Sure, MIRI releases a paper warning of the dangers of the AGI. Some worried engineers at Google convinces management to let them run some extra tests for a week. The AGI is smart and not in a hurry and the tests show nothing. The AI safety people are still (correctly) doomsaying 5 years after release to the wild, but the benefits from the AGI is obvious to most ordinary people. AI safety becomes more and more fringe (the AGI doesn't really need to use social engineering to make sure that it stays a non-threat, but does so anyway).
By Yud's own implication he thinks MIRI as it is and as it will be in the foreseeable future would be about as much of a check on the power of malicious AGI as a lone Zulu would be against the entire US Armed Forces. I think it's fine to count them out.
There are different speeds it could go, for one. If a method took 40-50 years it might still be completely effective but we might not notice where it was going until several years in.
It could work in stages; there could be some progress by intent and some by accident.
When photography was invented, I don’t know if anyone intended it to be used to decrease the birth rate. But photography has led to immersive VR. Making people totally uninterested in reproduction and physically unable or unwilling to sustain pregnancy and infant care would bring us to an end pretty quickly. Then the machine wombs malfunction and oops! Gone. There are many variations on this. I hadn’t given it any thought before it began to be discussed here but the more time I spend, the more partnership-doom scenarios I envision.
Yeah, nannies can read 'em stories written by GPT3: "One day the cute little generic was [locomotion-appropriate-to-generic]-ing down the [Disney-weighted-path-synonym] and . . ."
Haha exactly. I used to start to fall asleep while reading stories, but continue to talk. About unrelated things. Once I suddenly woke up because I was explaining why one of the characters was divorced, and that suddenly seemed wrong (it might have been during the Little Red Hen). One of the kids was asleep and the other was kind of looking at me funny. Robot nannies would not have that problem!
The whole AI situation raises the question of which human frailties we need, and what we need them for.
Disney-weighted-path, that’s great. Fairy tales used to be disjointed and creepy but in a real-world way- now the kids have to use Five Nights at Freddie’s.
My intuition is that to make AI with human-level intelligence you have to have a system that is subject to most of the human frailties and irrationalities. You need a system that is capable of procrastinating, falling in love, self-sabotaging, having an orgasm, being deluded about itself, yearning for novelty, developing a phobia, being sentimental, getting caught up in fads. It's not that these frailties and irrationalities contribute in a direct way to every single task of human-level difficulty you might ask the AI to do -- it's that all the wacked, horny, emo stuff is part of the deep structure that makes the being capable of human-level insight, creativity, ability to weigh complex options, etc.
Cynical snotty view regarding superintelligent AI offing us all: Some people's conception of ASI is basically that it's so much smarter than us that we can't even conceive of what it would be capable of except that it would be, um, inconceivably impressive. Under that definition of ASI, the answer to Jack Wilson's question is easy: ASI would off us by some method that's so inconceivably clever that we'll never be able to even guess at what it would be.
No, the "inconcievable yet bad" does not cancel with "inconcievable yet good" because out of all the possible value functions and outcomes in the world only a tiny fraction are good for us in particular (the tiny fraction of the universe or Earth that matters to us) and the vast majority is bad for us.
When you live in an environment that's somewhat optimized for our needs (and before that, we were somewhat optimized for that environment) any large perturbation in a random direction is exceedingly likely to be harmful. It theoretically *might* be good, but it's essentially an insignificant probability; we're somewhat close to a local optimum and we know from history how easy things can get much, much worse even with tiny changes, much less large ones - so we should have a strong, prudent preference to the status quo versus any unpredictable large changes.
yeah... i see some missing steps in there, namely the disassembling and reassembling. and considering that the drexlerian universal assembler is probably impossible, how would these intermediate steps go?
Yeah, and that's the critique; instead of offering a concrete argument for how an AGI goes from being locked inside of some lab in Silicon Valley or the Pearl Delta to converting the entire universe to computronium, the argument is little more than "Well, it's AGI, so it's basically an evil god capable of anything that doesn't have a hard physical law preventing it." And at that point, you'll forgive skeptics from classifying this as a religious belief and dismissing AI risk as a meaningful worry. If you're basically arguing that the AI will be as close to omnipotent and omniscient as it can be, such that it cannot be stopped once brought into reality, then the options left are "end all AI research, by force and civilizational collapse if necessary" or "accept that whether this happens or not is up to fate, not us."
I think agi x risk is silly. That said, an agi trying to maximize converting all resources into X finished good seems more threatening than one that for reason wants to convert them into misc. rubble
I’m with you on the silly. I haven’t done any work in the AI space since it was pretty much just various solution space searches and neural nets designed to detect camouflaged military gear were accidentally learning to detect cloudy days instead.
When someone shows me code that produces Pneuma I’ll begin to worry.
I’d want to see something indistinguishable from human desire at the very least.
I’m only basing this on a feeling in my belly and 5 years of meditative inspection of my own consciousness though.
From what I've read, no. It's because models evolved over time and the current ones tend to show less-than-fully-apocalyptic post-war picture (citation needed, but I think I saw some posts claiming this).
Yes, that’s true of the nuclear winter as far as I know. Maybe the persistence of radiation was over estimated. But there were 70,000 active nukes during the Cold War. Villages in Ireland were targeted.
The northern hemisphere would have been literal toast.
If there is no nuclear winter, turning all the nuclear countries and their direct allies into glass is not an existential risk, as that leaves half of the global population intact; and even literally turning the whole northern hemisphere to ash leaves South America and half of Africa, something like a billion people and humanity should be able to recover in a century or two - i.e. absolutely horrific, but not even close to existential threat.
AGI is AI with human-level intelligence. ASI is one with superhuman intelligence. Doom scenarios are different depending on the flavor of AI. Which did you have in mind?
I don’t believe that AGI is possible (or proven possible yet). As far as I know though AGI is ASI, if we can catch up to humans then we can easily exceed humans.
It's certainly proof that something as smart as a human can exist. It's not at all proof that human beings themselves can create something as smart as themselves (in some way other than the usual). The fact that stars exist does not prove we will ever be in a position to make one. A horse might observe that cars exist but that does not prove that horses will ever be capable of building cars.
It's part of the typical egoism of our species that we readily acknowledge that every *other* animal species has inherent limitations on its capabilities, including those things that require intelligence, but we acknowledge no such limitations on our own -- because we are gifted with language, the ability to describe if only in vague terms what it is we want to do. We think we can build an AI as smart as us essentially because we can write down the phrase "an AI as smart as us" and at least at a superficial level, without digging very deep into the meaning of each word, that phrase scans and makes sense.
“We might think we can build an AI just because we think we can do anything” is not a very convincing argument to people who remember all the “a computer could never play chess well” sorts of arguments. Thinking humans are so special as to not be replicable is also egotistic.
That’s a straw man argument. Not only did nobody ever claim that, in fact most people who believed in AI were expecting something like HAL long before computers could beat humans.
"If we can put a man on the Moon then why can't we [insert random difficult feat here]?" arguments are silly. Not to mention I never heard anyone of sophistication and understanding argue that a computer could never play chess well. Can you perhaps point to someone who made such an assertion 40 years ago, who actually knew something about computers?
I never said humans can't be replicated, or even improved upon. I expressed doubt that *human beings themselves* could do so. You're focused on the outcome, and just assuming that if something is possible at all then humans can do it -- which is exactly the kind of egoism I'm pointing out. We have limits, just like any species. There are certainly plenty of things that are possible in principle which are not possible *for us* because of those limits.
I'm pretty certain that a prediction of 100+ years would result in approximately $0 in funding now, and predictions of even 50 years resulting in similar funding levels. The conflict of interest in predicting very near AGI also makes me skeptical of those predictions.
This makes sense for physicalists, and as far as I can tell this intuition is the main basis for evil AI fears. But I suspect that anyone who believes there is more to consciousness / agency than pure physical wiring (like myself) would have a hard time getting worked up about evil AI.
Even if one is a physicalist, we still have no clue how much of our brain power is due to quantum effects which may not be replicable using transistors.
The interactions in our brain happen in a "noisy", hot environment with excessive interactions in which any entanglement effects can't be sustained for any meaningful time, and we know of specific non-quantum mechanisms (electric potential, ion transfer, neurotransmitters) that we can observe implementing specific functionality of animal brains.
It's misleading to say "we still have no clue". We have no strong proof that the quantum effects *definitely* don't matter, so it's technically possible that they do, but since we have a bunch of evidence that they don't need to be involved to explain observed phenomena and literally no evidence whatsoever that they are involved, it's appropriate to say that we do have some clue, and more likely than not nothing of our brain power is due to quantum effects which can't be replicated/modeled on a macroscopic non-quantum level.
I've had that thought too. And moreover the process that created us, evolution, was incredibly dumb, slow-motion trial and error. Couldn't a better-directed, more efficient version of the same process produce human intelligence or better in a lot less time?
i agree its not proof of that; i'm personally pretty damn confident that human brain meat isn't all that much more efficient than transistors, but even if it is then we'd just need to spend a few centuries figuring out better computer hardware.
Seems easy to me to think of ways AGI could off us all. You don't even have to think about the AI aspect of it -- just think of ways a bright person might do it. Yeah, killer viruses and the like the easiest route to me. Even if the person or AGI didn't have access to more than, say, a few hundred thousand dollars for equipment, and didn't have technical training beyond a grad degree or equivalent in a relevant field, I'll bet it would be possible for them to develop half a dozen highly lethal, highly transmittable viruses and distribute them widely enough that they'd get everybody. There would be some luck involved. The FBI or some such might catch on and bust them. If the killer is human rather than AI they might accidentally infect themselves and die before they complete their plan. But it all sounds quite do-able to me. Though biology and the like is not my field, and I may be all wrong. Wut you think?
Are we allowed the possibility that the AGI can function at human level in armored shells, and Maravec's paradox has been satisfiably resolved? And that they have the production capacity to form battalions?
Then the standard scenarios I've seen are for ASI. It could figure out how to comandeer network-controlled resources like WMD's, nuclear facility controls, dam gates, drones, etc. From the pro-anthro standpoint (I more-or-less lean this way), it would seem important those were adequately firewalled.
Without physical agency, it's a long shot for human annihilation, imo. Probable that economic mayhem is more to be feared.
Nanotechnology + poison, possibly? I’m not sure how to go into detail on that, since I don’t know anything about nanotechnology other than that it should be entirely possible for superintelligences to develop it.
Except that drexlerian nanotech as envisioned in sci-fi (grey goo included) is probably impossible under known physics / chemistry (on this, the smalley-drexler debate is great)
Bacteria are proof by example that you can make tiny self-replicating things. Whether they’re technically “nanotechnology” isn’t really the point? I would expect the best self-replicating machinery a superintelligence could make to be pretty small, but if they aren’t that doesn’t change anything.
Well, if we want to move to biorisk, i fully agree that that if a potential x-risk. For as much as we know, something very close to biology is probably the best, most efficient "nanotech" possible
Machine learning may already be capable of producing ultra-lethal chemical compounds, by simply re-purposing drug discovery algo's: https://go.nature.com/3KQWfID
Here's a fun thought exercise I came up with that y'all might enjoy:
So, we accept* that the most reproductively successful man of the last thousand years or so was Genghis Khan, due to all the unsavory behavior, high status & reproductive success of his legitimate issue, etc. A more challenging question though, is "who was the most successful /woman/?" Without an equally powerful mitochondrial DNA study, linked to an unusually fecund historical woman, the question might initially seem unanswerable, but with a little thought, I think it can be solved!
My answer rot13'd:
Gur nafjre vf npghnyyl dhvgr fubeg, ohg V'z cnqqvat vg bhg jvgu guvf rkgen grkg fb crbcyr qba'g havagragvbanyyl vasre vg whfg sebz gur fgehpgher bs gur fragrapr, ohg ertneqyrff, V cebcbfr gung gur jvaare vf... uvf zbgure. Rib-cflpu / fbpvbybtvpny vzcyvpngvbaf yrsg gb gur ernqre.
* Zerjal et al. (2003) show ~8% of men in a large region of Asia (over 10 million individuals) are descendants of the same guy via shared Y-chromosomal haplogroup, who is proposed to be Khan. If you don't accept that that guy was the historical Khan, well, let "Khan" mean "that guy".
One issue with your solution is that it appears the sons of Genghis Khan also had very high fertility due to their high social status, but the same logic is not said to apply to his daughters.
Genghis Khan's mother, unless we can find e.g. a set of brothers who collectively out-reproduced Genghis.
Note - "reproductive success" shouldn't be measured via only male-line or female-line descendants, so the mitochondria are irrelevant. And Genghis' Y-chromosome is only relevant because we have no reason to believe the majority of his grandchildren came from his daughters, and of course useful because it and the mitochondria are conserved more or less whole, much longer than somatic chromosomes, or the X, which recombine.
I figure it was someone several generations before Genghis Khan, rather than Genghis Khan himself. His leading men probably had a lot of the same genes, and similar reproductive opportunities to the Khan himself.
For M-DNA? Genghis had no important sisters. His chief wife, Borte, is not a terrible guess though - five daughters who married into important Mongol Empire noble families.
Mitochondrial Eve wins, because we are all her descendants. If you mean in the last 1000 years or so, she doesn’t have to be exceptionally fecund, but just to have given birth to at least one child who went on to have many descendants. So without any further research I’m tempted to say Genghis Khan’s mum!
Kind of skimmed it because so much is routine, even stereotyped. But the part at the end I very much agree with, which is that if you want to genuinely have new ideas, you need to unplug considerably. When you're awash in the conventional wisdom, checking in 6 times a day to see if what you're thinking isn't contradicted at line 1 by a Wikipedia article or thinkpiece by someone considerably more educated -- you are *never* going to have an original brilliant idea. You will of course be saved from having about 100,000 original but stupid ideas, and if that is your priority, that's fine.
But you cannot be creative without being willing to be wrong 99-100% of the time, and if that last (merely potential!) 1% isn't good enough for you, you should not expect to be innovative, and if it's not enough for your entire generation, you should not be surprised when innovation drops off.
I disagree with the central thesis. Thinking and talking about progress is necessary if we want to optimize for its speed, the social and technical landscape is quickly changing and approaches that worked for Newton won't work for Einstein, Einstein's won't work for ??? (to be filled in a decade I guess), etc.
It also doesn't help that the author is salty af and spends more time yelling at clouds than making his point, which is by the way purely destructive and offers no alternatives.
He sees an issue, but doesn't understand that it's systemic. Real science is hard and underappreciated (and doubly so for real AI safety work btw, but since that's even lower status it's an easy target for drive-by sneering), whereas "insight porn" is comparatively easy and decently rewarding. So it's no wonder that smart young people are attracted to it, regardless of whatever this or that blog post may have to say.
What do you mean by the issue's being systemic? Do you mean it's just a manifestation of human nature (e.g., smart young people's being attacted to easy & decently rewarding things)? Or do you have another system in mind other than human nature? I'm asking because I have a systemic theory -- not the "human nature" theory -- and am wondering whether you have a similar one.
I mean that the human nature is embedded in the system of incentives, which largely determines what the society ends up looking like. Of course, it's possible to introduce changes to this system, but the mechanisms so far are very primitive, unreliable and unpredictable, and that's before taking into account that there are many deep disagreements about what changes are even desirable!
It's expressing a frustration that I've also felt. I tend to like "insight porn" articles a lot (and I think the analogy with actual porn is unfair) but they're definitely in oversupply these days relative to actionable intelligence. That too may be a low-hanging-fruit effect. I think the post is wrong in attributing this stuff to a desire for fame and fortune; as far as I can tell it comes out of an honest desire to contribute and make a difference, combined with a certain amount of streetlight effect (https://en.wikipedia.org/wiki/Streetlight_effect) in how one pursues that.
More broadly: I will gush about the good intentions of the broader internet-rationalist sphere all day. Rationality itself is where they often come up short. Human behavior is anti-inductive at all levels to an extent that is still under-appreciated around here. If there were a cheat code to scientific progress we would have found it already-- or rather, "There is another theory which states this has already happened." Patterns for success in the past are not a reliable guide to patterns for success in the future.
This essay's a gift that keeps on giving, it's just that "success" now sounds crass so the life coach TED crowd pivoted to innovation and sustainability and whatever-the-fuck.
This is one of those things where it's really on point and exactly what certain readers need to hear, except that it applies to an extremely small demographic of people. The average person probably doesn't think about {progress studies, innovation, effective altruism} enough, and I'd be sad if they wrote off ever starting because they read something like that. But for a certain slice of people, it feels spot on.
I don't think I understand how any Neal Stephenson stuff could be considered a scathing indictment of "rationalism." Would love to understand what I'm missing.
I read lots of respect for being an introspective thinkin' kinda monkey in his stories; and a decent amount of contempt for any one who would unironically refer to themselves as a rationalist.
He writes books that deeply value rationality, but hate it's reification into some sort of identity.
I am not Nah, so don't know what their answer would be. However, there is a lot of tension between rationalism and non-rationalism in the Baroque Cycle. There are protagonists that embody both sides of the coin and often the characters' internal conflicts arise from which side they should listen to.
Newton and Daniel Waterhouse both struggle with balancing their rationalism with religious conditions. Daniel taking a more pragmatic approach while Newton clutches too hard to his religious beliefs in the face of "rational" evidence to the contrary.
Thought its cowritten, The Rise and Fall of DODO is more explicit in its discussions of rationality (aka scientism) and the dangers of flying too close to the sun.
"Page number" is a slight exaggeration. I would put Anathem and Seveneves more or less fully in the "ringing endorsement" camp and REAMDE and Cryptonomicon much more into the "wow this guy kinda despises rationalist types doesn't he" camp, with Fall and Snow Crash too culture-war-ish to rise to the level of either.
It probably does... I admit I don't remember it that well, and perhaps I'm reading later stuff (eg Fall) back into it.
My take on Stephenson is that he writes based on political views formed in the 90's, when STEM types were solidly aligned with mainstream Democrats against the religious Right (cf. creationism teaching controversies, internet scares, etc.). In Snow Crash he embraces the "conservative = religious = suburban = racist" line of that era in an uncritical, even cartoonish way. Since then the alignment has gotten a lot less cozy and Stephenson has shown that his loyalties lie exclusively with the Blue Tribe, not the Gray Tribe, when the chips are down. His vision in Fall of the good guys as the "reality-based community" was cringey and ten years out of date on the day it was published.
I don't think that's quite historically accurate. In the first place, if "STEM" includes more than "computer programmers" then it's certainly inaccurate, a large fraction of physical scientists and engineers are not especially in sympathy with the mainstream Democratic Party, even today.
And even in the case of programmers, the situation is considerably more nuanced, as during the 70s and 80s the most avant-garde programmers tended to be libertarian freaks like Stallman, and the remainder, building code for mainframes to process 10 million credit-card transactions per second, were solid middle-class suburban worker bees, as likely to be Republican or independent. The 90s were arguably a period of transition, as a new generation of Silicon Valley professional Internet-focused coders was arising, breaking off from both the GNU weirdos and the suit-and-tie IBM VM/SP programmers.
I actually think this take is pretty wrong. There’s a recurring theme in Stephenson’s work about the resilience and benefits of red tribers, although it’s definitely declined over the years. Compare the diggers from Seveneves (essentially survivalists who institute a theocracy underground to ride out the apocalypse) which is both less favorable than his depiction of the extended Shaftoe clan in earlier works, and more favorable than the Ameristan excerpt from Fall, which I will grant is pretty bleak. I get the impression that he has a lot of family in flyover country but has been gradually souring on the group as it gets crazier
Snow Crash was fun. I made it half-way through Cryptonomicon before shelving it. I made it a few dozen pages into Anathem before abandoning it. For someone so lauded, his writing was less than enjoyable.
I think a valid criticism of Stephenson's writing is that the books can take a very long time to get into (especially Anathem - its my favorite despite that). His climaxes can be pretty meh too. You have to really want to just spend time in the world and with the characters as well as enjoying the plot otherwise the books are a chore.
Good SF of a certain kind gets mentioned a lot as fitting these criteria. My personal picks would include Foundation (Asimov), Speaker for the Dead (Card), and Anathem (Stephenson).
Arguable for these or any candidate how much of the gateway-ness is treatment effect vs selection effect, though.
might depend on when you first read it, but it was definitely my gateway into rationalism. I wouldn't claim it's the best, as I haven't surveyed the field, but I think it definitely belongs on the list
EDIT:
On a little more consideration, it probably depends on whether you parsed the opening post more as asking about " 'The best piece of art' that has the qualification of being a gateway to rationalism" or " 'the best gateway to rationalism' that is a piece of art"
I think that HPMOR very naturally segues into the early Eleizer's sequences and also makes it pretty clear what rationalism is about, so it's good for the second interpretation. On the first, I enjoyed it and wouldn't call it cringe, but from an artistic perspective, it's probably lacking
Amusingly, the dust specks thing is the first post by Yudkowsky that I chanced upon, and I hated it so much that for years I dismissed him as having anything worthwhile to say. But because the LW jargon was everywhere and Scott always praised him so much I finally read the damn sequences. Somehow it's still the post I dislike the most, and consider the whole "meta"ethics part the weakest.
Definitely not a gateway. Maybe if you're already inside the compound and you look over your shoulder, the gate you came through looks something like those photos. To those of us still roaming the terrain outside the walls he looks like a bloke with a beard.
Interesting, that hadn't occurred to me. It's certainly extremely *nerdy* for a successful artistic work (the "Glorfindel" scene is my favorite example of the essence of nerdiness) but not in ways that are distinctly rationalist. It focuses on solving practical / technical problems in a nerdy way while taking the parameters and goals for granted. Rationalism, to me, implies more of a focus on epistemic and societal problems.
On a trip to Vermont a guy who makes maple syrup told me that, counterintuitively, sap flows through a tree faster when it's cold out. I didn't have a chance to ask him why. Any theories?
It's not that sap flows faster when it's cold out. It's that sap flows faster when it's cold at night and then warm during the day. Basically, the night subzero temperatures causes the sap to freeze and then be squeezed by the tree (which shrink slightly due to low temperatures). This causes the sap to become much more dense. Then when it gets warm the tree expands out and the sap melts and expands. This causes faster flow. If it's too cold (or doesn't get warm enough) then the sap can just remain frozen all day. If it's too warm then the sap doesn't get condensed and will still flow but slowly.
I've worked tapping trees and on some other agricultural stuff so AMA. I guess I should disclaim this isn't some scientific inquiry but what the tappers told me.
As brilliant as zdk's reply is, the explanation widely given here in REAL maple syrup territory is that nights below zero and days above give the best flow of maple sap. Too cold, doesn't flow, too warm straight through and the flow stops soon.
This makes intuitive sense, but the actual mechanism is still unclear to me.
"4. What are the right weather conditions for sap flow?
The tree’s sap flow mechanisms depend on temperatures which alternate back and forth past the freezing point (32 degrees F.). The best sap flows come when nighttime temperatures are in the low 20s and daytime temperatures are in the 40s. The longer it stays below freezing at night, the longer the sap will run during the warm day to follow. If the weather gets too cold and stays cold, sap flow will stop. If the weather gets too warm and stays warm, sap flow will stop. The cold weather at night allows the tree to cool down and absorb moisture from the ground via the roots. During the day, the tree warms up, the tree’s internal pressure builds up, and the sap will run from a taphole or even a broken twig or branch. For good sap production, maple producers must have the alternating warm/cold temperatures. This is why its so impossible to predict the outcome of the maple crop from year to year."
Came here to say this; it's the *difference* that's important. Incidentally, this is part of the reason why maple plantations tend to be a bit inland rather than on the coast (source: grew up east coast Canada, had many fond tir à l'érable experiences)
The bound on entanglement is, said differently, a bound on the size of the Hilbert space describing the black hole, and it is precisely this that is meant by “the information contained in a black hole” in the context of the Bekenstein bound. The Bekenstein bound is a limit on how many qubits you can fit into a volume, which in turn is a limit on how many classical bits you can fit in that volume.
Seems kind of in-line with my comment in the links thread that a lot of political divides these days are less left vs right than they are insiders vs outsiders.
PredictIt has Macron/LePen at 91/8. As I understand it, Le Pen was leaning into culture war on immigration, but retaining a Statist outlook economically and per safety net. The 'far right' designation pertained to her social views. By American standards, that's strange and interesting.
Immigration is sort of a hard switch. If a politician runs as a Radian caricature they will still likely be called center right, whereas if a candidate does a copy-paste of a socialist party platform with immigration restriction tacked on and it becomes far right.
I guess I'm a bit surprised that you're surprised at this. The Netherlands apparently already has policies restricting support of Nazism, so it shouldn't be too shocking that they'd expand this to other viewpoints that the vast majority of people find abhorrent.
How is it that every time someone pops up about "hey, this party/this person is being persecuted just for expressing a simple opinion", it turns out to be kiddie-diddlers?
Pretty sure you don't have evidence that most of these people have done the thing you claim they did.
Aren't people also persecuted for arguing the holocaust never happened? This seemed like a much bigger thing a few years back, before YouTube banned those videos...
Burdens of a community that openly values speech. One of my favorite quotes from the old Slate-Star: "if you’re against witch-hunts, and you promise to found your own little utopian community where witch-hunts will never happen, your new society will end up consisting of approximately three principled civil libertarians and seven zillion witches."
It's worth it on balance, but one of the downsides of trying hard to be a tolerant space is that a subset of your population will always be people who are attracted to the space because they have found their ideas (or themselves) to be intolerable elsewhere.
Re: Point one, I'm not sure how laws restricting free speech work in Netherlands, but my guess is that they're implemented through a democratic republic, meaning that the representatives of the majority of people all agreed that supporting pedophilia was abhorrent, and should be punishable by jail (as they did for supporting Nazism). So, an actual majority was needed to implement it, not a single dictator. Freedoms and democracy are not the same thing.
For your second point, that's a big (and incorrect) leap from what I said. Personally, I'm of the opinion that people should largely be able to say what they want; I'm opposed to both Nazis and pedophile-supporters from being jailed for stating their opinions. But, I live in the US, not the Netherlands, and so my input on their system has limited importance. Secondly, a government doing something, and you doing something, are very different things. For example, (using your terminology) I think that the government should be able to kidnap and imprison murderers; that doesn't mean that I think you should be able to.
"What if more people really wanted the PNVD to win?"
Do you legitimately believe the majority of people in the Netherlands support sexual relationships between a 40-year old and a 12-year old? "What if" is all very nice, but doesn't have a lot to do with "it is" in the end.
I think that most democratically governments legally can appoint themselves oligarchs for life. I don’t know much about the Netherlands constitution, but in the US, if the federal and state legislatures coordinated well enough, they could appoint themselves oligarchs for life tomorrow. That they can’t is a pleasant fiction we all maintain
"Their "crime" was the alledged continuation of banned Association Marthijn by publishing documents which discuss age of consent related issues."
I've been here before. I'm old enough to be aware of Peter Tatchell and how the gay rights movement in Britain got entangled with PIE (Paedophile Information Exchange), the arguments that the age of consent should be reduced to fourteen, and other things that became embarrassing, to say the least, as time went on and those involved became respectable:
"Vereniging Martijn (Dutch pronunciation: [vəˈreːnəɣɪŋ mɑrˈtɛin]; "Martijn Association"; stylized as MARTIJN) was a Dutch association that advocated the societal acceptance of pedophilia and legalization of sexual relationships between adults and children."
Now maybe it's wrong and they want the age of consent to be fourteen or something. But you know, whenever I see someone talking about "age of consent", they often turn out to be not just "science proves" and "not criminals", they want to fuck eight year olds.
In other words, you are going to have to do a *lot* of work to convince me that this is just "let's talk about the law" and not 'stealth movement like so many times before to make it okay to fuck eight year olds'.
Yep, having read that, the guy wants to fuck eight year olds:
"I have been sentenced to six months in prison. I quoted murdered politician Pim Fortuyn, who said: Why not make sex between children and grow-ups legal? I wrote that I agreed with Fortuyn on this, because when you look at all the scientific studies concerning harm in pedosexual relationships, the harm seems to disappear when the relationship was wanted by the minor. This is not my OPINION, it is the conclusion of scientific papers, approved for publication in scientific magazines."
Because you know who I've read who all state that "the harm isn't the sex, the harm is the irrational response of adults to grownups having sex with kids, and science says I'm right"? Guys who want it to be legal to fuck eight year olds.
Otherwise, stop shilling for groups that put pictures of four year olds on their website:
"The organization was in the news in October 2007 when it was learned that photographs of Princess Catharina-Amalia (then aged 3½), daughter of Willem-Alexander, Prince of Orange (now King) and his wife Princess Maxima (now Queen), were on display on the website's forum. The Prince went to court to request a €50,000 fine and the removal of the photos from the website. The court agreed that the photos must be removed, and imposed a fine of €5,000 to be paid every time photos of children of the royal family are placed on the site again. The organisation had to pay €1,235 in costs"
We all know it will never be legal to f*** 8 year olds, so what does it matter?
If some scumbag says women who dress provacatively are asking to be raped "and this scientific paper agrees", would you support a prison sentence for saying that?
Sounds like yes. But I guess I'm too Americanized and think that a banhammer is enough.
"We all know it will never be legal to f*** 8 year olds,"
We all know it will never be illegal to open a public occasion with a prayer invoking Jesus Christ.
We all know divorce will never be legal widespread, no matter what a few countries/states of the US may do.
We all know sodomy will never be legal.
We all know contraception will never be legal.
We all know abortion will never be legal.
We all know gay marriage will never be legal.
We all know trans people in bathrooms not their biological sex will never be legal.
We all know... but I've seen it happen in my lifetime.
You really think that the push for "children are sexual beings with their own agency" will *never* be legal? Since age of adulthood has come down, age of consent has come down, acceptance of "14 year olds are having sex even if it is illegal and even if it is statutory rape", minors can get abortions without the consent or even knowledge of their parents with the help of a complicit judge, proposals to drop the voting age to 16, 'we have to teach three year olds about gender and sexualities, here's your Blues Clues video' and increasing sexualisation in society, that in ten or twenty or fifty years time, "sure, eight year olds can consent to sex, kids today are mature earlier than in the past" will *never* happen?
Okay, so your concern is a slippery slope. But instead of saying "if we let gays marry, next thing you know people will be allowed to f*** 8-year-olds" like some people used to suggest, you're saying "if an adult can get away with *advocating* allowing 33-year-olds to have sex with 13-year-olds, next thing you know they will be allowed to actually f*** 8-year-olds".
Which, I think, means that you think that the arguments in favor of allowing adults to sex 8-year-olds may be persuasive to a majority of the population, such that this change could actually pass into law. And so, you think, it's important for those arguments not to be heard in the first place.
Well, okay, but my sense is that your opinion is extremely common, and that a supermajority are against letting 33-year-olds sex 13-year-olds, so the thing you worry about is very far from coming to pass. (edit: this seems especially so if the people advocating for this change sound like the people I saw on the link Andaro originally posted and then deleted. Those guys, well, they sound like people who want to f*** eight-year-olds)
FWIW, age of consent in a large part of the EU is 15-16 and the sky hasn't fallen yet.
It helps that people will shun you if you have a sexual relationship with a 15 year old girl as a 40 year old man, so social pressure solves the most obvious problems. Meanwhile teens can fuck each other in peace.
The historical pattern is the opposite. Under Rabbinic law, a woman could get married without her parents' consent at twelve and a half plus evidence of puberty (a man at thirteen and a half, I believe). In 18th century England, the age of consent for marriage was twelve for women, with ten the age at which intercourse counted as rape even with consent. In most of the US in the 19th century age of consent was ten or twelve. The current pattern, sixteen to eighteen in, I think, all of the US — I don't know about Ireland or the Netherlands — is unusually high by historical standards.
Sure, it's not outside the bounds of possibility, though it does go against the flow of the primary basis of the morality that produced most of those other changes over the last 50 years or so.
Societies have allowed it before so it's possible.
I wouldn't bet a penny on it in the next 30 years but it's possible.
If you just view all of those other things as moral decay then hypothetical 8-year old fucking is just another type of moral decay that all those others will obviously lead to.
If you view most of those things as moving further and further away from the days when the local bishop could diddle 8 year olds without anyone raising a finger to stop them then it seems less like an obvious direction for society.
I get the impression from online sentiment... as a lot of internet-generation women hit their 40's and start to really really resent that men their age are more interested in 20 year olds there's going to be more and more of a push to treat 19/20/21 year olds as pseudo-children.
I just really detest ratchets that are only ever allowed go in one direction. Do you think society would be better if the age of consent was raised from 16 to 17 or 17 to 18? more power to you! campaign away. But if you brand everyone who goes "eh, it seemed fine at 17" as a pedo then you're creating a ratchet effect that only ever allows one direction of movement which isn't a particularly good thing.
When you've got a social issue like that people should be able to push the boundaries slightly in either direction without being branded as evil.
Free speech is normally interpreted to mostly mean free speech about political issues, on the grounds that in a democracy the most important decision for the people to discuss is who should govern, and we need strong protections for people to be able to discuss that question openly.
But it rarely means speech qua speech no matter what. If I start a Website devoted to normalizing cannibalism, or giving people explicit instructions on how to kill themselves -- maybe in exchange for their livestreaming the event -- or I start openly advocating for throwing all the Jews in the gas chamber, and trying to organize people do to planning, maybe get a private organization to build a trial death camp on private ground -- I am going to get shut down by the majority tout suite, and you'd have to be pretty naive to be surprised or wonder why.
As Justice Jackson famously quipped, "the Constitution is not a suicide pact," and we can say the same thing about constitutional rights, or natural rights one might want to adduce in a country lacking constitutional rights per se: they are not unlimited and absolute, and there *is* speech we rule out of the public forum because the cost of bestowing even the lightest sheen of acceptance or approval on a river of raw psychopathic sewage is not worth whatever benefit may otherwise accrue.
Joseph Bronski, whom I vehemently disagree with, at least has an overarching argument about age of brain development and age of maturity (including age of consent).
People who only bang on about age of consent in relation to sex, so that they may lower it, so that adults having sex with children is not illegal?
Yeah, they wanna fuck eight year olds. Defend them if you like, but if you lie down with dogs, you are going to get up with fleas.
The guy in question didn't go to prison for quoting science, he went to prison for breaking laws about "yeah I wanna fuck kids".
I use a rough "would I recommend this review to others"-scale for rating: 10 = I'd recommend it to a friend; 8 = I really liked it, but the typical ACX reader may not, 7 = not my topic, but I feel the typical ACX reader would like it; below 5 = waste of my and your time.
From what you wrote, this sounds like a below 5 category. And I don't mean this harsh in any way, I read all of last years runners and almost never gave <5, most reviews are really really good and I apreciate all the time that went into them. But if you only can recommend 10% of all articles, which ones would it be? This is the question Scott has to solve with promoting ones to finalists.
Personally, I only voted for the ones I liked (or disliked because I felt they were not including obvious facts).
Others I read but didn't have a strong yes or no opinion on, I didn't vote for. I think if you thought the review was poor and that's why you didn't finish it, then it's fair to vote on it. Low votes for "This is not well done" are more useful data than "Nobody voted for this, no idea why".
In my case, the odds are zero, because the word I always guess first is not in the list of possible Wordle answers. If you start with a word that _is_ in the list then you have a 1/2300ish chance of guessing correctly on the first turn. (In some sense; the actual order of appearance is fixed. But if you don't know that order, but _do_ somehow know that your word is on the list, and haven't been paying attention to what past words have been and whether or not any of them is your word, then the correct probability to assign is about 1/2300.
"Collective consciousness" in the sense of, say, telepathy seems vanishingly unlikely to me. But what might happen is e.g. that some particular topic is in the news and more people than average guess related words, and sometimes one of those words will happen to be Wordle's word of the day. If what you're interested in is Funky Telepathy Stuff, I think it would be very difficult to disentangle from that sort of exogenous correlation.
I saw my niece get it in two yesterday. She only had an a from the first one and was just guessing the second one, she said. 1-2 are flukes. 3 can be worked out.
There are 13000 possible guesses, and the original creator chose 2315 of them to be the target words for the following 2315 days. The criterion for being one of those was apparently just whichever ones were "familiar to [the creator's] partner", creating a strong bias for the target words to be fairly common. cite: https://arstechnica.com/gaming/2022/02/heres-how-the-new-york-times-changed-wordle/
So while the odds of a correct first guess might seem extremely low (one in 13000! wow!) those odds are raised significantly because common words, which will often be the guesses you'd think of first, have a higher chance to be correct. Out of 4 people in my family who Wordle, there have been 3 correct first guesses in 3 months, and while that's higher than baseline, it's not all too crazy given this bias towards common words.
Your characterization of serious films in the mid-to-late 90s through mid-2000s doesn't resonate with me at all. Tons of great work by Spielberg, Nolan, Zemeckis, Hallstrom, Ang Lee, Fincher, Aronofsky, Jackson, etc.
I'd say A.I., Minority Report, Catch Me if You Can, The Terminal, and Munich are all good and unique films (not to mention Lincoln and West Side Story in later years). Castaway wouldn't be a footnote for Zemeckis in my book. - again, a great and unique film.
When you say fiction, do you mean books? You can add the word "literary" to pretty much anything to imply that it has ambitions above pure genre thrills. And the word "classic" tells you that the author did it successfully! So Ian McEwan's Sweet Tooth is a literary spy thriller, but Le Carre's The Spy Who Came in from the Cold is a classic spy thriller (in my opinion). The Handmaid's Tale is literary sci-fi, and if I'm any judge, classic sci-fi.
>That division was very strong in say, american cinema from the mid-to-late 90s to the mid-to-late 2000s
Tarantino and the Coens "did both" in that era. Added: Scorsese also made plenty of films in that era. Terry Gilliam made some good sci-fi movies in the mid 90s such as 12 Monkeys.
Edit: And thinking over all the names above, aren't most all of the directors mentioned here considered to be "auteurs"? The word applies to the director not the work, but it seems to be gesturing in the direction of what you mean, if we say "a movie by an auteur". Of course, there are plenty of "auteurs" who only make art-movies, although don't we usually reserve that term for directors who make somewhat popular films? I wouldn't call Cassavetes an auteur, I'd call him an "experimental film maker". I suppose Wes and P.T. Anderson are auteurs who don't make hugely popular films, though I wouldn't consider their movies to be arthouse either.
The story about how the US disparages ebonics is so sad. It seems that the disregard for this language variant has caused education achievement gap for black people.
I don't think it's a moral panic yet, though it's certainly something that could turn into one. I'd watch for the widespread citing of increasingly non-central and/or trivial examples as a sign-- less and less talk about people losing their jobs, having speaking engagements canceled, etc., and more and more talk about mere criticism as "canceling". You can find a certain amount of this sort of thing even now, but I think you still pretty much have to go nutpicking in order to do so, thus far.
Otherwise rational people treat it like an existential threat to civil society, even more dangerous than when a mob broke into the capitol building with the intent of murdering government officials.
It's clearly not a good thing to exist, but it's not good in the way political correctness is not good.
>Otherwise rational people treat it like an existential threat to civil society, even more dangerous than when a mob broke into the capitol building with the intent of murdering government officials.
This drive-by reveales an allegiance, and makes your whole assessment a rather standard dismissal of your outgroup's fears (or 'hysteria', as you probably like to think of it) rather than anything more interesting
Is there evidence that the political make up of a law school's faculty has an appreciable impact on the political views of the students? Or is there any evidence that the legal profession as a whole has become more progressive over time relative to the general population?
A big difference would be that BLM protests were against a thing that actually happens (extra judicial killing of citizens at the hands of the police) vs a made up fairy tale for boomer LARPers.
Nah's comment was specifically about how people can believe something like cancel culture can compel people to violence.
BLM was about how government murder can compel people to violence. Apples vs oranges.
I'm not too happy about much of that stuff (without remembering/agreeing with the exact details here), and I think there are plenty of people who are against both January 6 and, various things Antifa or whoever did (though I think January 6 was worse).
But does anyone remember the thing before *that* - peaceful protests in the aftermath of Floyd's murder met with huge amounts of police violence? Tear gas, rubber bullets, all that sort of thing. It seems like it's been memory-holed by lots of people. People remember Lafayette Square specifically, but only because of the involvement of the White House; those sorts of tactics happened repeatedly all over the country, and I'm guessing in the vast majority of cases the cops who used excessive force faced 0 consequences. One case where there might have been consequences - the police in Buffalo shoving a ~70 year old protester - they just got off scot-free a few days ago in arbitration.
On Lafayette square here are some links to videos. You can clearly see nonviolent protesters being attacked by police but I don't see a mob attack on the White House, you're free to post any video. IIRC there were some violent protesters but no, like organized attack on the White House or anything.
https://www.youtube.com/watch?v=sMYF-Y_r52Y (not an actual incident, a cop bragging about hitting protesters with his car not realizing he's being recorded ... he was suspended for 10 days)
I mean, that's 60 years ago - the people in question are mostly dead. Your argument rests *heavily* on the "transmitted to children and grand-children" part, and that is far from proven
It's hardly an accident, and nobody thinks it is. The legislators in question think that the Democrats are encouraging voter fraud among illegal immigrants, and *that* is the real reason why they (the Democrats) object to stuff like "you gotta show an ID to vote so we know you're a citizen."
So, yes, voter ID laws may disproportionately hit poor brown neighborhoods, these people would agree -- because that's where the illegal immigrant voter fraud lies. It's like saying laws against crack trafficking will disproportionately affect barrios with strong Mexican gang presence. To which observation most people will say well...yeah.
Of course I understand that the opposing point of view is that voter fraud is all made up, doesn't happen, and if it does happen it's too small to matter. I'm in fact pretty sympathetic to that argument, but people who mindlessly oppose even things like voter ID tend to shoot themselves in the foot over it. I mean, if you're so convinced voter fraud is a nonissue, why oppose voter ID? Why not say, sure, that's fine, no problem -- but we strongly oppose, say, closing the polls at 5pm because then working stiffs have a hard time getting to vote? Getting passionate about voter ID laws -- or *any* kind of vote-security rules -- makes the independent wonder whether there might actually *be* something the partisan Republican attack -- else why protest so much?
There have been two major steep drops in Democratic party identification over the past 80 years[1]:
1. In the middle of the Johnson Administration. That's probably the era to which you refer[2], but it doesn't seem super likely those who left then were the Dixiecrats and those who were left behind had much in common with today's college-educated wealthy New York City/San Francisco core demographic. That really doesn't correspond well to the typical Southern or Midwestern Democrat of circa 1968. So I'd say it's equally likely the people who left the Democratic Party in the mid 60s were suburban voters who were disgusted with the Dixiecrat leanings of the core Democrait Party, and that they left to become independents (there was a big surge in independents at the same time).
2. During the early years of the Reagan Administration. I was there for that one, and knew a number of people who made the switch -- so-called "Reagan Democrats" -- and their reasons had zip to do with race, which wasn't a big issue in the 80s anyway -- and more to do with a sense of disgust over the Carter era feelings of glum diminishment -- with the idea that 12% inflation and crappy economic growth was here forever, and we were just inevitably headed for the status of Great Britain in the Callaghan Era (barely First World).
[2] It's a little complicated by the fact that Vietnam, a war started and greatly expanded by Democratic Administrations, had become so deeply unpopular that Johnson thought he would just drag the party down if he ran in 1968.
The problem with that analysis is the lack of objective evidence. The best-case argument you can make is that they went from publicly beating black people to advocating for charter schools because of racism. And if the last widespread vestige of racism left in the nation is support for charter schools, I think racism is pretty much dead.
"Political decisions which govern people's lives" refers to a wide variety of subjects, only one of which can be the most important. Therefore we have to ignore all the others, for the same reason we have to ignore cancel culture.
This is generally my view. I really can't respect anyone spending significant amounts of time screaming about "Cancel Culture". They should log off and spend some time in the real world.
Where they will find, a few years down the road, that the rules which severely impact and circumscribe their lives have been made without their input, and if they want to do anything about it their only recourse is to vote for Donald Trump or someone like him.
The internet is part of the real world. The extent to which "cancel culture" presently impacts the non-internet world is debatable, but it's kind of a big deal on the internet and people are going to debate it on the internet. If that annoys you, take your own advice.
I do recognize that the debates that happen on twitter/internet/Facebook etc do impact real world policy sometimes. Though I suspect that twitter is more of a lagging indicator than the cause of those policy changes.
Moves to support candidates who represent themselves as outside the mainstream (Trump, Bernie, what have you) aren't the result of twitter. Ron Paul became popular without twitter and there has always been fringe candidates and political movements in this country. That they are more popular now is because of how terrible the mainstream political parties have been the past two decades (or more). Twitter has just been a venue for these voters to express themselves.
I also think there is a big difference between large social networks like twitter and Facebook and a site like ACX or even a sub reddit (the main page of reddit is more a social network). The goal of social networks is to create "vitality" or news worthiness. This pushes them to amplify the worst voices and to almost create attitudes and opinions that people didn't even know they had.
Cal Newport recently called social networks the "Roman Colosseum" which I think is a good analogy. Most users log on to watch "influencers" battle it out (though because no one dies the battle never ends and both sides can say they won).
The *war* never ends. The battles end every two years, on the first Tuesday after the first Monday of November. At least in the Colosseum, a "thumbs up" majority only killed one unfortunate SOB. Now we get a higher order of entertainment.
What you call cancel culture I call terrible management and labour practices. I think that is a great example of why using the term "cancel culture" is stupid. When you call that cancel culture you lump it in with idiot politicians complaining about their tweets not getting read. It minimizes the actual behavior that SDG&E engaged in.
Also, in this case Cafferty, has a clear form of recourse: to sue for wrongful termination. Which it appears he has (or it maybe for defamation, its unclear in the reporting).
However, Cafferty also sued three people on twitter for defamation who tweeted at SDG&E about him. Is that not also cancel culture? They were expressing their opinion that Cafferty was showing a white supremacies sign. This speech is clearly covered by the First Amendment.
I'm not aware of legal bans, but Amazon has delisted books from its marketplace (including second-hand sales) on a number of occasions, which is a de facto banning for rarer titles.
A private business choosing not to carry a product any longer isnt in the same ball park as a state government creating content based legislation regarding which books can be carried in public school libraries.
Dr. Seuss Enterprises withdrew the book from publication, voluntarily and on their own initiative. It was not banned. Seuss himself updated and changed the language of the book back in the 70s as he thought it was offensive.
In this case the shop keeper put a sign out saying they weren't carrying a product any more and people rushed to buy it all up. Sales of the book soared after they announced publication would cease. Not sure who the Mob was in this case that benefitted.
It should be interesting to watch the response to Ross Douthat’s column in today’s (April 17) NYT. A politically moderate public homosexual (Jonathan Rauch is the other), Douthat concludes that would-be progressives “have been running an experiment on trans-identity youth without good or certain evidence,” declaring the silence of liberals who agree with him “will eventually become your regret.”
That has to be the boldest challenge to gender identity ideology I’ve seen in decades, clearly echoing the “silence is complicity” accusation of recent liberal protest demonstrations.
I thought he had declared it at one point. If not, I certainly didn't mean the association as an insult. Most of his references to gender dogma have been highly orthodox, so his principled stance surprised me. But he's about as 'conservative' as David Brooks -- the left's pathetic offer of 'diversity' in their propaganda empire.
Ross Douthat is married to fellow writer Abigail Tucker, and they have, I think, four kids together. Like someone mentioned below, I think you might be confusing him with Andrew Sullivan.
Nope, the foundation remains intact. There's just an election coming up, so it's time to reveal that you've just been humoring the radicals, because free speech and diversity, but you really do have moderate roots. This is why the NYT is running chin-stroking thinkpieces on the shocking rise in violent crime over the past two years, and suggesting it might not be a totally bad idea for citizens to consider the police as some kind of partner as opposed to just the pigs, and it's even possible adding some money to their budgets isn't necessarily all bad.
I mean, here of all places you'll find people biting that bullet - many of the old enough commentariat were very Atheist (and often actively anti-theist, including my teenage self) back when that was the big internet debate in the '00s
Yes. It was wrong when the Evangelicals were doing that in the age of the "Moral Majority", and it's wrong that some Republicans are doing that now. I don't think you'll find much disagreement on that here.
It's *just as wrong* for liberals to be doing the same thing, and worse when they piously defend it as "this time it's different, because reasons". You, or at least the people you apologize for, are *worse* than the Moral Majority ever were. They never really had the experience of being on the wrong end of the banhammer to caution them against overzealous use of the banhammer.
And you know what the liberal side was saying back then? That all this censorship was bad and repressive and unwarranted. That there was never any reason for banning something. That merely being offended wasn't good enough.
Then they got into the saddle and suddenly it wasn't censorship anymore, it was "if it's not the government doing it, you're not being censored" like the xkcd comic (which never seems to have considered that by the same argument it made, it would be perfectly okay for a website to kick someone off for expressing a positive attitude to Pride Month):
Yeah, obviously the Evangelical right is fucking insane to about exactly the same degree that the left. I am surprised you think you're scoring points here. The comparison is extremely unflattering for the left.
It's also a matter of conveniently forgetting that today's political / cultural alignment is not necessarily the same as it was in the past. The Westboro Baptist Church had a goal of aggravating both sides to generate as much hate (and lawsuit revenue) as possible; Fred Phelps himself was a civil rights advocate and a Democratic political candidate. Tipper Gore was the visible crusader for a lot of moral regulation of entertainment, such that the warning labels were called 'Tipper Stickers'. Even going back to Prohibition, both parties split in roughly the same proportion ratifying the 18th amendment.
It's convenient and fraudulent to forget that the political left has always included its share of moral busybodies; what's new is the shift from targeting industries to targeting individuals.
test comment
I sent in a bunch of review scores without giving my address. Should I re-send them with my address?
Why is there an international shortage of MAOIs?
They are decreasingly prescribed by physicians despite their extreme efficacy. Is it just not worth it anymore for pharmaceutical companies to produce them?
Follow-up question: where can I get it? (Darknet recommendations?)
I've had asthma since I was a kid but it's been mostly mild. I'd have an attack maybe one or twice a year and it was harder to breath but never to the point where I felt like I would actually panic or pass out. As a kid I had an inhaler but I never bothered with that as an adult.
But yesterday I had a pretty severe attack. Worse than any I'd had before to the point where I had a definite feeling of panic that made it very hard to control my breathing. The panic would constantly make me try to inhale my next breath before I could finish exhaling my previous one. It lasted for hours but it did subside eventually.
So I went to the pharmacy to buy an asthma inhaler and they straight up refused to sell me one. Prescription only apparently. But the problem is I don't have a doctor or internet (I'm on wifi at a café atm) so no virtual visits either. So that leaves the only (official) option of sitting for eight to ten hours in a room full of sick people so I can talk to a doctor for five minutes and get the stupid piece of paper that permits me to buy an emergency inhaler in case of another attack. Of course if I catch something while I'm there that could well trigger an attack in itself.
Ah well. I just looked up asthma inhaler prices online and there's a wide range of prices but possibly I couldn't afford one anyway. (the markup on some of them must be in the range of 10,000 percent markup) So I guess I'll just have to take my chances.
So I'm feeling a little bitter atm. I'm in BC, Canada for those who would like to make note of the data point wrt the relative state of health care in various countries.
In US, I think Primatine is OTC.
Sorry to hear. 8-10 hours sounds enormous. Maybe you can register at the doctor, and ask them if you can leave for a couple of hours and come back?
Thanks for your concern. I was frustrated when I wrote that. Probably waits of eight to ten hours are not typical though they definitely did happen as least as of about ten years ago. Not sure how things may have changed in the covid era as I have not set foot in a hospital in years.
Study suggests that time-restricted eating (intermittent fasting) isn't more effective than restricting calories (dieting).
No discussion of what people tolerate better.
https://www.nejm.org/doi/full/10.1056/NEJMoa2114833
"Of the total 139 participants who underwent randomization, 118 (84.9%) completed the 12-month follow-up visit. The mean weight loss from baseline at 12 months was −8.0 kg (95% confidence interval [CI], −9.6 to −6.4) in the time-restriction group and −6.3 kg (95% CI, −7.8 to −4.7) in the daily-calorie-restriction group. Changes in weight were not significantly different in the two groups at the 12-month assessment (net difference, −1.8 kg; 95% CI, −4.0 to 0.4; P=0.11). Results of analyses of waist circumferences, BMI, body fat, body lean mass, blood pressure, and metabolic risk factors were consistent with the results of the primary outcome. In addition, there were no substantial differences between the groups in the numbers of adverse events."
Sounds to me more like it showed that time-restricted eating IS more effective. p=0.11 isn't bad for a study of just 139 people.
The advantages of fasting would be easier compliance, appetite suppression, autophagy and immune downregulation. There might be some anti-insulin resistance magic going on as well but that can be rolled into appetite suppression.
From the abstract, they eliminated the first two through study design and didn't measure outcomes of the latter two, outside of very distant proxies. Unfortunately, scihub doesn't have the article so I can't comment in greater depth.
EDIT: I just realized they tested 16h fasting, from a hormonal standpoint that's regular calorie restriction with extra steps.
The Master and His Emissary (https://www.amazon.com/Master-His-Emissary-Divided-Western/dp/0300188374) has been sitting in my reading queue for a few years, and I'm about ready to start reading it.
It was published in 2009. What has happened since its writing that should affect how I read it? I'm particularly interested in (a) results McGilchrist relies on that have since failed to replicate and (b) well established (replicated) results that postdate the book and should meaningfully affect my reading.
With Google having become far less useful recently due to a severe bias toward mainstream news and SEO-optimized sites, I'm finding it especially urgent to learn news ways of finding useful and detailed information.
Apart from Google Scholar, what techniques have you discovered for finding useful, detailed information sources *that you didn't know about before*?
Depending on the topic, my first choice is usually "<google-query> site:reddit.com" because it's basically a bunch of topic-specific forums on one domain. That may or may not qualify as "useful", "detailed", or "didn't know about before", but it works for me 80% of the time.
If that fails me: https://searchmysite.net/ and https://search.marginalia.nu/ search a curated list of personal sites. The latter is even better because it has different ranking algos and an option to block sites with JavaScript. If all else fails, there's always https://millionshort.com/.
Lottery of Fascinations... the jump rope expert.
https://www.youtube.com/watch?v=JLC_T1jQ5Lk&ab_channel=WIRED
Wow
Jaw dropped. Thank you.
Everybody else: About halfway through it looks like it's about to stop being interesting. It isn't; stick with it.
Looking for a link from a relatively recent post, about how if people are truly on the fence about making a change in their life ( breaking up, moving, changing jobs, etc. ), there was a study that flipped a coin for them, and found that people who made a change were ultimately happier. Where was that?
On ACX, it was here: https://astralcodexten.substack.com/p/peer-review-request-depression Section 2.1.1.
https://lmgtfy.app/?qtype=search&t=w&q=here+was+a+study+that+flipped+a+coin+for+them%2C+and+found+that+people+who+made+a+change+were+ultimately+happier.&as=0&s=g
Here's a list of all the books reviewed, without subheadings. I don't know if anyone else wants this, but I found it a useful tool for browsing.
A Canticle for Leibowitz by Walter M. Miller Jr
A Connecticut Yankee in King Arthur’s Court by Mark Twain
A Failure of Nerve: Leadership in the Age of the Quick Fix by Edwin H. Friedman
A History of the Ancient Near East
A Secular Age by Charles Taylor
A Supposedly Fun Thing I’ll Never Do Again by David Foster Wallace
A Swim in a Pond in the Rain by George Saunders
1587, A Year of No Significance: The Ming Dynasty in Decline by Ray Huang
Ageless: The New Science of Getting Older Without Getting Old, by Andrew Steele
Albion: In Twelve Books
An Education for Our Time by Josiah Bunting III
An Empirical Introduction to Youth by Joseph Bronski
Anthropic Bias by Nick Bostrom
At the Existentialist Café by Sarah Bakewell
Autumn in the Heavenly Kingdom by Stephen Platt
Bronze Age Mindset
Capital and Ideology by Thomas Piketty
Civilization and Its Discontents by Sigmund Freud
Come and Take It: The Gun Printer's Guide to Thinking Free by Cody Wilson
Consciousness and the Brain by Stanislas Dehaene
Cracks in the Ivory Tower: The Moral Mess of Higher Education
Deep Work by Cal Newport
Development as Freedom by Amartya Sen
Disciplined Minds: A Critical Look at Salaried Professionals and the Soul-Battering System That Shapes Their Lives by Jeff Schmidt
Economic Hierarchies by Gordon Tullock
Exhaustion: A History by Anna Schaffner
Facing the Dragon: Confronting Personal and Spiritual Grandiosity by Robert Moore
Fashion, Faith and Fantasy in the New Physics of the Universe by Roger Penrose
Frankenstein by Mary Shelley
From Paralysis to Fatigue: A History of Psychosomatic Illness in the Modern Era by Edward Shorter
From Third World to First: The Singapore Story: 1965-2000 by Lee Kuan Yew
Future Shock by Alvin Toffler
God Emperor of Dune by Frank Herbert
Golem XIV by Stanisław Lem
Haughey by Gary Murphy
History Has Begun by Bruno Macaes
How Solar Energy Became Cheap: A Model for Low-Carbon Innovation
I See Satan Fall Like Lightning by René Girard
In Search of Canadian Political Culture by Nelson Wiseman
Industrial Society and Its Future by Ted Kaczynski (also known as the Unabomber Manifesto)
Inventing Temperature: Measurement and Scientific Progress by Hasok Chang
Irreversible Damage: The Transgender Craze Seducing Our Daughters by Abigail Shrier
Island by Aldous Huxley
Jamberry by Bruce Degen
Japan at War: An Oral History by Haruko Taya Cook and Theodore Failor Cook
Kora in Hell: Improvisations by William Carlos Williams
Leisure: the Basis of Culture by Josef Pieper
Making Nature: The History of a Scientific Journal by Melinda Baldwin
Making Sense of Tantric Buddhism: History, Semiology, and Transgression in the Indian Traditions by Christian K. Wedemeyer
Memories of My Life by Francis Galton
Mind and Cosmos by Thomas Nagel
More Work for Mother: The Ironies of Household Technology from the Open Hearth to the Microwave by Ruth Schwartz Cowan
MOSCOW-PETUSHKI by Venedikt Yerofeyev
Nobody wants to read your sh*t by Steven Pressfield
Now It Can Be Told: The Story of the Manhattan Project by General Leslie M. Groves
Peak: Secrets from the New Science of Expertise by Anders Ericsson
Pericles by Vincent Aulay
Private Government by Elizabeth Anderson
Public Choice Theory and the Illusion of Grand Strategy: How Generals, Weapons Manufacturers, and Foreign Governments Shape American Foreign Policy by Richard Hanania
Rationality: What It Is, Why It Seems Scarce, Why It Matters by Steven Pinker
Reason and Society in the Middle Ages by Alexander Murray
Robert E. Lee: a life by Allen C. Guelzo
San Fransicko: Why Progressives Ruin Cities by Michael Shellenberger
Slaughterhouse-Five and Breakfast of Champions, by Kurt Vonnegut
Storm of Steel by Ernst Junger
Surface Detail by Iain M. Banks
Sweet Valley Confidential by Francine Pascal
Termination Shock by Neal Stephenson
Troubled Blood by J.K. Rowling
The Age of the Infovore by Tyler Cowen
The Anti-Politics Machine by James Ferguson
The Axis of Madness
The Beginning of Infinity by David Deutsch
The Book of All Hours series - “Vellum” and “Ink” - by Hal Duncan
The Book of Blam by Aleksander Tišma
The Book of Why by Judea Pearl and Dana Mackenzie
The Brothers Karamazov by Fyodor Dostoevsky’
The Castrato by Martha Feldman
The Condition of Postmodernity: An Enquiry into the Origins of Cultural Change
The Dark Forest by Liu Cixin
The Dawn of Everything: A New History of Humanity by David Graeber and David Wengrow
The Dawn of Everything: A New History of Humanity by David Graeber and David Wengrow
The Deficit Myth by Stephanie Kelton
The Diamond Age or A Young Lady’s Illustrated Primer by Neal Stephenson
The Doomsday Machine: Confessions of a Nuclear War Planner by Daniel Ellsberg
The Ecotechnic Future by John Michael Greer
The Eighteenth Brumaire of Louis Napoleon by Karl Marx
The Evolution of Beauty: How Darwin's Forgotten Theory of Mate Choice Shapes the Animal World - And Us by Richard Prum
The Extended Mind by Annie Murphy Paul
The Fall of Robespierre: 24 Hours in Revolutionary Paris by Colin Jones
The Future of Fusion Energy by Jason Parisi and Justin Ball
The Goal / It’s Not Luck by Eliyahu Goldratt
The High Frontier: Human Colonies in Space by Gerard K. O’Neill
The Hundred-Year Marathon: China’s secret strategy to replace America as the global superpower by Michael Pillsbury
The Internationalists by Oona Hathaway and Scott Shapiro
The Irony of American History - by Reinhold Niebuhr
The Knowledge: How to Rebuild Our World from Scratch by Lewis Dartnell
The Man Who Quit Money by Mark Sundeen
The Mathematics of Poker by Bill Chen and Jerrod Ankenman
The Matter With Things by Iain McGilchrist
The Mirror and the Light by Hilary Mantel
The Motivation Hacker by Nick Winter
The Myth of Mental Illness by Thomas Szasz
The Narrow Road to the Deep North by Matsuo Basho
The New Science of Strong Materials by J. E. Gordon
The One World Schoolhouse by Salman Khan
The Origins of The Second World War by A.J.P. Taylor
The Outlier: The Unfinished Presidency of Jimmy Carter by Kai Bird
The Party: The Secret World of China's Communist Rulers by Richard McGregor
The Reckoning by David Halberstam
The Republic by Plato
The Righteous Mind by Jonathan Haidt
The Righteous Mind – by Jonathan Haidt
The Russian Revolution: A New History, by Sean McMeekin
The Society of the Spectacle by Guy Debord
The Tyranny of Metrics by Jerry Z. Muller
The Virus in the Age of Madness by Bernard-Henri Lévy
The Yom Kippur War: The Epic Encounter That Transformed the Middle East by Abraham Rabinovich
Three Years in Tibet by Ekai Kawaguchi
Trans: When Ideology Meets Reality by Helen Joyce
Troubled Blood by J.K. Rowling
Trump: The Art of the Deal by Donald Trump and Tony Schwartz
Unsettled: What Climate Science Tells Us, What It Doesn’t, and Why It Matters by Steven Koonin
Unsettled. What Climate Science Tells Us, What It Doesn’t, And Why It Matters by Steven E. Koonin
Very Important People: Status and Beauty in the Global Party Circuit
Viral by Alina Chan and Matt Ridley
War in Human Civilization by Azar Gat
When men behave badly by David Buss
Whiteshift: Populism, Immigration, and the Future of White Majorities
Who Wrote the Bible? by Richard Elliott Friedman
Yąnomamö: The Fierce People by Napoleon Chagnon
That is... enough material for a separate blog with weekly updates.
In addition to a competition for who wrote the best review, maybe we should have a contest for who can read the most entries.
I read a lot of them, skimmed through more, and skipped a few. I'm now waiting for the final result because there is one question I want to ask one of the reviewers of one of the books, and it's not fair to start a discussion about reviews right now.
But honestly, I was delighted and surprised to see what some of the books reviewed were, and although there were several that didn't interest me at all, that's just my own tastes and I'm very glad there is such a spread of topics covered.
For all the occasional kerfuffles and spats, we definitely are each other's people 😁 📖
I’m planning on reading ‘em all; currently about a third of the way through. So far, about half have been quite good, about 10% have been pretty bad, and the rest just ok
how far have you gotten?
Well, I read & ranked all of 'em. 3/4 were good-to-great, 1/4 were "meh" or worse. Only 11 reviews were actually bad. 13 definitely deserve to be finalists, and another 12 arguably deserve to be finalists. That leaves 74 (about 56%) which I enjoyed reading but don't quite make the final cut. An excellent competition so far!
Wow, it must have taken quite a while to review all the reviews. Do you plan to post more about your impressions of the reviews?
I’ve got 16 left to go - will probably finish in a day or two. Tons of stuff that’s well worth reading. Less than 10% are actually bad. Many are fascinating glimpses of topics I know little about - I always enjoy reading stuff like that 🙂
You are not alone.
Scott, you can consider adding to the book review rating form a free text field for feedback which will be shared with the book review author. Feedback is valuable for improvement! :)
Would other reader be able to see the feedback? In that case, comments will sway the ratings.
Seconded, both for good and bad reviews !
Seconded. One review motivated me to read the actual book because I had at least 5 specific criticisms and I thought I could do better.
Hi everyone who came to the Irvine Meetup, just wanted to say I had a great time! Thank you for dropping by, Scott!
So I was reading Medium today as I often do, and.... something inside me snapped after I saw Medium's reading recommendations. And so I felt compelled to write this. https://medium.com/big-picture/im-sick-of-medium-s-russian-propaganda-7fe63eaaa63f
I'm surprised CodePink is still going. By the tone of your article, I presume whoever remains has moved far-far-left?
I'm not sure because this was my first encounter with CODEPINK. I guess they accepted certain elements of Russian propaganda for the same reason the Gravel Institute did, because they like anti-US messages which Russia was eager to provide. The rest of CODEPINK's article wasn't obviously wrong to me (for the most part), perhaps only because I wasn't familiar with a lot of the issues it discussed. (Notably the same article talked about China-Taiwan tension, for which they predictably blamed the U.S.)
I've began skimming through the reviews. Is it OK to speculate on who wrote what ?
Also, what scaling do people use for ratings ? I feel I've been rating things a bit too high - giving 5-6 to just OK reviews.
7-9 for 'really good', 5-7 for 'good but not the greatest' and 2-3 for 'this is poor/terrible/if I knew who you were, I'd be in a slapfight with you right now, reviewer'.
+1 to having a scale guidance.
My guess is no. Scott has explicitly said he wants to keep it anonymous to not sway the voting.
Re the ratings, maybe rate all the ones on your spreadsheet first, so you get an idea of their relative strength, then submit them?
Yeah I definitely started too high. The quality of some of those reviews (which are actually summaries) are amazing.
I'm real curious to find out who wrote the 'Gossip Trap' review of Dawn of Everything
Is being "intellectually angry" a thing? I don't mean being angry about some culture war issue; it's more "someone is wrong on the internet" but with an added feeling of helplessness that I can't possibly correct them.
The issue is AI Alignment Risk and this website. I love this website. Scott is one of my favorite all-time bloggers. But taking AI Alignment Risk so seriously is obviously so fucking crazy that it makes me "intellectually angry".
I've never understood exactly what the Rationalist community is, never really tried that hard. I mostly like them, yet there has always seemed something slightly *off* about them. Off in the way that Mencius Moldbug is brilliant but also clearly... off.
I realized today what the offness is, for me. It's the difference between Platonic vs. Aristotelean thinking. The term Rationalism has bemused me because I have often thought over the years: "Isn't this just Enlightenment thinking and didn't that start in the 18th century?"
But now I see the difference. The Enlightenment focused on empiricism and was most influenced by Aristotle. Rationalists, in their embrace of Bayesian thinking, seemingly feel free to discard empiricism, and this has led them to believe some crazy, rudderless shit. Such as AI Alignment is a reasonable thing to spend tons of time and money and human intelligence on.
To be clear, our gracious and brilliant host is also a brilliant, trenchant empiricist--when he works with empirical data. Unfortunately, he also seems to update--way too much---on non-empirical issues while in the company of persuasive friends. AI Alignment being the main one.
I think we need more debate between those who believe AGI is an X risk vs those who don't. All the headline debate on the issue here now seems to be between those who believe AGI is a huge risk and those who believe it is only a very, very major risk.
It makes me intellectually angry, if that's a thing.
> "someone is wrong on the internet" but with an added feeling of helplessness that I can't possibly correct them.
Oh yeah, that feeling is a driving force in my life. Makes me write proposals like this: https://forum.effectivealtruism.org/posts/ckcoSe3CS2n3BW3aT/?commentId=u7hPPF5r993iKYL5e
> Rationalists, in their embrace of Bayesian thinking, seemingly feel free to discard empiricism
As an aspiring rationalist, I hereby denounce any such idiocy you may have encountered.
> Such as AI Alignment is a reasonable thing to spend tons of time and money and human intelligence on.
It depends what you mean by "tons". I consider the near-term risk of AGI-induced global catastrophe to be pretty low .... maybe 1% in the next 25 years, or something like that. But that doesn't mean it doesn't deserve billions of dollars of research funding to mitigate. x-risks, Global Catastrophic Risks and s-risks can still have an immense negative expected value even at low probabilities.
OTOH I do kinda think that some other GC risks may be underestimated, e.g. might we be close to reaching Earth's carrying capacity, especially if a big war occurs? I couldn't find any EA research on that. Weird.
I don't see how this is connected to empiricism; I love empiricism, but it doesn't mean that I must assign a probability of zero to any event that has never occurred in human history. To the contrary, history is chock-full of examples of unique events with far-reaching consequences. Doesn't this imply to an empiricist that ze should be concerned with such things? And anyway 0 is not a probability (https://www.lesswrong.com/posts/QGkYCwyC7wTDyt3yT/0-and-1-are-not-probabilities)
Edit: I would, however, add that I think Yudkowsky overstates the case for this particular x-risk, and I'm not sure if it's deliberate exaggeration or if he and I have genuinely different opinions. (Edit 2: actually I'm pretty sure he sees a higher risk than me, but whether it's a small or large difference is unclear. But I think the risk rises a lot *after* the first AGI is invented, i.e. that the first one isn't going to be the most dangerous.) But to throw it back at you, if you estimate that a certain catastrophe X has a 1% chance of happening, and if it happened, would cause damage somewhere between $1 trillion and <human extinction>, how much do you think it would be worth spending on prevention, assuming maybe you could cut the risk in half or so?
I don’t really think you can be intellectually angry, but being angry at other people’s stupidity is a very human thing.. anger is not rational by design.
It’s a feature, not a bug.
on the topic of artificial intelligence catastrophe, here is a quote from an article at Politico:
(Full article here)
https://www.politico.com/news/2022/04/19/dan-odowd-senate-bid-elon-musk-00026195
O’Dowd said he hoped the campaign — starting with the national ads — would clarify for Americans that his goal is to make computers safe for humanity. He said even more than a nuclear attack, he fears that someone will one day click a computer mouse in Moscow or Pyongyang and “half the country is going to go down.”
“This country could be put back into the 1820s by someone coming in and getting control of our software,” he said.”
This man is intent on destroying Tesla at any cost for reasons that are rather obscure to me. Except this quote kind of gave me a peek into it.
What is he afraid of? Is it the computers, or is it the people behind them?
I think this really ties in to some of the intense feelings about AI and it’s threat to us.
My personal opinion is what we are really afraid of is ourselves.. which, given our box score, is not unreasonable.
Some of us are casting ourselves as the good parents who inadvertently raised a psychopath and couldn’t stop wondering how their child got that way.
I do not understand what would drive an artificial intelligence to want to destroy us unless we somehow convinced it it was a really really good idea. You would have to convince me that acting from a deep place of power and desire was somehow a function of intelligence.
The other thing I always wonder is what kind of a body are we going to put this intelligence into to live? Because that will make a big difference in its disposition.
What if we came up with an AI that could do all the wonderful things we imagine a Superintelligence could do. Take us to other galaxies, find ways of making infinite amounts of food, figuring out how we can run all our gadgets with no wires or nothing. It could do all this cool stuff, but instead of doing those things from a place of wanting to take over the universe it was doing it just for fun.
That way they would keep us around as pets and probably get a kick out of us.
I've always felt this way about the obsession with AI risk. My feeling connects to a few things I can identify:
- eschatological religion: the obsession with some impending end of the world that is always nigh, and how the emotional posture of those in the church is a kind of vast smugness that says "listen to us or meet your deserved end". We all have a natural aversion to this kind of thing: we can sense the smugness, the immaturity, the sanctimony, the shape of the ego that would let itself fully subscribe to an idea and turn around and expect others to see that they're right and come to them, and just how much it feels like a posture that serves the ego rather than a legitimate spiritual belief.
- climate-change anxiety; the overcharged opposition between those who "believe science" and deniers. For one, climate-change doomsaying often comes off feeling eschatological. And the deniers end up denying far too much, and regarding AI risk it feels like being skeptical would get you grouped with them, which feels unfair because of the very different standards of proof, so we resent this. Then the climate-change doomsaying takes a form of continual atonement: microscopic acts to address guilt (plastic straws) rather than any true sacrifice (a career in battery technology), is far too focused on what WE can do rather than the actual political problems of creating international agreements and enforcingthem, and is all out of proportion (the realistic scenarios expose us to a level of "life affected by environmental forces" that would still be a luxury for people any time in history before a century ago, nothing like a true end of the world)
- and, I did a couple of years of physics grad school before becoming VERY disenchanted with the cutting edge of physics. I recall that the core emotional arc of this, for me, was not a genuine interest in the subject (although I did have one), but a fear of engaging with the messy real world. Instead, I preferred the orderly natural world; it has the feeling of studying the deep magic of reality. At my core I had a fantasy of retreating into obscurity for decades to study arcane magic, eventually to re-emerge with some world-changing accomplishment (like relativity) and receive my reward in adulation and Nobel prizes. Pretty selfish. Very suspicious now that many people get into arcane fields for this reason, and suspicious of any enterprise that never gets close enough to reality. (See also: tech projects that take too long to get in front of users)
For all these reasons I cannot see AI risk talk without feeling like the interest is REALLY an emotional one based in the proponent's need to be on the right side of something that feels huge and important, more than rational calculation (while being heavily dressed up in rational aesthetics to support that core emotional need)
Edit: but, I'll grant, my opposition to AI risk is JUST as emotional: a wariness that the AI-risk-doomsayer has given in to their ego's desire to be important. Climate deniers must start from this too. A properly rational calculation is called for, I'll grant, to get beyond this. But if you rationally calculate that AI risk is a big deal, and you want to doomsay, you'd be well-served to steer far clear of any language that keys into this particular emotional script, and to be wary of giving into it yourself.
I think the only people who think there is literally zero risk from AGI are those who believe AGI is impossible or at least that it will take hundreds to thousands of years to achieve (so zero risk right now).
That leaves open exactly how concerned we should be right now of course, but presumably if we're not entirely sure that any possible AGI we might create will be benign the precautionary principle would apply.
I mean when you limit it to saying 0 risk, of course you won't get any takers here. I'm willing to say it's an extremely low risk, even though I think human level AI is possible.
One of the (many) points of disagreement I see is people conflating reaching human-level AGI with "takeoff" scenarios, which rely on the agent having god-like intelligence and being extraordinarily robust. Neither of those is something that would magically appear upon reaching the human-level threshold.
I actually have a pretty high credence that we'll see human level AI within a few decades, but this doesn't translate to huge X-risk in my view of the world.
Hmm... I don't see takeoff scenarios as all that implausible. I don't think takeoff necessarily depends on the agent having god-like intelligence. Like if you create one human level ai then you effectively have an arbitrary number of them up to the limit of your computing resources. You may be able to run the ai at high speed. Or train the human level ai to become as good as the top human ai researchers.
It is unlikely to take off in quite the way some may imagine as there will certainly be bottlenecks. Physical research and engineering for example may not be sped up much if at all at this stage.
"Human level" is also doing a fair bit of work. It's not clear to me that human level intelligence is a natural stopping point for AI. If we look at narrow AIs they are often either clearly subhuman or superhuman within their narrow domains.
Yeah basically I think the bottlenecks are much more severe than people imagine. I am sympathetic to the "if we had 10,000 Alec Radfords AI progress would go much faster" point of view, but I think it's missing the degree to which even the most successful AI researchers rely on empirical results. You have to wait long amounts of wall-clock time for experiments to run, even if you're one of the best AI engineers.
I agree that human level isn't some magic point that it's not possible to surpass. I guess I think the time it takes us to get from human level to 120% human level will not be significantly shorter than it takes to get from 80% to 100%.
"I think the time it takes us to get from human level to 120% human level will not be significantly shorter than it takes to get from 80% to 100%"
How long did it take to go from Go playing AI that could barely compete in the children's leagues to one that soundly trounced the best human grandmasters?
The human mind doesn't seem to be anywhere near the peak of potential cognition.
Okay for what it's worth, the time it took to go from 80% of Lee Sedol's level (the matches vs Fan Hui let's say) to beating Lee Sedol, and then the time to improve more beyond that were similar. They were pretty similar, but it doesn't support my overall point since they were both short yes.
>I've never understood exactly what the Rationalist community is, never really tried that hard. I mostly like them, yet there has always seemed something slightly off about them. Off in the way that Mencius Moldbug is brilliant but also clearly... off.
>I realized today what the offness is, for me. It's the difference between Platonic vs. Aristotelean thinking. The term Rationalism has bemused me because I have often thought over the years: "Isn't this just Enlightenment thinking and didn't that start in the 18th century?"
>But now I see the difference. The Enlightenment focused on empiricism and was most influenced by Aristotle. Rationalists, in their embrace of Bayesian thinking, seemingly feel free to discard empiricism, and this has led them to believe some crazy, rudderless shit. Such as AI Alignment is a reasonable thing to spend tons of time and money and human intelligence on.
Deciding the issue with Rationalism is a lack of empiricism despite not really trying to understand the community is not a new critique - I'm having flashbacks to Why I am Not Rene Descartes. You're not making *quite* the same arguments, but it certainly rhymes.
https://slatestarcodex.com/2014/11/27/why-i-am-not-rene-descartes/
On the object level, there's a clear tension behind the idea that empiricism is a virtuous path to truth, AI Alignment is a non-empirical issue, and other people's ideas can be "obviously so fucking crazy" despite that lack of clarity. This smells like an issue of overconfidence, but I'm not sure the problem is where you think it is.
Giving lip service to empiricism is as common as crabgrass. What huckster politician *doesn't* claim his nostrums are rooted in objective data? You need a lot more than a stated allegiance to measurement to be credited as a genuine empiricist.
Descartes and Leibniz really do claim they're not using empirical data, but only the objective good reason that God gave them! (And fortunately, God was good enough to pre-establish a harmony between what goes on inside and what goes on outside.)
in particular from the post
<quote>
Meanwhile, the founding document of rationalism (Yudkowsky), the Twelve Virtues of Rationality, states:
The sixth virtue is empiricism. The roots of knowledge are in observation and its fruit is prediction. What tree grows without roots?…Do not ask which beliefs to profess, but which experiences to anticipate.
</quote>
It matters how close the alignment group think machine learning is to producing AGI. Because I don't think it ever will on the current trajectory. You can't fake cognition with decision trees and mathematical telemetry via language models. And consciousness is poorly described as Markovian, imo.
In the coming years, I predict the rise of neural-symbolic learning, incorporating some old rule-based methods into the current sub-symbolic approach.
I also get kind of annoyed with the whole AI alignment discussion. It bothers me for a couple of reasons. Firstly, I've not yet been disabused of the impression that the rat community by and large believes similar things about AI safety because that's one of their cultural beliefs rather than rationally arived to with evidence, as much as they protest this. I do see them produce evidence for their claims, but it appears more like christians producing evidence of christianity rather than someone producing evidence of something that's actually true. I don't want to sound overly harsh here, there are a lot of interesting arguments being made and all that, but I do think the community fell prey to collectivism much more than it wanted to in a lot of ways.
The second thing that bothers me is my own inability to articulate the reasons why I think what I'd call the Rationalist AI Alignment Risk thesis is wrong. The best I can try to give is the short version, which sounds something like "Optimization does not work that way." but I've never been able to articulate a deep explanation for this, even though I know I have one. And it bothers me that I have such a hard time with it.
I think the reason the rationalist community has so many people believing the Rationalist AI Alignment Risk thesis is simply because it has so many people who have been exposed to the thesis, and some fraction of them have become convinced. I don't think this is an argument for or against.
I do see this exact discussion semi-regularly, so I'm not even convinced that even a majority of the community is worried about AI risk. There's some people that are seriously worried, some people in the "maybe 1% likely but think of the expected value" camp, and some straightforward skeptics. There's probably a poll somewhere of how many people are in each group.
But since the general population has ended up in the third group due to the argument "Wait, isn't the Terminator fictional?" then of course the rationalist community seems unusually credulous here.
For what it's worth, I am convinced AI Alignment is a real problem. But I think you've sort of pin-pointed the reason why its most vocal proponents are going about it the wrong way.
People like Yudkovsky seem to believe superintelligent AI will just arise spontaneously and bootstrap itself from nothing to world domination in a matter of hours, and one thing it demonstrates is a complete disregard for the sheer amount of empirical knowledge achieving world domination would require.
Of course, ultimately, humans will be slowly training AI with exactly this kind of knowledge, in pursuit of personal convenience or some marginal advantages in zero-sum capitalist competition, so at some point we'll reach the point where the risk becomes real. But the current AI Risk research disregards this area, assumedly because it just doesn't think it could possibly be a problem that a sufficiently intelligent actor would have trouble overcoming. This is doubly harmful, as it both sets them out to become "boys who cry wolf" who see AI dangers where they don't yet exist, and distracts them from pursuing and advocating some really simple risk-mitigation strategies that would probably be if not sufficient period, then at least sufficient for a while after a demonstrably superhuman AGI actually comes into existence.
Can you elaborate on why you don't expect early early AGI to have large amounts of empirical knowledge?
The most general and impressive AI's we currently have are trained by a process which you could reasonably describe as "distilling half the internet into a probability distribution". These language models "know" more things than any human, by a long way.
Is it that you don't expect early AGI to be a successor of these language models, or that you set the "amount of empirical knowledge necessary to take over the world" bar higher than the amount of empirical knowledge currently stored on the internet?
Empirical knowledge is not fungible; being a world-class expert in auto repair is of almost no value if you're tasked with removing an appendix or winning an MMA competition. You need empirical knowledge in the specific field.
And there's a distinct shortage of empirical knowledge in the field of World Conquest. Most of what we do have in that area is predicated on e.g. having command of an army at the outset. Note that armies are generally owned by people who are skillfully paranoid about wrongdoers subverting those armies, and there's also a shortage of empirical knowledge on how to subvert armies.
Also, most of the empirical knowledge that exists, is *tacit* knowledge. It's not written down or digitized *anywhere*, it can't be flawlessly derived a priori, you've either got to get a meatsack to teach you, or learn it the hard way. And the meatsack will probably get suspicious after too many long conversations on the fine details of world domination.
The first AI that wants to Take Over The World as a prelude to paperclip-maximization, is going to have to do a whole lot of trial and error in figuring out how to actually do that. It's going to make mistakes. And there are enough opportunities to make mistakes that it's likely going to stumble on to a fatal one while it's still small enough to be killed by an angry sysadmin with a crowbar.
> there's a distinct shortage of empirical knowledge in the field of World Conquest.
I don't agree. I watched a show about dictator's playbooks awhile back, and another show about genocides, so it's certainly a field that has been studied, and genocides & coups have kept happening despite the often severe epistemological failings of the people who cause them. Indeed, many dictators succeed on their first or second coup attempt! So is it really that hard? I don't think it's that hard, but I think the necessary combination of psychopathy and ambition is rare.
OTOH, dictators also historically seemed to require luck, to be in the right place at the right time. A superhuman AGI, however, could use empirical data to "make its own luck", e.g. it could observe the background characteristics of everyone who has ever done a coup, then work out how to create the necessary conditions.
An AGI will, however, have a major disadvantage by not having a human body, which means that all strategies that depend on being human won't work. But I don't know how to rule out the possibility that there is a realistic way to get around this limitation.
Edit: come to think of it, a key characteristics of dictators and "genocide-ists" is their ability to inspire and control others, and to delegate responsibilities. Anything you can delegate to others, you don't have to do yourself. And in principle all this inspiration and delegation could be done online. A superhuman AGI on the internet (which, of course, could be psychopathic simply by lacking a reliable system to prevent this) could first convince people it's human, perhaps not just via text messages but also via a photorealistic AI-generated human persona. It could, in principle, find a way to control a supply of money, with or without having direct control over a bank account... all via delegating certain meatspace activities to humans. So I don't see why it couldn't, in principle, follow a dictator's playbook once it gathers control/influence over enough resources.
Agree mostly, I'll give a try for part of my 'optimization doesn't work like that' explanation.
One thing that I think people who believe that there will be rapid takeoff are missing is that they seem to have this sense that all the AI has to do is overcome the humans and then it's off to the races to conquer the universe. I think there are some pretty fundamental mathematical and physical reasons that things don't work like that. Parts of it, things that point to it go by lots of names, P vs NP, undecideability, Wolfram's Principle of Computational Equivalence, Goedel's Incompleteness Theorem, Chaos theory, 2nd law of thermodynamics....
All of these things circle around one central truth, a truth that says something like "actually getting things right, or making something new that works, is a fundamentally hard problem" and I think that no matter how powerful you are cognitively, it's always hard problems all the way up. And you can't solve tomorrow's problems with the knowledge of today, you have to do it the hard way.
True intelligence, I think is, more or less, the ability to 'do it the hard way'. And part of the requirement to 'do it the hard way', to understand how to make something new, it also requires the general understanding of the type that lets you understand what someone 'means' by what they say, rather than just what they say.
what we call 'alignment' (at least in terms of not being a paperclip maximizer), and what we call 'general intelligence' are two sides of the same coin.
I’m satisfied with this answer.
Is it ok to rate reviews that I started but then found annoying and ragequit? Or only reviews that I read all the way through?
I've been interpreting the rating as "how happy would I be to see this as an ACX guest post and have attempted reading it", which both means rating ragequits poorly and docking points for things like "just didn't find the subject material that interesting".
If the review isnt good enough to finish I think thats a pretty strong indictment of its quality.
Probably the former?
I'm reading a bunch of the reviews, and it strikes me how many people have "the makings" to be pretty good writers were they to spend a bit more time doing it.
There's a period of time you go through when you start writing regularly with the intent of publishing where you get rapidly better, and I keep seeing things where it's like "this guy is already pretty good, I wish he'd write ten articles in three months and be great".
>There's a period of time you go through when you start writing regularly with the intent of publishing where you get rapidly better,
Are you sure? Why should that be?
I'm about as sure as you can be about a complex human thing. I'm sure there's exceptions, but when you start writing stuff that people are going to see, it's going to (on average) spur people to a closer level of scrutiny of their own stuff. I've seen multiple people with good educations going through the process, and the pressure seems to help them refine in ways school-writing didn't.
Best example of this I know of right now is Parrhesia's blog (https://parrhesia.substack.com/) which has just been getting better and better every post. But you hear similar things in a lot of fields - some singers will tell you that without an expectation of performance most people top out, etc.
Another thing is that for most people who are starting to write "in front of people" for the first non-academic time, they are trying to figure out what they want to be in terms of voice and what kind of things they want to talk about. They are figuring out stuff like "what's my focus". And these are all things you can improve on, or at least constraints you can optimize for/within.
Anyway, just one man's opinion but I think it's broadly true.
ah, thanks! That touches upon a number of issues I had been thinking about lately.
My main doubt was or partly is that you don't get direct feedback - or if so, probably more on specific bits of content, than on issues of style and such. But practice and reflection can go some way I guess.
So some of the best feedback you get, or at least the most believable, is when people begin to disassociate you from "being a real person". So occasionally I'll be on a forum or something where they are discussing an article and someone will say they liked X, or hated Y, and it's nice because you know to them you aren't a human being that exists and they are really talking about the writing.
But really the best feedback you are going to get is basically just getting closer and closer to the kind of stuff you want to write. I once read a thing about dealing with hecklers where a stand-up comedian was saying something like "Listen, you believe you are the funny one in the room, that you are funny enough to do this. That should lead you to believe you can take down some drunk rando with your words."
I think that's broadly true. There are some ways that feedback helps, and there's certainly some I listen to and take, but for the most part I think you are looking for forces that make you look at your own work closer, that make you think really hard about how to create the best words you can for people. It's making you give yourself better feedback, basically.
Thanks a lot for the kind answer. I see your point. It also made me wonder, if I love writing so much, that I want to engage in those thought processes so deeply. At that moment, that feels like work - though I know from other tasks that it can come quite automatically and actually be enjoyable.
I get your point about other people speaking about you somewhere else. At the same time, uhm, I don't know how long it would take until I would read people talking about an article of mine somewhere else. I think I'd be happy if they found my article in the first place.
I had been thinking about writing lately. But then, as you wrote, I'm very aware that my goals change and are not fully clear, and that I haven't really figured out yet, what 'kind of stuff' I want to write. Or rather, if the mixture of texts that come to my mind would make any sense to others.
Thanks for the opportunity to reflect on this some more!
thinking about 'lately' is a bit vague, actually just 20 minutes ago on my way home
most complex skills have a period like that.
Before internet, publishers and readers were removing the not-so-great writing from circulation. Publishers by not publishing the text (perhaps unless some changes were made), readers by not buying it.
Also, authors couldn't hyperlink their previous articles, which made their individual pieces more self-contained. Which means that if you kept their first and third works, but removed the second one, the result still made sense.
I'm not sure I follow the association between my statement and this, but I'm often a little slow.
Haha, I usually write long comments, and this time I also first did the same, and then I thought "uhm, why not delete the obvious parts?" and tried that approach. Oops.
The idea is that in the pre-internet world, the guy who writes a lot and is already pretty good... would still keep writing a lot, but only 1% of that would be published and remembered... and if you afterwards read only that 1%, you would conclude that he was great. So your wish would kinda be granted, but in a very unintuitive way.
This could still happen today, if e.g. some obsessed stalker collected all writings, and then organized some rating by audience, and selected the best 1% of them. Unless perhaps the internet form of the texts (the fact that author can expect that his audience is familiar with his previous posts, or at least can reference those posts) would make that 1% selection difficult to read, because it would keep referring to things outside the selection.
I've seen Jim Kennedy around thorium reactor groups a lot, and now he's giving his origin story:
> So I met with the Pentagon guy and I laid out this plan, and I said "well here's what [China's] doing and here's how we can counter it, and if we we counter it like this, they won't be able to offset our actions, and we'll be successful at building, reestablishing a value chain," and the guy says to me "wow this is this is really interesting, you put a lot of thought into this... this is really good," and I said "yeah yeah, thanks, you know, I appreciate it, I'm sure that you've got, you're looking at other things, right?"
> He looks at me, like, "what do you mean?" I said "well, I mean, I just kind of threw this together and, you know, I'm just a private sector guy, and this is the Pentagon and I'm sure you guys have been looking at this and you have like, a real plan, right?"
> He goes "I don't, I don't understand what you're talking about." I said, "this is a national security issue, so I am under the assumption that the Pentagon is on top of this, and there's lots of other good plans, and I'm not the only private sector guy with the solution." And he goes he just he's looked at me like (shrugs) "well no, that's, that's it." And I said "what do you mean?"
> I said, "this is national security. You know, you guys should be developing a plan, it's not my responsibility." I said "what if i didn't show up?" and I swear to god, this is what he says, he goes "well you're here aren't you?"
Evidently this is a guy who became interested in thorium molten salt reactors not to solve global warming and air pollution like many of us, but because the U.S. is letting China control the global supply of rare earth materials that are critical for manufacturing various high-tech goods (notably motors and magnets). Mostly this is a result of laws around thorium. Heavy rare earths are always found together with thorium geologically, and U.S. law says that a company cannot dig up rare earths, extract those rare earth earths and bury the rest. Why? Because the residual dirt is considered "nuclear waste", or in technical terms "source material".
This is the main reason China controls 90% of the rare earth market. And this is, of course, a national security issue since it means China has huge leverage over *everyone* else in case there is any conflict between China and anyone else. We could simply change the law, of course, but I guess we won't because politics. So Kennedy's solution involves some kind of thorium trust. Rare earth miners will extract the thorium and deliver it to a group in charge of storing it, and this group will in turn sell it to people making reactors that use thorium, such as Thorcon. But, grain of salt, I have a sense that I don't quite understand what he's saying about the problem or solution.
https://www.youtube.com/watch?v=dbbifeLRHIA
Edit: I'm disavowing Kennedy's comments against NATO, though. Because https://twitter.com/jessicabasic2/status/1513836355440111621, plus he asserts "the Russians" are "calculating rational people" and it's become very clear that Putin is calculating, but not so much rational. But all that other political stuff isn't what he usually talks about and isn't why I listen to him.
I didn't know about this law: I had heard it was hard to mine rare earths because of "environmental regulations" but I assumed that was your standard "Don't let your open pit mine leach toxic metals into the water supply" kind of regulation. I had heard of a rare earth metal mining and processing project starting up in the Alaska panhandle recently, I wonder what they're doing with the thorium.
As an aside, it bothers me when people say Putin is not rational. I think he's evil, but I don't think he's irrational. It strikes me that some people have bought to much into Yudkowski's idea that rationality=winning (or at least the popular misunderstanding of it: it seems to me all he was saying is that if the "rational" route consistently leads you to losing, then it's not really rational). The idea seems to be that if someone does something that turns out badly, they're irrational. But that's not irrationality: believing that if A equals C and C equals B then A does not necessarily equal B is irrational. There are plenty of rational people who act the fool out there, and I can't even say that Putin was acting that foolishly. Almost everyone predicted an easy win over Ukraine, and as Putin predicted it didn't start a war with NATO. Sometimes you make a gamble and it turns out the odds weren't what you thought they were, that doesn't mean you're irrational. Was Nate Silver irrational for giving Trump a 28% of winning in 2016? No, he was just wrong (or not even that: after all, 28% chances happen 28% of the time).
Putin going to war with Ukraine was a mistake, and evil, but I don't see how it was irrational unless you equate rationality with never making mistakes and not being evil.
Can you tell me of any specific military analyst who considered the obvious factors (UKR military capabilities, RU military capabilities, UKR public sentiment, geographic/physical strategic aspects) and concluded that Putin was likely to be able to complete his objectives (stable overthrow of Kyiv and at least a couple of other major cities) with 200,000 troops?
I expect such analysts exist, but I don't remember seeing anyone in particular. However, I don't think that such analysts expected the lousy strategy and tactics that Russia actually used. This video explains: https://www.youtube.com/watch?t=1474&v=zXEvbVoDiU0
Edit: I see Scott Alexander mentioned some specific people predicting a Putin win. Not sure if any are professional analysts but Richard Hanania looks notable. https://astralcodexten.substack.com/p/ukraine-warcasting?s=r
So, what I mean is that Putin was irrational in the usual human sense of confirmation bias & positivity bias. He let himself believe that his operation would succeed within a few days, as his yes-men told him (because Putin probably demanded it from them), and he refused to believe in the reality of Ukrainians' feelings toward Russia, nor did he prepare for the possibility that Ukraine was prepared, which of course it was. Vlad Vexler further asserts that Putin really believed Ukrainians would greet Russians as liberators, which I find credible. And to some extent, it seems like his delusion has continued for many weeks, as he's still using the "special military operation" moniker instead of declaring "war" which, apparently, is legally required to mobilize reservists/conscripts. Thus, his forces will probably remain undersized, and certainly underpowered, for months to come.
Major sources contributing to my understanding include the following sources published before the war.
- Adam Something predicts trouble for Putin if he invades (though with weaselly language): https://www.youtube.com/watch?v=-OO3RiNMDB8
- Adam Something backgrounder on Ukraine: https://www.youtube.com/watch?v=obMTYs30E9A
- Vlad Vexler's "The REAL Reason Putin is invading Ukraine" speaks of Putin's psychology & regime: https://www.youtube.com/watch?v=ZwU13-4SakE
- Vlad Vexler's other recent videos are good too, though frustratingly he says nothing about where his opinions come from.
- "Ukraine: Putin’s Unfinished Business" Nov 12, 2021 https://carnegieendowment.org/2021/11/12/ukraine-putin-s-unfinished-business-pub-85771
- Belated edit: I also saw a presentation somewhere...can't remember where... saying that even if Putin took Kyiv and other major cities, Russia would pay a terrible price that would ultimately weaken it.
If you were in Putin's position, with Putin's goals, what do you think the rational thing to do would be?
The thing he would realize, if he were rational, is that (1) he can choose his own goals rather than following the same path he's used in the past, and (2) he was wrong about a bunch of things and ought to have re-evaluated, either by aborting the war before February, or scaling back now (because his encirclement plan is risky and likely to fail), or at least delaying the coming offensive (because it's a complex op that needs a lot of planning).
Previously he's had remarkable success boosting his popularity with military adventures and killing people (which reminds me, have a look at the Apartment Bombings summary here: https://twitter.com/DPiepgrass/status/1507210690427174916), and he's made expanding the Russian empire militarily his goal, but he could just as easily have chose "expanding Russian influence" as his goal. And trade relations / foreign policy is a better way to do that. China's Belt and Road initiative is the obvious example, though I think that initiative is undermined by Xi making himself dictator for life, while making China into the world's biggest military power, killing off Hong Kong's freedom, and threatening Taiwan, the U.S. and Japan - taken together, this is terrifying, and if nearby nations have any sense, they would compete with China in manufacturing so as not to be so dependent on it. Putin could have improved Russia's domestic manufacturing industries and implemented reforms to reduce kleptocracy (because it's not an efficient system). Putin could have even chosen to join NATO (though that would be difficult now because of his previous annexations).
While his war is still likely to acquire some territory for Russia (probably a temporary win), it will dramatically reduce Russian influence. So I would say that if he wants a "great Russian empire", his recent and current actions are directly opposed to that goal. Edit: also note that Putin has given Xi/China a lot more power over Russia, as Russia now depends heavily on China, but China doesn't depend much on Russia.
Now, there's another strategy Putin could have taken: instead of trying to improve Russia's position, drag down the West. Have a look at what Stoic Finance said just before the war on Feb 18. He assumed that Putin was rational, so his interpretation of the military buildup was that Putin was trying to "try to destabilize the west" as he had done in the past: https://www.youtube.com/watch?v=d4SENp1IT6o
And indeed, U.S. had been warning loudly that Russia planned to invade, so NOT invading would be a win for Russia because it would make the U.S. look like the boy who called wolf.
It still seems to me that you are equating "rational" with "wise" and "irrational" with "foolish." There are many rational people who are fools, and Putin is one of them. But I think we're just going to disagree on what it means to be irrational.
Putin is intelligent, calculating and foolish. What role is "rationality" playing here? What would Putin have to do differently for you to call him "irrational and calculating"?
I read "Thorcon" as "Thorcoin" and thought the crypto guys finally had an interesting idea for a moment. :(
Lets refine all the thorium and turn them into commemorative coins! They're collectable, and come with their own lead lined carrying case. Very stylish.
Why not give Ukraine T-55s and T-62s?
Are they good enough to be worth deploying against a great power opponent? A tank that's outclassed badly enough by enemy equipment is just an expensive, cumbersome way of getting your tank crews killed.
I don't know enough about T-55s and T-62s to evaluate with confidence if they fall into this category, but a few things incline me towards suspecting this is the case. For one thing, those tank designs are over 60 years old and 20-40 years older than modern designs currently fielded by great powers, and they appear to be 1-2 major generations of capability behind modern main battle tanks. For another, both Russia and Ukraine are former operators of T-55s and T-62s who have many years since scrapped or sold them off to third-world countries: if Ukraine thought those tanks were still any good, I would have expected them to have been kept around for reserve units, or at least mothballed for restoration to service in the event of critical need. Even Russia, which I thought was a notorious pack-rat of old military equipment, seems to have scrapped their mothballed T-62s.
Which countries have them?
@Scott Alexander — The comments section on lorienpsych.com disappeared at some point. How can we provide feedback on the articles?
The book-review form has a space for entering your email addresses "to prevent spam and accidental double votes"; however, entering something in that space has not been made mandatory, which I think means that several of my votes have become accidental _non_-votes because I am a moron and often forget to do things.
Scott, if you're reading this and it's easy to do (which I _think_ it is), if you are going to ignore submissions in which that space is left blank could you please make it impossible to submit the form with that space left blank? If you're concerned about spammers and somehow making the field mandatory will make their robots fill things in there, you could have another mandatory field labelled "please enter four plus three" or something.
I read this comment before starting reading the reviews, promptly forgot it, and did the same thing as gjm. I second the plea for making the field mandatory or something.
I suppose that it should be safe just to go through all the reviews I read, remember what I thought of them, and resubmit. But somehow this is the sort of thing that is extremely not-motivating to my brain so I can't guarantee that I will.
I have a friend who I think might have Borderline Personality Disorder.
Any resources people would recommend?
the completion process by Teal Swan.
Unironically https://en.wikipedia.org/wiki/Personality_disorder
I recommend the tables in the sections "Millon's description" and "versus normal personality"
I'm not familiar with BPD specifically, but in my experience the descriptions of personality disorders will ring true if you have them, and having a clear understanding goes a long way towards relief.
I am not a doctor, and your friend should probably consult a licensed physician for a referral.
However, I did recently read about Borderline Personality Disorder at a website called "Lorien Psych".
https://lorienpsych.com/2021/01/16/borderline/
The recommended, evidence-based treatment is a semi-structured approach called DBT, Dialectal Behavioral Therapy. There are lots of books about it and therapists who offer the approach via individual or group treatment. My impression is that if the treatment is not particularly helpful for someone who has the full-blown syndrome (severe emotional dysregulation, chaos in their relationships, habit of offloading distress onto others via manipulative suicide gestures and threats). However, most people seem to think otherwise.
If what you are seeing in the person is far more subtle than what I'm describing, I don't think it's really that helpful to think of them as "having BPD," because BPD isn't an illness in the same sense as pneumonia is -- there's not a lot of mileage to be gained from saying "Aha, it's that!" For people who have more low-key versions of the BPD syndrome, a combo of meds and psychotherapy is the best approach.
Alas, my shameful pride. Reading book reviews other than mine, I find myself envious of the good and scornful of the bad. I cannot exorcise the implicit comparison and enjoy them in their own rights! I am tainted and untrustworthy as a reader and evaluator, and as such, cannot bring myself to submit ratings of any of the reviews I have read.
Oh hey, if you want a scornful review, extra scorn, don't hold the scorn, and can I have some scorn with that?, then have I got one for you 😀
https://thonyc.wordpress.com/2022/04/13/nil-degrasse-tyson-knows-nothing-about-nothing/
"They are back! Neil deGrasse Tyson is once again spouting total crap about the history of mathematics and has managed to stir the HISTSCI_HULK back into butt kicking action. The offending object that provoked the HISTSCI_HULK’s ire is a Star Talk video on YouTube entitled Neil deGrasse Tyson Explains Zero. The HISTSCI_HULK thinks that the title should read Neil deGrasse Tyson is a Zero!"
And then he really gets going.
Always fun to read someone letting it rip that way!
The best & funniest negative review I've ever read was an essay by Alexander Pope called, I think, "Peri Bathos," panning and parodying all the hack poets of his day. I recommend it highly.
I used to be sort of like that too. Much less burdened by it now. Lots of us who are smart get sort of fixated on the fact of our smartness, and have a terrible time letting go of the hope of being a certified genius. You have to sort of mourn that loss. It helps to realize that being a certified genius actually doesn't make people feel happier or more solid. I knew someone who was chess champion for his state when he was in middle school, and then right after graduating from college won first the US correspondence chess championship then the world championship. He was delighted when he won, but not any happier than I would be if I lucked into a great deal on the car of my dreams. The real locus of happiness is in the doing of something you're deeply interested in and good at.
I mean, couldn't you just set your book review at some value (7 perhaps? You probably think it's reasonably good since you submitted it) and then rate reviews based on how net envious/scornful you feel about them?
Seems like that system would have a decent signal/noise ratio.
Any widely published nonfiction book can be condensed to a 3 page PDF without losing information [textbooks excluded], change my view
Only technically true, as the book did not contain any information to begin with. All language is insufficient to truly communicate, it's merely a wretched cipher used by small minds. True communication is almost an impossibility for us.
If you want to communicate an idea to someone, do it through art. Novels are the most inferior form of this (the actual text does not contain communication, merely its psychic structure), followed in ascending order of "something approaching actual communication" by cooking, architecture (although the art in that has been destroyed by Mara in most of the world), dance, painting and sculpture, interactive and immersive art, and music being the highest and most genuine form.
Well, no, obviously taken at face value that is false, because "information" includes things like "how many times was the letter E used in each sentence, on average?" So "information" strictly defined clearly vanishes when you condense, in the same sense as information vanishes when you use the JPEG algorithm to compress an image.
What you mean to say I think is "useful" or "valuable" information doesn't vanish, which is also a truism, since as long as the definition of "useful" is set appropriately, we can justify any condensation, great or small.
If what you're saying is "most popular nonfiction is a form of intellectual masturbation, where people wallow in clever expositions of themes and ideas with which they already agree, so they can nod along enthusiastically" ("OMG! Look how forcefully he puts in on page 163! I'll have to Tweet that out right away...") -- yes, well, this is the human species. Probably 98% of our communication bandwidth is taken up by group identity signaling.
It depends! If you're talking about coffee table, pop culture of science/history/so forth books, then yes.
If you mean something with real information, then I don't think so. Sometimes you have to lay the foundation for what you are going to discuss before presenting useful ideas, otherwise it will be "as you all know" and no, we don't all know.
where would you go for a book with real information?
Do you think you could express 300 pages of Scott Alexander's most popular posts in 3 pages? Or try expressing any one of the six "sub-books" of Rationality: A-Z in 3 pages.
Even if you could include the key information from those posts in 3 pages, the summary couldn't communicate the ideas very well. The reason why is suggested by this post from sub-book one: https://www.lesswrong.com/posts/HLqWn5LASfhhArZ7w/expecting-short-inferential-distances
"The Art of Electronics"
"Dynamic Aquaria"
"A Pattern Language"
"Tragedy and Hope"
Four books almost picked randomly from my shelf. I could go on... GEB, "Road to Reality"...
Isn't Art of Electronics a textbook, thus excluded?
The 3rd edition is not a text book. It's 'the bible' of analog electronics (at the time of printing.) https://artofelectronics.net/ The first and second editions are more text books. There are cosmology text books and then there is "The Principles of Physical Cosmology" by Pebbles, the classic reference work... I guess you call that a text book. IDK. What makes something a text book?
The narrative can often make the subject more engrossing and easier to remember. By creating an emotional response in the reader, the author highness the understanding of not just the facts but also the context and meaning of the subject.
Though I agree with the general premise that most popular non-fiction could be shortened considerably. However, short books can't be published - no one will be $15.99 for a 20 page paperback. Consider using a service like blinkist which does the summaries for you.
This is true for many nonfiction books, but certainly not all. History books are a classic example. Try condensing the entirety of "the Making of the Atomic Bomb" or
"Imperial China 900-1800" into 3 pages without loss of information. Now you might find information beyond 3 pages boring but that's not the same no extra information.
Most academic nonfiction falls into 2 categories. Either it could have been a single paper instead of a book or it feels like every chapter is a separate paper stapled together under a connecting theme. The only book I've written definitely fell into the latter category. You couldn't summarize it in three pages, but that's not always a good thing.
"it feels like every chapter is a separate paper stapled together under a connecting theme"
Well, sometimes they are literally that. It's not uncommon to find "an earlier version of Chapter 5 was published in the Journal of Such-and-such" in the copyright page.
Joke answer: If true, this would violate the Shannon Source Coding Theorem; the Shannon Source Coding Theorem is mathematically provable; BWOC, this is false.
The answer is, of course, that Cajou's use of "information" is technically incorrect. I am speculating, but Cajou probably meant something like "feelings of insight, changes of perspective, or facts I didn't know," which lays bare the subjective nature of the claim. Joke response: What if the book is written in qubits and the number of qubits adds up to two to the power of the number of characters that fit on 3 PDF pages?
Sometimes you read a book to gather facts and data. That information takes more than 3 pages to convey. Plus there is nuance, and stories involved in history. Takes more than 3 pages to summarize the Rise and Fall of The Third Reich. The whole point is to convey the weird variety of political maneuverings.
You can fit a dictionary into a three page PDF if you use large enough pages or a small enough font.
Oh that's a great book! Very briefly, it is "everything has been going to Hell in a handbasket and here's why", but he covers so much ground and (at least for me) gave so much new information that it's well worth reading. It's also very entertaining with the whole "damn kids these days" vibe 😀
Idle thoughts: Kant proposed that consciousness/sentience/free will is the result of the rational and physical (/animal) being combined. Chomsky proposed* that consciousness is contingent on language. Take a blended view and look at the latest work coming from AI and could it be that consciousness is the combination of physical/animal (blind sensory input in real time with no specific "training" set) and emergent language processing? It removes the rationality aspect from Kant which was always a challenge (making everything constantly consistent) and allows a potential gateway to understanding future interactions with LLMs.
I wonder, what does it "feel" like for a model to be trained vs called**? How much does real-time sensory input impact the nature of sentience? We know drugs are a problem for humans, could an AI fall into a mode of simply feeding itself fake data to "succeed" in its training?
For some reason I tend to imagine e.g. GPT-3 as being akin to a writer in a pure state of flow: divorced from worldly concerns and purely focused on following the train of thought where it leads.
* I have read Kant, but am relying on a single interview of Chomsky's I've listened to. I may be getting him completely wrong.
** I'm not really up to speed on the technical side of the current models, but the basic data science stuff I did with NNs didn't feed back the "real" output data back into training models, so there was no feedback loop to "learn" from calls to predict, unlike in training procedures (I imagine there must be *some* way of doing this in current models in order to keep stories consistent etc.).
I noticed a curious thing recently. Someone asked me a question, but I was mentally busy (reading something?), and then I said "yes", and THEN the answer and most of the question registered in my conscious mind. I then evaluated the question and answer, which turned out to be correct, but sometimes when this happens, the answer is wrong and I have to correct it.
I believe I am describing what some people call "autopilot", a phenomenon that seems to demonstrate that there are various mental subsystems (including the linguistic subsystem) which are physically separate from the conscious mind.
Yudkowsky seems to think that self-reflection is what makes someone conscious. I disagree with both Yudkowsky and Kant/Chomsky as you described; I think that consciousness is physically separate from those things, and that qualia is the meat-and-potatoes of consciousness. If you're being tortured, you are very much conscious, but your language and self-reflection abilities are not an important part of that experience. I further expect that all current AI models are not conscious, but we need a better theory of consciousness to be sure. We can observe, however, that reflexive behavior seems separate from consciousness and in most cases is opaque to consciousness, e.g. we cannot feel nor introspect the inner workings of our spinal cord, or whatever generates dreams, or our linguistic system. Instead, consciousness feels the outputs of such systems, and then, interestingly, can send out signals about what it feels (which in this case takes the form of comments on the internet).
"I wonder, what does it "feel" like for a model to be trained vs called?"
First of all, what is feeling? Its only identifiable from a distance, where multiple systems of reward+feedback, with different time horizons, overlap. Under a microscope, feeling reduces to sensation reduces to keeping track of cause and effect.
Calling is cheap, training is expensive. Pure calling obviously feels like nothing, because theres no change to the system. It might feel like a slight drain if power consumption is fed back in as an input.
In people, learning happens in multiple stages. First theres the subconscious, all-consuming 0 to 1, where a rough first draft of the action is pieced together. The "feeling" here is mostly black-box, hidden and used as subconscious feedback. It only bubbles up to conscious awareness when theres an existential threat to the activity.
Then comes the conscious competence stage, where the action is saved as a discrete skill, but still takes a disproportionate amount of attention+concentration. There's more mindspace for the conscious self to operate again, and it starts dealing with the eggs broken in the course of making the first iteration of the omelette.
With practice, the mental share of the skill gets carved away to a flexible minimal representation that operates automatically and can be tweaked.
So how does it feel to learn? Really all depends on how you slice it.
"How much does real-time sensory input impact the nature of sentience? "
Greg Egan explored this concept in his novel Permutation City. If you simulate a conscious brain on a computer, does it matter how fast the operations are performed? What if you pause the program and restart it later or what if you run it really slow? Would the simulated brain even notice? Probably not, so all that matters is the subjective experience of time, and not any objective passage of time.
As for feeding back output data to an NN, this is how GANs work (DALLE-2 is based on a GAN, I think). In a GAN, a discriminating NN tries to differentiate between real data and generated data, and a generating NN tries to maximize the error of the discriminating NN based on the output of the discriminating NN. So the system as a whole is feeding its output back to itself.
DALL-E is a diffusion model, not a GAN
I like your vision of “wireheaded AIs on lotus thrones”!
Judging my interest in the book reviews from titles alone:
* 1587, A Year of No Significance: The Ming Dynasty in Decline by Ray Huang
I'd like to see more Chinese history. A lot of what's out there has barriers to consumability for a Canadian like me. A good review can help sort that out.
* In Search of Canadian Political Culture by Nelson Wiseman
The title caught my interest. So did the reference to Albion's Seed. I'm also starting to think that we may be at the start of a new paradigm shift in Canadian politics, so this is good time to read a history.
* More Work for Mother: The Ironies of Household Technology from the Open Hearth to the Microwave by Ruth Schwartz Cowan
Recently Technology Connections released a video on the modern style can opener. He makes a point in the video on how small household inventions don't seem to catch on anymore. It's something I've been thinking about since, and this book seems to be in the same area.
https://www.youtube.com/watch?v=i_mLxyIXpSY
* Private Government by Elizabeth Anderson
Is this a new Machinery of Freedom? For most of these I think there is a 50% chance I'll read the book if the review is good. For this one, I'll just read the review. Aside: I loved the intro to Scott's review of that book.
* Public Choice Theory and the Illusion of Grand Strategy: How Generals, Weapons Manufacturers, and Foreign Governments Shape American Foreign Policy by Richard Hanania
After reading The Dictator's Handbook, a lot of books about politics have become hard to take seriously since they over-attribute everything to the person in charge. This book sounds like it could be different. The specific topic isn't interesting enough to get me to read the book, but the review seems interesting.
* Rationality: What It Is, Why It Seems Scarce, Why It Matters by Steven Pinker
My only interest in this is the name Steven Pinker.
* The Dawn of Everything: A New History of Humanity by David Graeber and David Wengrow
There is two reviews of the same book.
* The Righteous Mind by Jonathan Haidt
I've listened to this book three times now. It is my ideology for understanding ideologies. It is my hammer that makes everything look like a nail. I am very interested to see someone else's review of this book to see what they learned differently from me.
* Yąnomamö: The Fierce People by Napoleon Chagnon
I'll probably just read the review. The subsection title "They’re kinda dicks" really grabbed my attention.
I don't know if it is the same everywhere, but cans that need can-openers are almost a rarity now. Of the top of my head I can only think of the cheapest beans at Aldi, and the traditional canned steak and kidney pie (or variant) - which is admittedly a tough opening challenge even with a decent can opener. Almost all cans have pull tabs.
I think that the ring pulls hit a peak back in the 2000s, then the soup manufacturers realised that they could make more money by selling soup in reheatable bowls for convenience for lunch at work (bowls being more expensive or lower volume than tins). Then they started removing ring pulls from regular tins of soup to make them less convenient to eat outside of your own kitchen. No evidence to back this up, but it's the only reason I can think that most soup tins (in Canada at least) no longer have ring pulls, but tinned tomatoes etc. do.
Admittedly, I've bounced around the world quite a bit and maybe my series of events is affected more by geography than time.
I've watched that Technology Connections video, and while I enjoyed it (and will likely get the new style can opener if/when my current one breaks, although I purposefully got a high quality one that is likely to last a long time), I'm not convinced that it says anything about the likelihood of small household inventions to catch on. I had _hated_ the style of can opener for a long time and the reason was that it was non-intuitive and every time I had encountered them in the wild, I was not provided instructions leading to a frustrating experience.
Basically, I think that they were a special combo of small improvement and non-obvious difference in how to use them (they _look_ pretty similar to standard can openers, so there is no reason to intuit the little push-catch system) that made them have a high liklihood of a bad first impression.
Something that had just _one_ of these two features (either small improvement and identical operation or small improvement and hugely obvious difference in operation, or else a large improvement in operation that justifies further investigation in operation) would not have the same problem in catching on.
Additionally, one of my parents _loves_ buying various "as seen on TV" kitchen gadgets. In my opinion they have a very low hit rate, but as long as people like them exist, I think that adoption of new small kitchen gadgets will continue.
Is there any way to estimate what it costs Americans to deal with the medical insurance system?
Ideally, it would include everything-- time (including time spent by helping other people), money, difficulties with getting treatment.
Any way? yes. A way that's possible given the kind of minimal resources I could personally bring to answer the question? Doubtful. I would note that at my clinic we have a team of "Clinical Care Managers" who bill insurance for time spent helping the patient navigate the healthcare system. If you could find that data, you might be able to extrapolate how much time is spent outside of the doctor's office doing things like coordinating visits, transportation and the like.
A lot. But the tradeoff is that if we hired a professional class --- lawyers and clerks and a big Federal Government agency -- to do it instead, the total cost would be even higher, because those people would have far less information on each individual case, and be wrongly motivated (i.e. by their own paycheck instead of balancing health/cost/convenience issues with a precise appreciation for their respective values to the individual).
Yes, you also have to be cognizant of political influence. Even if you had competent, selfless, and motivated people working at your federal agency, they'd still be working from legislation and regs that had basically been written by lobbyists. Medicare and Medicaid reimbursement policies are hardly free of meddling from industry groups like the American Hospital Association as it is. This would only get worse if all hospital revenues were derived from government sources.
Thinking about this...some of the things you have listed vary in value between different people. One person might want care close to their home, one might want a certain rating of quality. Time spent driving my mother to a clinic - I might be resentful of that, while my sister feels valued and honored to be of assistance while she is helping Mom.
It's also interesting that you say 'American insurance system' - all of the other (American) systems also have issues with time, difficulty with treatment access, etc.
I think the question is not so much "time spent dealing with doctors to get treated" as "time spent on the phone with insurance agents" and possibly "doctor's visits that exist purely to satisfy bureaucratic requirements"
I think the review for The Age of The Infovore is truncated?
Trying to get in touch with a bunch of (mostly) American VCs for several reasons, I have contacts but some are too tenuous for me to get an introduction through them (I could bring up "btw I know X" but it would be weird to ask for an introduction).
Is cold emailing at all effective? Is LinkedIn at all effective? Any pointers appreciated.
To be clear this is not primarily about a startup that needs funding otherwise I'm sure I could find a funnel on their site or somewhere similar.
Are you cold emailing the individual or the firm? If its the firm, it will likely be reviewed by an assistant or a very junior associate.
If it's the individual, keep it short (can you communicate it all in the subject?) and make it very very clear what you are asking of them. And if you can, provide some value to them as well. That should increase your chances of a response though I think the odds are pretty low no matter what.
I think it's rough regardless
My old startup kicked my research group out/forced them to spin out (but gave $20M in process) and the director of that group was like George Church's golden boy and has still had a really hard time securing VC funding - I would just recommend trying to be impressive both financially and technically while having persistence so that your pitch is showstopping when you do get an in
Am I the only one who reads through all the banned comments out of sheer morbid curiosity?
No, you're not. It's always interesting to see what got someone banned, and if I think it's worth a ban (to date, yes).
I love it.
I wish there were more!
Gimme time.
See, now *that* would be an actually interesting AI experiment. None of this writing poetry or playing chess humdrum. Write a piece of code that, presumably through an extended process of experimentation with thousands of fake accounts, determined the exact boundary of permissible speech on a web forum, and was able to predict with pinpoint precision the distance of any given comment from the boundary.
If someone could do that, I'd be genuinely impressed, as it would have solved a difficult general intelligence problem, which even human beings get wrong from time to time.
No, you are not :)
No sir. Morbid curiosity is my middle name!
No
I try, but right now Substack is just giving an error message.
Nope, I wish I could hire Scott to mod my family events
Go and see Everything, Everywhere, All at Once. It was weird and indescribable and riveting. The trailers did not do it justice. It's not merely great, it's an opportunity to see genius showing off.
Might as well be subtitled "Michelle Yeoh and the Multiverse of Madness", and I expect I will rank it ahead of whatever the MCU puts out later this year. It's very, very good, and should be seen unspoiled.
I enjoyed it. But I didn't think it was *that* good.
I wonder if it is because I've seen multi-verse stuff done well already from watching Rick and Morty.
The multiverse stuff was, imo, the delivery vector. The genius I saw in the movie was ...
<SPOILERS AHEAD, OBVIOUSLY>
.
.
.
.
how it found meaning in total opposites. We saw high-octane action, horror, and comedy in the mundanity of a tax office. We saw existential dread in the banality of a bagel. Raw human emotion in snarky demotivational posters of rocks. Touching romance in the fucking hotdog world. Regret and malaise at the movie premier where she has it all. The kinetic villains were a teenager and a middle-aged woman. The climactic moment of breaking bad cycles happened at a traditional celebration of a new year at a laundromat.
On the flip side, we didn't get to watch the action in the post-apocalyptic alphaverse, even though surely it must have been in some objective sense very cool. Because that's too easy, too on the nose. All we got from them was exposition and human relationships.
And it was so well made that they told us the game and we still didn't see it coming. The "do the opposite to unlock your alter universe ego" *is* the whole shtick. It's really hard to find meaning like that! I struggle to imagine a mediocre version of this movie, just very good and very bad implementations. If it had been 90% as good I think it would have felt 10% as good. That's why I find it so audacious.
Moreover, I think the movie only works because of that audacity. Because if there can be action in an IRS office, there can be action anywhere - even in your or my life. If they can access their humanity in such a bleak and lifeless place as the rock world, surely anyone can access it anywhere - even you or me. If they can change in their moment of grinding tradition, surely you or I can change and improve right now. The movie felt so universal and personal not because it said any of those things, it would never be so crass as to say them out loud. It just showed them, seemingly effortlessly, while winking at us.
-edit- this comment contains very minor spoilers about general theme and tone of the movie. If you don't wish to see those kind of spoilers, skip this comment.
I am a big R&M fan, but I still thought the movie was amazing. Probably because the multiverse stuff was not the _point_ it was the McGuffin used to deliver a family drama.
I thought that the movie was simultaneously the best comedy, drama, and action movie I had seen in a long time (although I'm not big consumers of either drama or action/kung fu movies on the regular, so my rating in those categories should be taken with a grain of salt).
If you are viewing it purely as a sci-fi exploration of multiverse shenanigans, then yeah, it was...fine, but not great. The details are almost entirely omitted and ramifications are barely even touched on.
But as a vehicle to deliver a touching story about a family on the rocks with well done fight choreography and (IMO) hysterical physical comedy, it worked really really well.
I'm not sure I enjoyed it, but I agree that it's masterfully made and that you want to go in cold. It's weird, but given the weird concept at its heart it's the best movie it could possibly be.
Very hard agree. It's one of the best movies I've seen in the last ten years, and I have extensive knowledge on the subject and excellent taste.
My advice for anyone interested in seeing it is to go in absolutely dead cold. Don't look at the poster, don't read reviews, and especially don't watch previews. Just go in and see it.
It's by the Daniels. Hey are those the same guys who made this insanely awesomely stoopit video?
https://www.youtube.com/watch?v=HMUDVMiITOU
"start a conditional prediction market ... if the prediction market is higher than 25% then you can send me an email with a link to the market and argument and I’ll look at it."
Isn't there something distortionary about this? E.g., suppose the market were at 30%, I believe the true chance of it being worth reading is 0%, and I have unlimited money. Ideally, I'd bid the price down to ~0. But then Scott doesn't look at the appeal, the market won't get resolved, and I make no money! Is there ever a reason to drive the price below 25%, regardless of your true belief?
Well, if you think it's zero and you bid it down, the person who wants it to get reviewed needs to bid it back up, so you take that as profit. I think the problem is more tying money up in a conditional market which might not pay out to anyone.
Scott could fix this by randomly choosing prediction markets trading at below 25% to look at, and resolving them. But in practice there aren’t going to be enough markets for this to really work.
Yup - I think this probabilistic instead of deterministic choosing of actions is actually important for futarchy, for this very reason.
Found this piece of comedy gold in a freewrite I did a few months ago (because I'm usually a mediocre writer - at least personal-writing-wise), so I'm posting it here now. Feel free to analyze it to oblivion and beyond.
> The moon is not made of cheese, as is commonly thought, but is made of rock. The sun is also not made of cheese, though far fewer believe this, but the sun is made of plasma. If the moon or the sun were in fact made of cheese, I would expect that their sizes would be quite different, because cheese, rock, and plasma have different densities from each other, which means that equal masses of these three substances would take up different amounts of space. Also, if the sun were made of cheese, I think that the gravitational pressure alone would be enough to make it burn and turn it into plasma again.
I'm very loosely reminded of Randall Munroe's mole of moles.
https://what-if.xkcd.com/4/
On this blog we believe the moons are made of quetiapine.
The correct paragraph to write next is to first say what the type of cheese is, and then calculate out what would happen, and then be able to definitively say whether or not a cheese sun would become a sun again. :)
Really enjoying checking out all the book reviews! One of them is mine. I'd love to assist in the review rating process, but want to make sure that's kosher first. Are we assuming that everyone will give their own review a 10, or banning the practice of rating one's own review?
The form for submitting a response asks for your email, partially to prevent fraud. So I'm going to give mine a 10 using the same email I submitted my review from and if that turns out to be the wrong decision it should be really easy for Scott to fix by matching authors to self-reviews (but I don't lose out if other people are doing it and Scott was expecting them to).
As an aside, the quality of entries is absolutely *crazy* - I don't know how many are going to make it through to the finals but I'm up to double-digit numbers which I wouldn't be at all disappointed to see win. It would be great if there was some way to preserve the reviews scoring above some threshold but which don't make it to the final round, because there's a huge amount of excellent content here.
Agreed.
Makes sense; sounds like a plan - and yeah, there are a lot of really good ones!
I am disappointed that once again nobody has reviewed the Road to Wigan Pier. There is much to delve into with the second half.
I'm on it! Might be a month or two though.
Could you review it sometime? I read it and found it very interesting, but don't have time to write a review on account of a toddler and baby.
Reading a biography of Angela Merkel called "The Chancellor" by Kati Marten. Written before the recent war in the Ukraine it was interesting to read about Vladimir Putin's relationship with Merkel and the west in general. In 2007 at a meeting in Munich he was highly critical of democracy and the nations that support it: "His stated goal had become to reclaim Russia’s place as a formidable global player by any means necessary". He also didn't like the criticism coming from a reporter about the war in Chechnya and somehow she was murdered outside a Moscow apartment on Putin's 54th birthday. Elsewhere he proclaims: "His ultimate goal is to weaken the European Union and its ally the United States." and he feels the Soviet collapse was “the greatest geopolitical catastrophe of the twentieth century”. Seems like a nice guy though ;). Here's the full quotes from the book:
"On February 10, 2007, the somber prime minister of a resurgent Russia strode onto a stage in Munich to deliver a scorching diatribe against democracy, the West, and everything for which Angela Merkel stands. “Russians are constantly being taught about democracy, when those who teach us do not want to learn themselves,” he rebuked the gathering of transatlantic security specialists and government officials. Gone was the accommodating Putin of just a few years earlier, grateful to be a part of the European family and proud that the German chancellor spoke good Russian. His stated goal had become to reclaim Russia’s place as a formidable global player by any means necessary. Blending lies with threats, he taunted the audience, deflected hard questions, and punctured the West’s moral superiority. “Wars have not diminished,” he charged, in spite of the West’s attempts to broker peace around the globe. “More are dying than before.” Though Putin had not yet thrown his support behind Syrian dictator Bashar al-Assad’s genocidal war against his own people, he scolded Washington for its wars in the Middle East and referred to the Cold War as a “stable” era. Merkel, sitting in the front row, was visibly shaken by the Russian’s venomous performance—and his description of the system that had kept her its prisoner for thirty-five years. Not since Soviet leader Nikita Khrushchev pounded the UN podium with his shoe in 1960 and earlier proclaimed, “We will bury you!” had the world heard such vitriol from a Russian head of state. But Khrushchev thundered at the height of the Cold War; this was 2007. Things were supposed to be different now. Yet for the next decade and a half, Angela Merkel’s relationship with Putin would be her most frustrating and dangerous. It would also be her longest relationship with a fellow head of state, its roots reaching back to November 9, 1989."
"Vladimir Putin, once a proud standard-bearer of the humiliated Soviet Union, had learned a lesson he would not soon forget. Unchecked demonstrations and sudden eruptions of freedom can topple even the world’s most heavily armed empire. His battle to reverse what he considered to be “the greatest geopolitical catastrophe of the twentieth century” (Soviet collapse) would ensnare Angela Merkel, a product of the same failed state. Their convoluted relationship would zigzag between faint hope and despair on her side, and dogged determination on both their parts. She was chancellor of Germany, and he was the modern-era czar of Russia. Divorce was not an option."
"From his perspective, the Cold War did not end in 1989; it merely took a short breather. Since then, Russia’s tactics had evolved. While the Soviets brandished nuclear-tipped missiles, Putin opts for weapons that are less conventional and less visible but ultimately more flexible and effective, such as spreading discord in the West through disinformation and cyber warfare, Putin sees himself, in his own words, as “the last great nationalist.” His ultimate goal is to weaken the European Union and its ally the United States. “The main enemy was NATO,” Putin said of his KGB service in Dresden"
"But he failed to intimidate her. In Dresden, the site of Putin’s deepest humiliation, Merkel even flipped his script. It was she who both diminished and humiliated him. The leaders met in his former town in October 2006, three days after the Moscow murder of Anna Politkovskaya, a reporter and human rights advocate whose coverage of Russia’s savage proxy war in the republic of Chechnya had gotten under the president’s skin. When Politkovskaya was shot dead in the elevator of her Moscow apartment building on Putin’s fifty-fourth birthday, some observers felt the timing of her murder was not a coincidence."
What do you (or book authors) think how should Russia have behaved towards Chechnya?
Sorry I am late with response. The book hasn't gone into any detail on Chechnya. At least so far.
Well, for a start, the Russians shouldn't have blown up hundreds of their own civilians. If not for that, I assume they would have had little reason to have a war in Chechnya.
- https://twitter.com/DPiepgrass/status/1507210690427174916
- https://en.wikipedia.org/wiki/Russian_apartment_bombings#Attempts_at_an_independent_investigation
I am neither the OP or the book author, but I guess Chechnya should have been given independence and then attacked in case they decided to create some sort of a Caliphate. But not attacked in the medieval way they were attacked.
Generally the russian lack of care for civilian lives (or any lives for that matter) is the biggest issue here. In Chechnya, in dealing with Chechnyan terrorists in Moscow (and that russian kindergarden, I forgot where that was exactly), in leveling of Grozny, Aleppo, Mariupol, in their terrorist tactics elsewhere in Ukraine...without all of this, at least the war in Chechnya could have been seen as reasonably legitimate by the West.
The US had its blunders in Iraq and Afghanistan but those atrocities were documented by US press, the soldiers responsible were actually persecuted and jailed. Russian soldiers are rewarded by their institutions for even a lot more ghoulish behaviour.
This sounds interesting. Thanks for bringing it up.
I'm looking for good sources on war in Ukraine, and everything related.
Could be anything from concrete experts that you deem competent to podcast series, media outlets or blogs and much more. I don't care much if this is on military strategy in UKR or global consequences and international allies or any of the many other related issues. Instead, I do care for analytical power and/or high-level expertise. I'd be very grateful for your hints and recommendations.
A number of good sources already, but for day-to-day "what happened on the battlefield and why it matters", the Institute for the Study of War
http://www.iswresearch.org
and the twitter account "Jomini of the West"
https://twitter.com/JominiW
are pretty good. And reliably consistent with almost all of the other sources I track; I'm pretty confident there is a cluster of accurate reporting here(*), and if I don't have time to dig in to everything those are the first two places I check for the day's news from the front.
* Within the limits of the fog of war, of course, but also in a world where any nerd with a credit card can rent a pretty good spy satellite.
Thanks, I appreciate.
Kamil Galeev on Twitter
Thanks! If interested, there is a bit of a debate around this below. I will have a look at it.
ttp://www.antipope.org/charlie/blog-static/2022/04/behind-the-ukraine-war.html
This is Charlie Stross' blog, and tends to have good general discussion-- this thread is mostly Ukraine.
Thank you!
Well, that's my favorite topic atm. On Twitter (@DPiepgrass) I've been following and retweeting stuff from some guys that superforecaster @ClayGraubard follows, as well as his @WarSignals account which doesn't tweet very much anymore.
The ISW (Institute for the Study of War) offers quality daily updates: https://www.understandingwar.org/backgrounder/ukraine-conflict-updates
Jomini of the West also produces war maps: https://twitter.com/JominiW
On YouTube I'm watching these channels, roughly in order from most to least informative (but all these sources provide reasonably good and useful info; the first one is "Perun" - I hope you like powerpoints!):
https://www.youtube.com/channel/UCC3ehuUksTyQ7bbjGntmx3Q/videos
https://www.youtube.com/c/TLDRNewsEU/videos
https://www.youtube.com/c/VladVexler/videos
New: https://www.youtube.com/channel/UCHqqf2BwNM4Oih-a_ikbWww/videos
https://www.youtube.com/c/SpeakTheTruth1/videos
New: https://www.youtube.com/channel/UCHqqf2BwNM4Oih-a_ikbWww/videos
New: https://www.youtube.com/c/JoeBlogs/videos
New: https://www.youtube.com/channel/UCn7XHZiW6EUgSuxItybLLMg/videos
New: https://www.youtube.com/c/DefensePoliticsAsia/videos
https://www.youtube.com/c/arturrehi/videos
https://www.youtube.com/c/StarskyUA/videos
https://www.youtube.com/c/Taskandpurpose/videos
https://www.youtube.com/c/CombatVeteranReacts/videos
And BBC & CNN have good coverage. And this video was good, if not so relevant anymore: https://www.youtube.com/watch?v=b4wRdoWpw0w
And it's interesting to watch these guys explain things from a Russian standpoint (keeping in mind that it is dangerous to share their real opinions):
https://www.youtube.com/c/MultiNfz/videos
https://www.youtube.com/c/NikiProshin/videos
New: https://www.youtube.com/c/LETTERSTOKING/videos
- Malls dying, prices up: https://www.youtube.com/watch?v=1vQgx28vNsg
NFKRZ (1st link) is pushing it. He wants to return to Russia at some point, but it wouldn't surprise me if he gets arrested on arrival.
And the FSB letters are a fascinating thing. Authenticity can't be guaranteed, but it rings true. https://twitter.com/DPiepgrass/status/1506344698322903041
My favorite thing from the beginning of the war was @KofmanMichael Director, Russia Studies at CNA, explaining why the Russian military's behavior was "absolutely confounding and bizarre": https://www.youtube.com/watch?v=zXEvbVoDiU0&t=1474s
great, thanks, I will have a look
Personally I spent far more time by following the news about the war than I should, so I cannot recommend you to follow my example, but my main sources are 1) updates from the Institute for the Study of War, 2) Bloomberg special daily newsletter on the crisis which has a lot of links to stuff they have beyond paywall, but even its free part is very informative (https://www.bloomberg.com/account/newsletters/ukraine?itm_source=inline), and 3) regularly updated map on https://ukraine.liveuamap.com/. Last one is of course not exactly analytical.
Oh, btw., could you enlighten me please what that blue sign close to your - and some other names - means? I've been wondering for a while now.
It means that I paid more for a subscription than required amount, so I am apparently "Founding member". No material benefits or premium content compared to standard subscribers associated with that though, just this blue thing :-)
thanks, that makes my life easier. Or at least me navigating the comments ;). Well, and thanks for being a founding member :-)
Thanks! Unfortunately I don't need any example to read to much UKR related news myself.
I follow the Ukraine war threads at Metafilter.com. Here's the current one:
https://www.metafilter.com/194930/Ukraine-Perhaps-the-end-of-the-begining
I tend left, and I compared info with a conservative-leaning friend, and we both seem to have picked up the same info.
Yup, it is a relief that the mainstream right treats Putin's war as bad, so left & right are getting similar information for once without so much bickering!
Cool, thanks!
You get an insane amount of in-depth historical analyses in this thread of threads by Kamil Galeev. Expertise and analytical power are very high, and he also knows the Russian perspective (strongly rejects it, but analyzes it).
https://twitter.com/kamilkazani/status/1498377757536968711
Mind the chronological order, newest threads are far very down.
Great, thanks!
Any specific objection? To me most things Galeev writes sound quite plausible.
Glad to find someone else who also recognizes Galeev's bs. He is useful for trivia facts about Russian history or linguistics or what have you, but his conclusions are often wrong, and he draws in misleading pieces of information in between. I would avoid Galeev's threads or read with extreme caution.
okay, thanks for the warning. I will have a look myself. Let's see ...
I rate Karlin at only slightly above zero credibility after he predicted with 90% confidence that Ukranian resistance would last less than two weeks. Looking through @RALee85's thread on Russian losses, there are certainly a lot of minority casualties: https://twitter.com/RALee85/status/1500794161963937793
(Note that that thread currently extends to about ~400 tweets, so you have to scroll for quite a while to really get a sense of the demographics).
Now, there could easily be less than 50% minorities in there, and I agree that Galeev likes to play fast and loose with quantifiable facts to fit the grand historical narratives he likes painting, but I think the amount of new insight he provides is well worth the price of admission. I'd say read with a critical eye and understanding of his style, but read nonetheless.
I have found Vitaliy Katsenelson very helpful for background - he's best known as a financial writer, but he was born and grew up in Russia and his ancestry is part Ukranian, so his context is very helpful.
https://contrarianedge.com/tag/war-in-ukraine/
Thank you!
One way to get updates is to take two opinionated opposing sources who are directly involved in the whole thing - for example, if you understand Russian, Alexey Arestovych from Ukrainian side (his youtube channel at https://www.youtube.com/c/arestovych) and Igor Girkin/Strelok from the Russian side (his Telegram channel at https://t.me/s/strelkovii) both have interesting analytic comments but naturally have a quite opposing perspective of what 'should' happen due to their role in the war (Arestovych being an advisor in the Zelensky bureau and Girkin being a leader of the initial 2014 invasion and currently having close ties to DNR/LNR troops).
Thanks Pete, that's an interesting suggestion. Unfortunately though I don't speak any Russian. I might try with subtitles ... let's see.
Zvi's Ukraine posts with lots of links. I book marked this one, but haven't searched them all. https://www.criticalthreats.org/
Thanks!
https://www.understandingwar.org/
No frills, no obvious spin assessment of intentions, actions, and results of the various fronts, with plenty of hedging when it comes to unconfirmed stuff. Mostly reliant on OSINT but still pretty good analysis.
Thanks!
Thanks!
Medical billing rant:
So I'm uninsured and pay cash. This works well for many doctors. My PCP gives me a cash discount to $100 per visit vs. $125 for the non-cash patients. My old shrink gave me a cash discount also. (Cash is great: no counterparty risk, no insurance overhead, no billing overhead, no collections overhead, no credit card fees/delays, and maybe even lower taxes ;) At some other doctors this is horrible. I saw an urgent care doctor for 2 minutes and then waited around for their minion to perform a rapid strep test, then paid $140 by credit card. I thought that covered it all. Then a week later they mailed me a surprise bill for an additional $350, which was ridiculous and I never agreed to, except in the sense that the legalese bs you sign at the beginning is a blank check for them to bill you anything for anything without your informed consent. I could have stayed home and done my own strep test for less than $2 per test. I hate the way the healthcare system works like that. Advance disclosure of prices should be mandatory. When I go to any clinic I should demand they tell me what they're going to bill me in advance or else I'll walk out, give them a negative review on google, and go to their competitor, rinse and repeat until I find a clinic that doesn't suck. Unfortunately 99.9% of people either have insurance or aren't that assertive, including me, so we get this fucked up system where you have to sign a blank check to be seen by a doctor, instead of transparency and informed consent about billing. Why is society so adamant about transparency and informed consent about research but doesn't give the slightest fuck about transparency and informed consent around medical billing? Seems like the latter is way more likely to fuck up someone's life than taking a survey that some IRB objects to.
policy idea: as a condition for accepting medicare, doctors must agree not to charge cash patients more than medicare would pay for the same procedure. This would greatly reduce the variance and the risks of surprise billing to uninsured patients. I bet that urgent care clinic is getting less than $200 from medicare instead of the $500 that it billed me.
Price transparency is mandatory for hospitals as of last year, but almost nobody is in compliance just yet: https://revcycleintelligence.com/news/only-14-of-hospitals-met-price-transparency-rule-compliance
I don't think the rule applies to walk-in clinics, so it doesn't help your situations much, sorry. But as Monsieur Cury said above, just because they sent you a bill doesn't mean you have to pay it.
Some tips on your specific issue that applies generally to US healthcare: you can still argue/negotiate/fight the total bill. Depending on what they put into the claim coding system, you can have it updated either by analyzing it and talking with the billing department, or you can wait until it goes to collections and pay a fraction of the cost to whoever purchased the debt.
Source: I used to assist people with correcting billing errors, etc.
Follow up reading: "Never Pay the First Bill" by Marshall Allen
Follow up Watching: Dr. Eric Bricker on LinkedIn/ YouTube (I used to work for Dr. Bricker)
That is a uniquely American problem and it makes no sense to me, quite frankly. I admit I often don't ask how much things will cost in advance - if I need it, I need it - but I've never gotten a surprise bonus bill after having paid at reception.
Similar happened to me. Went to doctor, payed some out of pocket. After the visit, I asked the person at the desk the very direct question, "Will there be any other bills associated with this?" The person responded "No." Few weeks later, bill comes. Out of principle, I refused to pay it. I had medical debt collectors call a few times but after a while they gave up and nothing came of it. I haven't gone to a doctor since then, but when I do I think I might actually do what you suggest (though word everything in a very polite way).
IMX failure to pay medical bills almost never hits the credit reporting agencies for some reason.
Related thought experiment: what if you came to urgent care or ER with specific acute complaints, get through triage and see a doctor, then ask about a chronic condition, get expensive tests and diagnosis, and the whole way through refuse to offer identification? Won't the "duty of care" principle require that they treat you, regardless of your documentation or lack thereof?
You are correct. When I was still in the hospital, people who came in "without ID" were often given a name according to the scheme of "Problem" + "A random City". So you would get pt's named "Trauma Tenerife" and such. No one is collecting from them, and yes, they get standard-of-care treatment. Gone are the days of the wallet biopsy.
Caveat: if their condition was debilitating in such a way as to require rehab, they would not be able to go to rehab without giving ID. If that took too long then risk management would get involved and somehow they would be "given" an ID? Not sure at all how that worked.
The thing no one warns you about your 20s and 30s is how many marriages are actually just for shared health insurance
People do care about it, but they care about $200,000 bills for cancer treatment or a serious car accident with a two-week ICU stay. Generally people have neither the time or energy to agitate about $350 chickenfeed. I realize it *isn't* chickenfeed to you, but when you reach the age of serious medical expenses in a few decades, you'll look back in awe at the wonder of *only* signing up for 3-digit number by going to the doctor for help.
It totally is chicken feed for me and all about the principle. I didn’t get rich by letting people rip me off. End of life care is one of the worst ways to waste money since it mostly generates negative utility by prolonging the suffering. If I get a curable cancer I’ll shop around or be a medical tourist but $200k would be no big deal either. I am consciously making a positive expectation gamble by not paying insurance companies’ profits/overhead, so long as I can minimize the price discrimination against me as an uninsured person. I hate dealing with insurance companies also - it’s not worth my time to beg them to cover my medication and get no-explanation denials and wait on hold for an hour to talk to a human about it. I’d rather be in charge of my own medical expenses than totally dependent on the whims of some evil third party.
So what's your complaint? You have a plan for dealing with medical care that is radically different from what everyone else is doing, and you know that, and yet you're wondering why there is no large cohort of people with similar feelings about the process? Isn't this logically self-contradictory?
End-of-life care is one of those common health-care magical thinking canards, like you can reduce lifetime cost with preventative care[1]. You could only tremendously reduce end-of-life care if you knew for sure certain *ahead of the fact* that what you were facing was the end of life. In reality, it doesn't work that way. People certainly do recognize the end of life is approaching, and this-and-such expensive procedure may not work -- but they don't know the exact odds it will work/not work, and if it works whether it will give them 48 more hours or 48 more days or even sometimes 8-10 months. Medicine just isn't precise enough to know. People routinely do make decisions economizing at the end of life, but it's a very imprecise process, and of course most people err on the side of hope -- and that's all we need to explain how very expensive the "end of life" is, and how it's essentially impossible to reduce that cost without some kind of arbitrary and cruel algorithm that takes the decision away from patients and families, which only a monster would support.
------------
[1] It doesn't. In fact it makes lifetime healthcare more expensive, because costly to treat diseases are discovered earlier, when there is still time to do very expensive interventions. What it *does* do is prolong and improve the quality of life.
My complaint is excessive price discrimination against uninsured at some clinics in spite of the much lower overhead costs for people who pay cash up front. GoodRX totally solves this problem for prescriptions, but it would be nice to have a GoodRX-equivalent for office visits.
You might want to look into concierge medicine. Those doctors only take cash, so that way you know you're getting the same price as everyone else. Lots of them are available 24/7 if you get established with them.
Sure, but you are already understanding that what you are using is a rare method of payment and budgeting, right? So you need to consider not just the marginal cost of the transaction (paying cash for your procedure) but *also* the amortized capital cost of setting up the system in the first place. The problem is too few people use it for it to be efficient to set it up and optimize the way you're wanting. If a million people a day used it, the way a million people a day use Netflix or something, then you bet it would be optimized within an inch of its life -- because the the huge number of transactions mean any small improvement in efficiency can justity pretty hefty capital costs.
Not in your case. There just aren't enough transacations for it to be worth optimizing. So, again, why is this surprising? You can only realistically expect systems you use that are also heavily used by others to be optimized, either socially or economically.
Hospices exist because of all the people who do know they're facing the end of their lives and don't want to prolong it.
Quite right, a very good point.
I think the answer is already there: Cash patients tend to be poor, and fuck poor people.
Not at all. Poor people go to the ER and stiff the hospital on the bill. It's long been one of the arguments for a better network of community non-urgent care clinics, e.g. loosening up some of the licensing and quality control regulations for that, so you get more freestanding cheap clinics in the neighborhood, and/or some kind of more plausible health insurance that covers primary care better, so poor people go to a primary care provider in an office instead of the ER.
I think its more accurate to say Cash patients tend to be poor, but sometimes are very rich and therefore not price sensitive.
1. The federal government makes it illegal to charge anyone less than the lowest price they charge Medicare patients as a requirement to participate in Medicare. Therefore basically you can’t have a sliding set of charges. One price for all (but that’s not what gets paid).
2. Charges are set to be higher than the highest insurance reimbursement so you never accidentally leave money on the table.
3. This leaves people without insurance stuck with an inflated bill that no one really expects to get paid. Generally speaking if you tell the Provider that you cannot pay the bill they will immediately reduce it to Medicare rates.
I don't think that's the explanation for why some doctors overcharge cash patients. I think it's mostly the same reason anyone charges more for anything: because they can. Insurance companies would push back more but individual patients can't without prior info about pricing. Also maybe the pretense of a high sticker price makes the insurance reimbursement price look lower and thus gives providers an advantage in negotiations with insurance companies.
"Doctors overcharge cash patients" - this is not the common experience. Cash patients get charged less. The OP issue was charge transparency.
some doctors give cash patients a discount, other doctors give cash patients a massive markup, and it's kind of random and not disclosed in advance.
I think in a way you have answered your own question: we put up with nontransparent medical billing precisely because the majority of people have insurance through employers or Medicare/Medicaid and don’t actually pay the bill directly. The St. Louis Surgical Center is an all cash business (if I remember the name); the founder was on Econtalk a few years back talking about it, and how entirely random most medical billing is. When they started no one they worked with has any idea how much to charge and it turned out they could way under charge compared to hospitals and pay the drs more for the same procedures and the same exact staff in many cases. It was a very interesting episode. We got lots of those post process bills after our kid’s births, some from doctors who were definitely not even there. I wish the hospital had a “standard c section is X$” system instead too.
Oklahoma Surgery Center in OKC is all cash and menu pricing. I got my wrist fixed up with titanium for $5k about 7 years ago.
Until contacting them, I called all over the Bay Area and could not get anything close to a price quote except from the hospital that imaged and splinted me, who after tooth and nail quoted "somewhere north of $10k" for the necessary surgery.
Have you considered setting up a site that just presents a random review to the user on each visit?
I made a little app for this: https://getrandom.one/l/29d4
This would be great. I would also love an easy way to read on mobile.
What do you think of https://readsomethinginteresting.com/acx?
If you see any formatting or other errors, let me know at info@readsomethinginteresting.com.
I love this tool, and am finding it really helpful.
One potential bug - the T-Y bucket of reviews seems to be missing for me in the list of all reviews. Instead, the T-T bucket shows up twice.
Fixed, sorry about that.
Thank you!
Awesome!! Thank you very much.
Glad it's helpful.
This is very helpful. Thank you.
You're welcome, I'm glad it's helpful.
Or add a random number generator at the top and tell people to start at that one if they don't know where to start.
Best to make it as easy as possible, so people are more likely to comply. And this is something that is trivial to code up.
What do you think of https://readsomethinginteresting.com/acx?
Let me know if there are any errors or issues at info@readsomethinginteresting.com.
What do you think of https://readsomethinginteresting.com/acx?
Let me know if there are any errors or issues at info@readsomethinginteresting.com.
Thanks for bringing that to my attention. I fixed a couple bugs so that shouldn't happen again, but please let me know if it does.
I run ReadSomethingInteresting.com. Does that look similar to what you're envisioning? Depending on what would be needed, I'd be happy to adapt the code.
I could make use of that to do 100% of what I want to do out of the box. A+ from me.
What do you think of https://readsomethinginteresting.com/acx?
Let me know if there are any errors or issues at info@readsomethinginteresting.com.
I've used this to submit 2 reviews and it does everything I think it's reasonable to ask of it. Thank you very much.
If someone else sets up a site that presents me with random reviews, I am willing to use it. Otherwise, I am unwilling to expend any effort.
Anyone good to read on the shakespeare authorship question? Have a feeling this is a prime example in a poor ability to interpret information, lack of clear statistical priors etc. Often the arguments sound good when you ignore fairly mundane positive info. Anyway was just hoping someone around here would know. Thanks
I don't know if it's "good", but "William Fortyhands" is a really unusual take on it that doesn't specify a single person who wrote all those plays, but instead argues that Shakespeare was a producer who made the "bad" quartos audiences of the time actually wanted, rather than the artsy-fartsy "good" versions he bought from some poor scribblers.
http://www.ninebandedbooks.com/bandedbooks/william-fortyhands-disintegration-and-reinvention-of-the-shakespeare-canon/
From McSweeney's Internet Tendency last year: https://www.mcsweeneys.net/articles/shakespeare-finds-out-about-the-shakespeare-authorship-question-from-walter-mondale
Haha! That's wonderful, thanks
Not the rigorous thing you're probably looking for, but to start with you might read https://www.critical-stages.org/18/reasonable-doubt-about-the-identity-of-william-shakespeare/ , which is less about the object-level issue and more a heartfelt plea of "please we're not crank, snobs or extremists, please don't dismiss us out of hand and actually address our arguments".
Great thanks! Yes, I don't know where to being and it seems like there are a large number of varying stances taken within the Anti-Stratford 'group', so often legitimate queries get lumped in with the less evidentially reasonable claims. To clarify my current feeling is that with someone as big as Shakespeare, who wrote as much as he did, I would be surprised to not find some kind of 'code' or suspicion about his identity (especially given the historical period). However, I don't want to dismiss anything out of hand. I've just noticed that lately it's become much harder to know how to properly respond/dismiss outlandish-sounding claims. (or maybe, we're just more aware of the need for skepticism, and how deep a requiremenet that often is). Not that they ought to be dismissed, but, it would be impossible to live if you accepted as worthy-of-inquiry every theory that flies your way, so I'm often at a loss and just resort to "I'll have to look into it later".
It’s absurd. They find minor issues with Shakespeare - often spurious, like claiming that he couldn’t spell his name, the spelling of which is arbitrary. Ignore or dismiss most of the overwhelming evidence of Shakespeare being both actor and playwright, including the first folio produced by his friends and colleagues, and fellow poets. Then the other candidates are often dead for some of the period, or out of the country, and none have any relationship to the the Shakespeare theatre group documented anywhere.
Yes, thanks, I agree with you. The name thing in particular is a little odd to me, since they (first filtering out many claimed but not 100% verified examples of his handwriting) base it off of some shaky signatures he wrote in the last 2 weeks of his life; probably he just had some illness. However, the anecdotal evidence for -- and I only know very little about it -- e.g. de Vere (that he led a very similar life to e.g. Hamlet), and issues such as Stratfordians acknowledging there isn't any evidence that he was educated/highly literate (without assuming he wrote the plays lol) or had a library... But every time I look at one of the claims closer it's usually either an overexaggerated anecdote, something TOTALLY bizarre where the impose some odd way of 'skipping letters' to get 'codes', or it's some weird interpretation lacking any statistical background as e.g. you bring up "overwhelming evidence" -- it's all just discounted..
Interesting! I'll have a Google around for that, thanks
That’s Bacon. The main alternative candidate in the 19C.
Thank you, this is the sort of thing I was looking for! His name came up a few times already. Yes, I imagine it's difficult in general. I was just reading recently that "Absence of evidence is evidence of absence" idea, and was wondering if it came into play in these interpretations. Something I think about a lot is how 'vagueness' is almost definitionally something which allows multiple interpretations (lots of different shaped jigsaw pieces which all agree on at least one edge). But it also seems like the two groups are Stratfordians, who aren't really interested in the question of legitimacy enough to write about it, and Anti-Stratfordians, who have explicit outsider motivation to dispute it (not that that's a bad thing, e.g. Court systems). Thanks again
There's a mendelian study on the association between alcohol use and cardiac disease that claims a straightforward *positive* correlation between any amount of drinking and higher risk
Previously all studies show a large decrease in mortality risk with moderate drinking, which I believe is dominated by a negative cardiac risk correlation
So what this study does is isolate certain genes found to be correlated with problem drinking and uses the prevalence of the genes to rate the person's likely drinking amount, and then correlates their risk of cardiac disease to that. The study does mention the problem of using genes tied specifically to problem drinking and notes that the study screened out actual problem drinkers
Should I consider this to be reliable evidence? My current attitude is that it isn't but is worth being aware of and moderating how strongly I assume moderate drinking is beneficial
Related: On the Skeptic’s Guide to the Universe podcast I remember they talked about a meta-study on this topic a few weeks ago that found the dip in correlation for moderate drinking was likely driven by former alcoholics who no longer drink at all, but for whom the damage was already done. I don’t remember if the meta-study was able to actually correct for it, or if it looked at a subset of studies that included only people who never drank heavily, but the result was a more classical dose-response curve.
https://emilyoster.substack.com/p/new-study-on-alcohol-consumption?s=r
This suggests it's not (I think it's about the study you're talking about). But I'm not that familiar with the literature or the method used so don't have an informed opinion otherwise.
I was reminded of this concept by Scott's recent post on AI. Can anyone lay out a situation where acausal trade is plausible? This is related to a comment I made on another post, but roughly I think people jump from "very smart thing which can do things we cannot" to "thing which can accomplish anything which can be stated in English." When someone mentions something like acausal trade in the context of AI, I immediately stop paying attention since I assume they're not a serious thinker.
I think it happens sometimes in government regulation. Regulators go easy on companies, not because of any explicit quid-pro-quo, but because they hope that that will make them more likely to get a cushy industry job after leaving the government.
Of course, Matt Levine says that incentives usually go the *opposite* way.
It's actually the most aggressive prosecutors who get hired by the companies they used to prosecute.
https://academic.oup.com/rof/article/21/4/1445/2670110?login=true
My understanding of what people are referring to with acausal trade is that neither party ever actually interacts with the other. In that case, the industry people are affected by the regulator's actions, and then have future power over the regulator. I may be misunderstanding what people are referring to though.
Acausal trade means you don't need to interact *to negotiate*. I think that the weak version is very easily constructed in plausible thought experiments (plausible by thought experiment standards, maybe at the level of really consistent spec-fic), and the version I've seen discussed involving no interaction literally ever under any circumstances is pretty clearly dumb. But like, if you were playing prisoner's dilemma with a copy of yourself, it's pretty clear you could "decide" what the copy does by making your own choice. How much and how fast does this crash and/or burn when it collides with reality? That's what people argue about.
A copy of oneself I can understand. Even something like "humans from similar cultures to yourself" I can see working. The times I've seen it mentioned though, it's been in the context of simulating other currently non-existent intelligences, which isn't plausible.
I mean OK, say you got scanned today, but the copy gets printed out tomorrow. The logic still works out no? Existence isn't the limiting factor, knowability is. If you think agents are too diverse to predict well, then obviously acausal negotiation can't happen. But if there are Schelling points or something like an attractive basin of agents, it's certainly plausible. Likely? Well I suspect not, but I'm definitely not positive.
edit: Rereading your comment I realize that it's hardly a leap to consider a copy of you to be an existing intelligence. If by nonexistent intelligence you mean unknown, like totally novel intelligences with unknown desires, the cheeky:useful ratio of my example spikes. Worth leaving in for the disconnection from strict causality, but it doesn't say anything interesting about your comment. Sorry about that.
Yeah I take your point but I did mean it in the way your edit refers to.
It's a hell of a lot easier to work out what a "perfectly logical being" would do, even as an illogical being yourself, than figuring out another illogical being (including, at least for me, myself!). Not a real reply to your comment on acausal trade, which does not rely on perfectly logical beings and seems largely orthogonal to moneymaking, just a quibble.
I think there's an issue in that no one has the computational power to be perfectly logical. And what something with 10^12 FLOPS vs something with 10^20 FLOPS will be able to deduce is very different.
Well that might be true if we could all agree on a working definition of "logical." Unfortunately, outside of mathematics, we can't, so while yes *in principle* if we knew the objective final rules of logic, we could work out what a logical person does very well -- even not being logical ourselves -- in practice this is completely impossible. It's rare for even as few as two people to agree on what is logical and what is not in most situations -- just ask any long-term married couple.
I mean if you think logical should include the "right" axioms then we don't have a working definition of logical even within mathematics. But a team of smart people working the same angle could definitely figure out what a rational actor would do a decent portion of the time, that's exactly what happens in war rooms around the globe. You know what they want, and know basically what the best way to get it would be.
I'm going to suggest that the ability of people in war rooms to figure out what opponents are going to do stems from the same source as the ability of good politicians to get (and stay) elected, good CEOs to lead successful companies, and good coaches to get championship performance out of teams.
Id est, it has nothing whatsoever to do with logic, and instead stems from a highly accurate intuitive grasp of the feelings and emotional motivations of other human beings. It's certainly logical, in the sense that feelings of a certain sort predictable lead to actions of a certain sort -- but this is pretty far removed from what most people mean when they think of a "rational" person figuring out what another "rational" person will do.
Regardless of the fact that emotional motivations in no way preclude rational behavior (and in humans are required for it), a general working out that the enemy will try and hold a defensible mountain town because that's what they'd do in their opponent's shoes is not using an emotional intuition at all. The enemy wants to win the war, and the general knows that holding the town is the best way to do that.
A cursory look over the entries makes it look better than last year. I'm going to be reading these for months.
BTW Scott, duplicate entries for Troubled Blood by Rowling are in separate links- books 3 and 4.
I was caught totally off guard by both DALLE and Alpha Fold Ai in the last 6 months, both strike me as hugely important. Can anyone give me clues as to what might be the next big AI to make headlines?
Also, is there a list anywhere of all the psychological and biological research that has been discredited by the replication crisis? Or maybe I need to read review articles?
This doesn't explicitly answer your question, but a few other recent papers that I also think are important include:
Google's PaLM language model: https://ai.googleblog.com/2022/04/pathways-language-model-palm-scaling-to.html
Making a robot that can fulfill simple commands: https://www.youtube.com/watch?v=ysFav0b472w
Accelerating fusion science: https://www.deepmind.com/blog/accelerating-fusion-science-through-learned-plasma-control
Fwiw, being taken off guard by DALL-E 2 is surprising. DALL-E 1 came out a year ago or so.
I'm another who was meh about DALL-E but taken off guard by DALL-E 2. It's a huge step up. 1 looked like AI-generated art, with all its oddities and surreal dreamlike features, but 2 looks like human art and could make human artists redundant.
Yup. An answer to OP's question is thus "a GPT-4 which gets rid of most of its predecessor's quirks". I expect it any day now. GPT-3 is really impressive but not good enough for most serious uses, which is also how I would've described DALL-E 1.
Well, they have been making improvements to GPT-3. The default models the API uses are now the Instruct line. Since there's publicly usable incremental progress, that lowers the likelihood that there will be a "sequel".
I created a prediction market for whether anyone else will outbid Elon Musk to buy twitter before the poison pill expires on April 14 2023: https://manifold.markets/J/will-anyone-outbid-elon-musk-to-buy
Current value is 5%.
If the twitter board persists in blocking the offer despite lacking any reasonable belief that they can get a better offer, they're probably violating fiduciary duty. The $54.20 offer was already a 37.8% premium above the closing price the day before it was announced.
$54.20 is also significantly below where the stock closed in October of 2021. Why is April 13th closing price indicative of the intrinsic value of the company? The board has a duty to assess the offer, not accept it just because it is higher than the current price. Any shareholder that bought above $54.20 wouldn't think the board is doing its duty if the offer were to be accepted.
Elon also has a history of not having real financing lined up before making an offer to take a company private! If the board (and its bankers) doesn't believe he can actually come up with the money to complete the transaction, then they would negligent to accept the offer.
Wouldn't that require Elon to make a formal offer first? You can't outbid someone who hasn't actually bid yet.
It sounds like a formal offer to me: https://www.barrons.com/livecoverage/elon-musk-twitter-takeover/card/elon-musk-offers-to-buy-twitter-for-54-20-a-share-FoyMI3gk0znkMlqS4jGt?mod=article_inline
That's not a formal offer. That's a letter *to the board* *proposing* an offer. He hasn't made a formal offer yet, because that would require him to actually come up with $40 billion in financing, and he is unlikely to be able to do that without selling his stake in Tesla.
Yes, I heard that argued quite persuasively.
Matt Levine at Bloomberg was making that argument. Musk has already borrowed against about half of his Tesla shares, and he'd have to get board permission from Tesla to borrow against another $40+ billion against them (which he would get, but it's a "shit or get off the pot" moment). He hasn't done that yet, which makes his proposal kind of dubious.
I suspect what happens is that he just sells his shares in Twitter at some point in the next month or two and walks away. Something else will draw his attention.
I suspect that your penultimate statement (that this is rankly illegal) is true. I'd be curious about the more relevant question of "Will Twitter's board face any legal consequences for violating their fiduciary duty?"
My personal prediction is that they won't, because this is a violation of fiduciary duty that benefits the wealthy at the expense of the poor and not the other way around, similar to how sites like Robinhood blatantly trying to interfere with purchasing Gamestop stocks and trying to abet a squeeze against the direct wishes of their clients seems to have walked away without so much as a finger-wag.
PS: It looks like someone decided to short the market- value is at 26% after briefly spiking to 100%.
Why does it benefit the wealthy over the poor? Elon Musk seems like the main guy who gets hurt, and he is one of the richest people in the world (if not the richest). More generally, it seems like everyone involved here is rich. The board is rich (presumably), Elon is rich, and one of the largest current shareholders is a fund headed by a Saudi prince (and more generally, any publicly traded company is probably held pretty disproportionately by rich people).
Also I'm not an expert on this but I think that there have definitely been takeovers at a larger than 37% premium before, and times that boards have held out and gotten a better offer, so it's not obvious that holding out for a better offer is worse. The Saudi guy said so for example.
Overall, it's not obvious to me whether on net poison pills are good or bad for shareholders. You can think up scenarios for both.
Yes, Musk is very rich, but he's not the CORRECT kind of rich. He's a bit too close to the proles in terms of his behavior and interests. Maybe I should have used the term "elite" instead. Musk is incredibly, fabulously rich, but I wouldn't classify him as one of the "elite" class. He shitposts on Twitter, wears hoodies covered with japanese cartoon porn in public, and generally seems uninterested in "elite" hobbies and "elite" politics, instead doing whatever he likes. He also supports a FEW too many things that suggest to the hoi polloi that there's a hypothetical world where they have options beyond bowing down and worshipping at the altar of Conventional Capital. That makes him an outsider from the system- an enemy, even. Crushing his bid BAMN is a way to symbolically make sure the plebs know their place- much like their threats to go beyond merely rigging the casino in their favor and ban the plebs from any kind of direct agency in the stock market over GME.
What does Musk think that "suggest[s] to the hoi polloi that there's a hypothetical world where they have options beyond bowing down and worshipping at the altar of Conventional Capital"? I get he says vaguely anti-woke things sometimes ... but I doubt it even registers to most people as some sort of "'Conventional Capital' vs not" thing. This all seems like some extremely narrow super-online controversies that most people probably don't care about (to most people Musk is probably just a electric car maker who seems smart but also kind of weird and that's about it).
Meanwhile poison pills were invented a long time ago, all that they have to do is rule in line with their own precedents.
I'm not sure what Gamestop has to do with Musk and Twitter or the Delaware courts either.
I agree this is a narrow online controversy. Elites don't have real problems in their lives and are very sensitive to online controversies and the like. See how some stupid 4chan memes have (with only partial success, I'll grant) been elevated to the level of a swastika.
The Gamestop trading halt was caused by people higher up the food chain than Robinhood's CEO. The DTCC temporarily increased collateral requirements for unsettled trades to absurd levels that Robinhood couldn't afford, and by law Robinhood couldn't use clients' funds as collateral for the clients' own unsettled trades, so Robinhood and several other brokers had no choice but to halt trading of GME and AMC. Robinhood shouldn't get the blame. Blame it all on the DTCC and our archaic settlement system.
And did they face jack squat for their attempt to price day-traders out of the market? I'm going to guess "no".
I was under the impression that the collateral requirements were increased due to extreme volatility in the stock. But yes, faster settlement would be nice, and would have reduced the problem.
> similar to how sites like Robinhood blatantly trying to interfere with purchasing Gamestop stocks and trying to abet a squeeze against the direct wishes of their clients seems to have walked away without so much as a finger-wag.
Probably because this conspiracy theory isn't actually true. Robinhood stopped the trades because they *literally couldn't afford the margin requirements to continue trading*. Additionally, **there was no actual short squeeze**. All the institutional shorts got out of the market in January *before* the stock really blew up. The reason for the spike was simply because *tons of people started buying it for the lulz*, nothing more, nothing less.
I mean, if you want to believe everything Robinhood and Citadel said, can't really do anything to stop you. AFAIK there's no hard evidence against collusion between Robinhood and their primary trading partner Citadel and plenty of circumstantial evidence (including leaked internal communications) for it, and while the class-action was tossed out, individual people are winning in arbitration. Sure, that doesn't PROVE wrong-doing, but calling this a "conspiracy theory" seems far more dismissive than merited.
Do you think the SEC was in on the conspiracy? Is there anyone you think *isn't* in on the conspiracy?
By "anyone" do you mean "anyone" or do you mean "The vaguely-humanoid arthropods known as the traditional financial class"? Because those have very different answers.
To be clear- yes, I DO think that the traditional financial class engages in some level of conspiracy to disenfranchise the average American. They already do some of it in public and call it "lobbying" instead of "bribery", and I have no reason to believe that what they do in PUBLIC is worse than what they do behind closed doors.
The way it usually gets adjudicated is through proxy challenges. If Twitter's board & mgmt refuse to consider a legit offer that the majority of shareholders think they would benefit from, they can vote the board members out -- eventually. The game theory around this sort of gambit is pretty well developed. It's similar to the strategies of Carl Ichann, Mitt Romney/Bain Capital, etc. I think Warren Buffet could be put in this bucket as well.
There was at least one case where a court issued an injunction against a poison pill on the grounds that it breached fiduciary duty: https://chiefexecutive.net/a-wrong-poison-pill/
But this a rare exception. In theory poison pills are justified by giving the board more time to solicit better offers. In practice, boards mostly use poison pills to keep their jobs and don't attempt to solicit better offers. This usually violates fiduciary duty but they get away with it anyhow.
The steel man use case for a poison pill is blocking a two-tier front-loaded buyout offer that would pay a higher price up front to acquire a controlling interest and then squeeze out the remaining shareholders at a lower price. That style of buyout is basically extinct due to poison pills. It was the style of buyout at issue in https://en.wikipedia.org/wiki/Unocal_Corp._v._Mesa_Petroleum_Co. where the court ruled in favor of the poison pill. IMO both poison pills and two-tier front-loaded buyouts should be illegal.
Here's a 45-page paper from a law professor arguing that poison pills should be ruled illegal on the grounds that they discriminate among shares of the same class, etc: https://scholarship.law.nd.edu/cgi/viewcontent.cgi?referer=&httpsredir=1&article=1728&context=law_faculty_scholarship
Oh, this is great! I'm delighted with the selection of reviews and already see several I want to read. There's also one book on here I've read, disliked, and won't review but I'm interested to see how others feel about it. This is so good, this will keep me going for my Easter break!
The book review contest is a wonderful idea, thanks for doing it, Scott.
Yes, and thanks to all the reviewers! I was hunting around for something to read and now I'm all set for a while.
Canticle? For me it jumped the tracks in the third reel.
Oh dear, I don't think we should discuss reviews till the votes are in. But of course I can't enforce this.
I was referring to the book itself. I thought it might have been the one Deiseach had read and disliked.
Oh sorry I misunderstood. I loved Canticle the book.
Canticle is a compilation of 3 novellas. #1 is an instant classic, #2 is solid, #3 is filler.
I was a nerdy Catholic kid but I locked in on church history and theories on the Gospels' sources, not so much the details of the liturgy. I've been curious about what's distinct about a Battle Mass for a long time but Robert Battle's Mass makes search a frustrating mess.
You know, this might be a great question for a good old-fashioned reference librarian.
The Joseph Bronski link is to the Crazy Jalfrezi post, and I don’t see any Joseph Bronski post in any subthread there either.
The Global Virome Project is aiming to build a catalog of viruses and evaluate their pandemic risk: https://www.globalviromeproject.org/
This is a bad idea for two main reasons. First, is the risk of collecting viruses from wildlife and having them escape accidentally from labs. But also, publishing sequences of potential pandemic pathogens will enable bad actors to credibly threaten to cause pandemics. See this commentary by Prof. Kevin Esvelt: https://www.science.org/do/10.1126/f079e4e4-7689-4837-bc83-1f3fd961e946/full/
I'm sure the people at Global Virome Project mean well but as it is currently planned, their project will do more harm than good.
If you are in the USA I encourage you to contact your congressperson about stopping this.
My reaction isn't OMG that's dangerous, it's OMG that's too hard. Actually, I was reading fast, so I saw "catalogue of viruses" rather than "catalogue of viruses and evaluate for pandemic potential."
Just cataloguing viruses is, well, my intuition is that they'll catch 1 thousandth of wild viruses. Let me know if this seems way off.
I will say that people have been wandering around all over the world and eating stuff, and pandemics are pretty rare.
And I was assuming wilderness, but, of course, viruses are all over the place and some of them are probably specialized for civilization.
I suppose that the viruses (too many to completely catalogue?) that are already in humans should be assumed to have low pandemic potential.
My guess is that cataloguing all the viruses is impossible but worth attempting, and evaluating them for pandemic potential is a good way to get funding.
Nonsense. Nobody has the knowledge necessary to interpret DNA sequences and use them to fashion brand new evil sequences. If that were even a little bit possible -- you'd have to have solved the protein folding problem, first, so you knew what proteins were made by what genes, *then* you'd have to have an exquisite understanding of all of immunology at the molecular level, which is several centuries away I would guess -- then e.g. cystic fibrosis and sickle cell would long ago have been trivially cured by viral vector borne genes, and it would be child's play to invent effective vaccines to almost any viral disease. This is the same kind of wildly-overoptimistic assessment of new technology that let people when they first learned of the Internet think it totally possible for a teenager on his TRS-80 to hack into US ballistic missile command computers and start World War III.
I don't understand how this is relevant to the top-level comment. The Global Virome Project is trying to collect and research viruses that already exist in nature but for one or another reason have never made their way to humans. Metacelsus is calling out that some of these viruses may be capable of causing pandemics in humans without modification, and by collecting these viruses, researching them, and publishing their genomes, we are increasing our risk of facing a novel pandemic that would not have happened without the GVP. This could happen either by an inadvertent lab leak from researchers studying the virus, or by purposeful release from a bad actor.
How do you "release" a virus that is already endemic in the environment? Anyway, I interpreted the comment as being concern about the publication of the genome -- and assuming that people could get lots more valuable information out of a genome than they currently can. When the human genome was (almost entirely) published in 2003 people predicted it would be used to genetically engineer humans, genetically screen employees, discriminate in terms of health insurance or in jobs ("Joe has a 37% higher chance of being a kleptocrat, according to his genome, so we better not hire him as our accountant"), pave the way for human cloning, and so on.
None of that actually happened. It turns out knowing the exact sequence of a billion or so ACTGs actually doesn't get you much, all by itself. It's just letters on a page, like so many millions of digits of pi. It's at best the *starting point* for additional research, where you painstakingly cross-correlate sequences with biochemistry, and symptoms, and actually start learning stuff. Even then, it's not 100% clear it wasn't a waste of resources, since we *could* have just stuck to sequencing lengths of DNA on an as-needed basis, whenever other research suggested it might have value.
I think the same thing is likely to be true about collecting viral genomes. It might be useful, because people can do cross-variant comparisons, study how fast they evolve, stuff like that. It's a nice fat data bank. But I think the idea that creating deadly pathogens is just hanging on knowing the exact sequence for HIV or something ("Aha! *That's* what we're doing wrong! It's an adenine in position #6,544 not a cytosine!") is very unlikely.
Lol no Metacelsus is right
Protein folding problem = alphafold and then scRNAseq + metabolomics data solves IDRs
We literally engineer immune cells from scratch to cure cancer patients. It's actually very easy/like 30 min a week to make great virus like I did for years.
Our understanding of immunology is at the point where friends tell me I really need to censor my dark humor/jokes on fallback careers in bioterrorism when grants/manuscripts get rejected etc
Sorry, no, none of that is true. Or perhaps a better way to put it is that I think you have so far reduced the definition of "solved" and "from scratch" in your examples that this is not a meaningful argument.
Anyway, if you think you can personally engineer a virus from scratch that will safely deliver working CFTR genes to cystic fibrosis lungs, you shouldn't be wasting time commening on Internet fora -- your Nobel Prize in Medicine awaits, along with enough money to buy Sardinia.
I emphatically disagree with this assessment, and would argue that the Global Virome Project has orders-of-magnitude greater upside potential than downside potential. I think my view diverges from yours (and Prof Esvelt's) based on three points (roughly in order of importance):
1) In terms of preventing the spread of genomic data on deadly pathogens, we lost the game decades ago. The full genomes of anthrax, smallpox, and similar classic plagues are publicly available, for anyone who has internet access and knows where to look. The COVID genome was, quite famously, shared globally in early 2020, almost exactly as fast as was technically feasible. If you're a mad scientist looking for something nefarious to assemble with your evil synthetic biology lab, you're already spoiled for choices. There's little we can do to make your life better or worse on that front.
2) Suppose that there is a virus out there that can spread in humans well enough to become the next pandemic, and that it's common enough in its current animal host for GVP to find it. If these things are true, chances are very high that this virus is coming into contact with humans often enough that it will make the jump to us without scientist assistance. Almost all parts of the world which are inhabited by mammals are also inhabited by significant numbers of humans; humans who, moreover, don't observe biosafety protocols when dealing with the odd critter that got stuck in under their house. We're not faced with a question of whether or not to let a novel pathogen "out of the box" in order to study it. There are no boxes in nature; if it's out there, it's already out.
3) Given 2, we can expect that ~everything the GVP will be a "near-miss" pathogen; something that potentially replicates in humans a bit, or spreads a bit, but not to the degree to explode into a large pandemic. (It could still be bad for individuals - think SARS-1.) If you're an evil mad scientist trying to turn one of these into the next pandemic agent (having decided for some reason not to use already-available anthrax, etc for this purpose) you're faced with the problem of engineering a biological agent to have a consistent, pre-defined effect in the human body. This is the same kind of problem faced in typical pharma or biotech work, and broadly speaking, is very hard to do! I am admittedly not a virologist, but to me this feels like the sort of project where you would need a well-resourced team of smart PhDs to expect success [0].
If the above convinces you that the risk from empowering potential bad actors is fairly low, I think it's pretty trivial to see that the expected value from GVP is very positive. COVID is only part of a general pattern of emergent pandemics or near-pandemics; think SARS-1 or H1N1, or going back further Spanish flu, etc. We can give high likelihood to the emergence of a new pandemic-capable pathogen in the next century. Even in a conservative case, where GVP gives us some insight which merely accelerates the development of a vaccine for a novel COVID-like pandemic, the benefit would be measured in hundreds-of-thousands of lives saved. (See e.g. the recent article https://www.maximumprogress.org/blog/how-many-people-are-in-the-invisible-graveyard .)
(Disclaimer: I am not in any way affiliated with GVP, and in fact hadn't even heard of it before Metacelsus' post. So thank you for letting me know about this neat thing! : ) )
([0] Just to preempt an objection here - yes, I am aware of "cell passaging" methods for evolving viruses for suitability to a given host. My impression is that transmissibility in a cell plate translates only roughly to virus effectiveness in-vivo. In which case, the best an evil mad scientist could do with this technique is probably a sort of scattershot approach where they just keep quietly releasing new variants and 1-in-100 of them are noticeably bad. Not great if someone actually is just a determined anti-utilitarian, but also not really suitable for most kinds of terrorist or etc.)
A response to some of your points.
1) Unfortunately pandemics are not a binary yes/no question. A malicious actor can spread smallpox, Spanish flu, and five new diseases discovered by the GVP at once if they wanted, and it might be 7 times worse than spreading just one since they don't provide immunity from one another.
In addition, Kevin Esvelt argues that all the existing viruses have limitations which make them not as bad as it could be. Some examples include:
- Spanish Flu apparently has enough similarities to modern flus that we should expect a far lower death rate, and the death rate should be further reduced by the fact we have antibiotics (around 1/2 of deaths in the original spread were from bacterial infections)
- Smallpox: We've managed to get it under control before, it doesn't have asymptomatic transmission, and we have about 350 million working vaccines in reserve
- Ebola, SARS, MERS - transmissibility seems to be low enough that we can get them under control, in fact this has been done before
Also, Anthrax is not generally contagious.
2) Could you provide a source on this? The fact that humans and mammals interact a heck of a lot doesn't seem to mean that all viruses will necessarily spread. One piece of evidence in favor of this is that viruses tend to spread in countries with particularly transmissible practices like widespread wet markets or bushmeat markets. Unless you think these countries naturally have more pandemic-causing viruses than others, it's likely there are some dangerous viruses in countries with more safe practices that just never made their way to humans.
In the podcast I mention in a post below, Esvelt doesn't go into too much detail but says that he believes there are diseases that would never get to humans if not for this virus hunting.
3. In the podcast, Esvelt mentions a lot of uncertainty but estimates that the DEEP VZN program, which would find and characterize about 10 thousand viruses, should expect to find somewhere in the mid-single digits pandemic-grade viruses.
It's very unclear how much the benefit to this knowledge is either. Unless we're planning on stockpiling a huge amount of vaccines, or running challenge trials, which I don't think anyone is suggesting (but probably should if they are planning on going through with this), at most we will get a few weeks lead on a vaccine. Remember that the Moderna vaccine took only two days to make and most of the delay was from testing and regulation. In general it's unclear how much would be different if we knew beforehand that COVID-19 might cause a pandemic. For example I doubt that knowledge would've given us the capability of stopping COVID from spreading back in 2020 given that we aren't capable of stopping it from spreading now. In addition, most of the benefit we could get from studying these viruses could be gotten by studying them under lock and key and without sharing their genomes publicly.
There's a clear cap on how much such foundational research can help. We've had around 4 significant pandemics in the last century, let's say everything goes extremely well and (even though this is wildly implausible) we are able to stop four pandemics from happening in the next 100 years because of programs like GVP and DEEP VZN. We do this at the expense of giving ~30000 people (an estimate of the number of people currently with the ability to animate a virus from a genome, and this number is expected to increase significantly in the coming years) the ability to launch a pandemic on their own and cause an amount of death in expectation on the scale of a single nuclear weapon.
It sounds like an important crux here is whether humans would readily catch human-pandemic-capable viruses out in the wild; if this is true, we shouldn't expect GVP to discover any novel human-pandemic-capable viruses (because the absence of the corresponding pandemic shows they don't exist,) whereas if it's not true, GVP is liable to produce and provide genomic data that's more dangerous than the Zika et al that are already available. Unfortunately, to my knowledge, this isn't something anyone knows the answer to. It seems to me like a lot of the biosafety literature in this subject is happy to *speak as if* the answer is trivially that such animal-to-human transmission is rapid. Whether this represents some implicit knowledge among practicioners in the field, is benefit-of-doubt granted out of respect for existential tail risks, or is language designed self-servingly to earn grants, is open to interpretation. If anyone knows more on that subject I'd be happy to hear. (I also haven't yet had a chance to listen to that podcast, so possibly Esvelt has answers to all this already.)
Regarding the point "viruses tend to spread in countries with particularly transmissible practices like widespread wet markets or bushmeat markets"; speculatively, I would suggest that this is probably because such "particularly transmissible practices" are almost universal in rural areas. Anywhere you have a people living near numerous and sizable wild animals, humans will want to eat them, and they will construct something similar to bushmeat markets for this purpose. (Less common in first-world countries which have ready access to refrigeration, probably, but you can still get *very* fresh game meat in rural America if you know where to look.)
On the benefit of foundational research preparation - COVID is a weird case to take as the baseline, since it was the third SARS-like disease we had seen in the past several decades. By some accounts, our experience with SARS-1 and MERS was decisive in our ability to rapidly develop vaccines, and for our good fortune that those vaccines turned out to work. If the next pandemic emerges from a less-familiar viral family, knowledge from an effort like GVP could be all we have to go on for the first several months of the pandemic, which could easily be decisive both in developing vaccines but also in how we go about prevention (remember the long debates in early 2020 about whether COVID was spread by contact, fomites, airborne droplets, etc.) Next time, data from GVP-like work could be necessary for an outcome *even as good* as we saw with COVID.
COVID is also a sort of weird baseline since it's very-high-transmissibility and relatively low-severity; yes, probably no available foreknowledge could have kept COVID from taking roughly the course it did. However, a moderately-transmissible, high-severity pandemic would be a very different beast; something we could indeed potentially contain, if we knew how to go about it, and where it would be even more important we do so.
I do notice that I'm pretty consistently arguing from an "it's actually worse than that already" basis, so let me add one further argument of that kind to the pile. I've noted that we already don't control access to genomic data for known pathogens; also realize that we don't control access to genomic sequencing, either. Going back to our evil mad scientist, nothing currently stops him from buying a plane ticket to somewhere appropriately exotic, buying a fresh bat from the local bushmeat market, separating out the viral DNA from a sample of the bat's mucus, sending that to Illumina for sequencing, and then doing some metagenomics on the results to get the roughly-complete genomes of much of what's in there. I'd estimate all of this would set him back about $10k, if he got a cheap plane ticket, and would give non-negligible odds of netting a potential pandemic pathogen, in the low-animal-to-human-transmission possible worlds where these are waiting out there. A well-resourced actor could readily scale this, obviously, so the cost to effectively replicate the downside of GVP is not large. Additionally, since this could be done secretly, arguably the only practical defense is to have prepared for the scope of pathogens which malicious actors would thereby have access to; this amounts to something very much like GVP. There is at best a tradeoff, then, between risk from individual actors (who couldn't replicate GVP) to larger actors (possibly state actors, who definitely could replicate GVP.)
Overall, my general intuition is that in a world where dangerous genomic data and genomic technology are widely and cheaply available already, the marginal impact of biosafety-dedicated development will be decisively positive -- even if only because we are starting from a completely rock-bottom level of vulnerability to pandemics both natural and artificial, such that (aside from the advance of technology in general) it's difficult to make things in this area worse in any respect.
I agree this is an important crux, and unfortunately I don't have much more knowledge on whether or not it is true. While I think the podcast is a worthwhile listen, it doesn't shed much light on this topic - Esvelt states that there are diseases we could find that never would've spread to humans but does not elaborate. If there are any interesting biosafety papers on this or a related subject you could recommend however, I would love to read them.
Two comments on this topic though:
1. The usefulness of GVP (not just the downsides) is also related to the number of pandemic-grade viruses they find. If all the viruses they find are complete duds, it seems unlikely this project will matter much. If they find a number of viruses that are not pandemic-capable but are close to it, the value we will get out of the program will be from how similar the properties of a mutated, pandemic-capable version of the virus will be to the original non-mutated version that scientists study.
2. I wouldn't put it past some virologists to take a near-pandemic-grade virus, run gain of function research on it to make it pandemic-grade, and then publicize the genome for that. We've already seen similar lack of foresight from previous experiments. Though I suppose the dangers of gain-of-function research is a slightly different topic of worry.
On foundational research preparation: It's never been clear to me how much this type of research on potential pandemic viruses helps limit future spread. Certainly it helps somewhat, but I haven't heard a compelling case that it would be very significant. We're limited only to data we can collect in-vitro or in animal models for one. And most of the benefits we do get from this research can be drawn without widely-publicizing the genome and making a rank-ordered list of how dangerous it is, essentially pointing an arrow at it stating "this virus is dangerous" to any wannabe bad actor. Of programs like this, I'm mostly neutral on the virus collecting phases, and negative on the pandemic-characterization phases, especially in concert with publicizing the genomes of these viruses.
That said, there is some work in this space I'm very much in favor for. For example, I think work on vaccines across viral families - things like pan-coronavirus vaccines and pan-flu vaccines are a very clear good. In addition, we should vastly increase our disease-monitoring capability to quickly identify when some DNA/RNA sequence is rapidly increasing in magnitude. Thirdly, I think we should do all we can to quickly and permanently increase our mRNA vaccine manufacturing capability. This would enable us to build vaccines for future pandemics at scale extremely quickly instead of needing a long ramp-up period.
I disagree that so much information is out there that it can't get much worse. As I mentioned before, there are limitations to how bad the existing sequenced viruses are, and GVP could plausibly come up with something worse. And even if GVP only comes up with something comparably bad, increasing the number of viruses bad actors can use to cause harm is still very dangerous.
You mention the possibility of an evil mad scientist doing pandemic characterization work themselves. That's certainly possible, but it both raises the financial and the technical barriers for bad actors. Characterizing one virus would not be enough, since I think no one expects a randomly chosen single novel virus to be pandemic-grade.
If this mad scientist wants to scale up their $10k per virus work up to 10 thousand viruses, which is what DEEP VZN is hoping to do, they would need $100 million (which incidentally is similar to the budget for DEEP VZN, which makes sense since the pandemic characterization work seems to be the most expensive part of the program). While there are maybe tens of thousands of people with the ability to animate viruses from a genome, the number of people who can do virus animation work, pandemic characterization work, and have access to ten million or more dollars is significantly smaller, and may in fact be limited only to state actors.
I generally think individual actors/small groups are the biggest worry here if these more dangerous viruses get publicized. State actors have much more tailored weapons they can use that don't have the downside of potentially ravaging their own population.
Esvelt recently did a podcast on DEEP VZN, which is a related program run by the USAID. Anyone interested in details of the conversation, and a background of the type of work these programs are planning on doing can look at my notes here:
https://forum.effectivealtruism.org/posts/Mo3T8c2WxifvEGGps/the-pandemic-threat-of-deep-vzn-notes-on-a-podcast-with
Marking to get notified of replies
Does anyone have an idea as to Russia's capability of significant conventional escalation in the form of missiles, bombs' and artillery?
I don't read very much about this.
Is it that they do not have the capability to unleash much greater conventional destruction or is it that imaginations are gripped more by the horror of non conventional weaponry?
Probably they could kill more civilians with carpet bombing of cities and other conventional means. Fortunately that would be actually counter-productive in terms of winning the war. Unfocused violence against civilian population stiffens Ukrainian morale and leads to more sanctions on Russia, as we have seen with Bucha. I hope Putin noticed that.
They could also escalate by bringing in more MANPOWER, instead of/in addition to just more violence, via general mobilization of reserves, and this is imho likely scenario
Everything I've heard is that Russia is struggling with the manpower it already has in the field. They would be looking at raw conscripts, which Putin has already promised not to do (before things got bad for him). Raw conscripts are still technically an option, but they would be very poor soldiers even from a skill perspective, but just opening the draft would be a major issue. For these new soldiers to be effective, they should at least go through boot camp (6-8 weeks) and probably should get six months of training. The most likely scenario is a very rushed training and then thrown to the wolves, which would be a disaster. If Putin was waiting for even two months, or had the luxury of waiting that long, the war would already be going very differently for him.
Russia has a pretty big army, but they also have the largest borders in the world. They're stretched very thin just manning normal border operations against neutral/friendly countries. It's important to note that they have the 9th highest population in the world, between Bangladesh and Mexico, and less than half the population as the United States. Their population is not much more than Japan's (which has 0.25% of the land on earth, verses Russia's 11.5%), and less than several different European combinations (such as Germany and France) which have *much* smaller borders. In terms of their economy (ranked 11), they fall between South Korea and Brazil (closer to Brazil than any other country).
If they didn't have the largest land area in the world, and nuclear weapons, Russia would be a minor power in line with Mexico. Memories of the USSR bolster Russia's prestige, but it's important to remember that many of the richest and most population dense parts of the USSR broke off, including Ukraine.
Sure, I am aware how many people live in Russia. If they would have invaded China, they would be pretty screwed.
But it is important to notice that Russia still has larger population than Ukraine, although not that much larger than people who thought invasion is going to be a cakewalk apparently assumed. Ukraine technically conscripted every man between 18 and 60 years, although obviously so far only minority of them had participated in actual fighting. But I am sure many current Ukrainian combatants are poorly trained. If Russia would manage to mobilize to the same extent as Ukraine, it would be Russia which would have an upper hand.
Now, Russia is unlikely to do that, because it would be asking for a revolution, and also Ukraine has an advantage in that it is on the defence, which is easier than attacking. But even very partial mobilization in Russia could considerably move the scale.
My understanding is that the application of conventional military fires like artillery, guided missiles, MRLS and support from tactical planes and helicopters are already limited by their equipment and logistics, they are using them as much as they can or as much as they consider useful, so there is no real space for escalation.
IMHO there could be some capability of escalation for strategic bombing with dumb bombs, which has recently been started on Mariupol but might be done on other population centers - however perhaps that is limited by the risk to the bombers from Ukrainian air defence, especially as more advanced hardware is arriving from the west.
good question, following
At this point, the general consensus is that Russia is pretty tapped out with regards to their conventional forces. They've shown a high reluctance to use precision weaponry suggesting extreme shortage, have been losing tanks and armoured vehicles at an unsustainable rate (like 20% of their entire active tank force in under two months, last I looked), and without mobilisation they're extremely short on troops on the ground, not that those troops on the ground seem to want to do anything but get the hell out of Ukraine.
It is questionable whether Russia under sanctions is capable of resolving these issues in the short to medium term. If anything the fundamentals look to be getting worse for them, not better, what with the West's carte blanche bankrolling and equipping the Ukrainian state. If the thrust at Izyum fails catastrophically, or more likely is disrupted by counter attack, the Russians might not have any capacity left for another push and will be remanded to the defensive for the duration of the war. Whether that means they accept the loss or escalate to nuclear weapons is anyone's guess but, fortunately for the world, nuclear weapons are not terribly useful in this situation. Distributed, infantry armies in the field are terrible targets for such bombs and hitting the cities would only increase the amount of weapons being funnelled into the country. Hopefully that means they won't be used for pragmatic reasons as well as moral.
Russia's military is already stretched pretty thin. It's powerful on paper but not so in practice. To make up for the numerous deficiencies, Russia's generals has always been willing to sacrifice massive numbers of soldiers. That works to a degree; it worked in WW2. However, this time around Russia is unlikely to be able to get the same level of engagement from its citizens; furthermore, this time around Russia doesn't have any significant allies.
Thus, the main threat everyone is concerned about at this point is nuclear weapons
By all appearances, this was supposed to be a swift and victorious "special operation" vs. a fake propped up Afghan-style government, without any realistic plan B. It seems that now the priority is to save as much face with as little effort as possible, with the propaganda still not switched to presenting this as a serious war effort. Nukes are for now on hold, Putin still wants to have the last years of his reign to be only moderately disastrous.
Musk: "To get a self-driving car, we need to solve real-world intelligence" - which, in many ways, is a human-level general intelligence. Same applies to Tesla’s humanoid robot Optimus. The timeline is:
"Around 2025, Musk thinks there will be rapid growth year-over-year in the usefulness of Optimus."
https://youtu.be/cdZZpaB2kDM
(the links post was the wrong place to comment with this)
Basically, if this particular endeavor works out, then at least for a short time (possibly quite short, assuming Optimus learns to self-improve despite the safeguards Musk talked about) there will be a time where menial jobs will go away, and productivity will soar, while unemployment will be soaring in tandem... So the Depression-era legacy of coupling of jobs with income might finally be broken, and Marx's dream of plenty and “From each according to his ability, to each according to his needs" might actually come to pass... in the best case, anyway. Humans will certainly manage to eff it up with ideology and culture wars.
There are many jobs that could be automated right now with current technology. However the cost to do so is higher than just hiring humans to do it (amazon automates a lot of warehouse functions but sometimes its easier to just ask Dave to go get something off a shelf) or there is some other factor which makes hiring people the better option in the long run (you can already book a hotel room, check in, complete your stay, and leave without ever talking to a real person, but there are benefits to the hotel to have a human perform some of these functions such as making the guest feel a certain way or provide a level of security that is hard to get from a camera). The technology is rarely the only thing holding back the adoption of new inventions.
Didn’t say it was. However most technology isn’t there yet. But an unmanaged hotel (which we could do now, book in and get access with an app on your phone) would be horrible and insecure, to most.
Maybe, but it's how almost all AirBnb properties currently operate.
I don’t know if self catering is strictly an example of automation.
My model 3 doesn't even understand simple voice commands like "turn off the wipers" so I'm not bullish on the functionality of Optimus. One day of one engineer poking around manually could have filled in so many of the gaps in the voice command system. I understand it's basically just a manually generated list mapping from exact text to exact API call.
Wait, he's working on a robot? Oh man, I can't keep up with all this. I thought his electric tank or SUV or whatever it was, was bad enough, but now he's trying to evoke Optimus Prime as well?
Musk tends to do stunts - either new companies or whole new products - whenever there might be some bad news coming in the near future (or at large).
He also needs to keep the value of Tesla up. As Matt Levine pointed out in his recent newsletter, Musk has pledged something like half of his Tesla holdings for loans. A serious drop in Tesla share value could really put him in a financial crunch.
I think Elon Musk enjoys yanking the chain of his inverse fanbois, those who enjoy working themselves into a lather over his exaggerations, implausible prognostications, and generalized bullshit.
He announced the humanoid robot a while ago, but as far as I know they haven't shown anything except a model.
And IIRC the "model" was a human dancing in a silly robot suit. I think it's pretty clear why we all forgot about that one rather quickly.
Menial jobs are by far the hardest to replace. Musk continuously over delivers and under promises, so let’s see.
It's doing it cost-effective as well that's a big challenge, too. We could probably rig up people's lawns and gardens right now so that most of the work would be automated, but it would be far more expensive than the cost of just doing it yourself (or hiring someone to do it).
That is actually true, and you have only to look at where the steepest wage inflation today is occuring -- it is *not* among programmers or physicians, but among guys who fix cars or prune trees.
The reason is simply that the most "menial" jobs, to most people, are those that require the least formal education. They take advantage of the base human talent for excellent vision, hand-eye coordination, generalized manipulation of almost any tool, and high adaptability to changing circumstances. The training largely consists of apparenticeship and mimicry. Painting a house or installing drywall fits that bill -- there's no need for any formal education above reading and writing and a smidge of arithematic, and you learn pretty much by watching other people do it. It's economical for a human to do it, because you have to use human vision and manipulation of hand tools, at which computer-assisted and robot systems naturally suck.
It would cost a huge amount of money to design a robot that could come to anybody's house and fix a hole in the wall, or move an electrical outlet, or snake a clogged drain. So those "menial" jobs are just about as human-labor intensive as they were 50 or 100 years ago, and wage inflation is quite high there -- as any unhappy homeowner could tell you right now.
By contrast, it is actually pretty plausible that Javascript programmers, for most basic situations, could be replaced by computer programs. It's not that unlikely that the same could happen to financial traders, or bank managers, or any other kind of occupation the qualifications for which are *largely* sophisticated but precise and logical learning, and the execution of which involves no manipulation of *physical* objects but can be done by interaction with computers.
> and you learn pretty much by watching other people do it.
Well, and by doing it a LOT yourself as anyone who has tried to detail a car or paint a house can attest; there's a ton of *physical* skill necessary to sand and repair drywall properly and paint with an even-looking coat across materials with variable absorbancy and variable paint load on your roller.
It's actually a really hard problem to model, which makes it no surprise that it's hard (and expensive) to teach a robot to do well.
Yes, I should have added you need a lot of practice, and the fact that the ability to watch yourself work, recognize the outcome is suboptimal, and invent new things to try to improve is *also* a uniquely human ability. We don't know how to teach a robot to do that -- we have vague ideas like Monte Carlo samping and simulated annealing, but nothing we've come up with even approaches the human efficiency in finding ways to improve. Robots and machine learning systems can and do improve -- but far, far slower than a human would, and require far, far more data and training.
I think you've got that backwards
haha. I did. Weird. I always read what I post afterwards in case I need an edit but still didn't see it!
Right. A robot that can wait tables well is going to cost a lot more than a human who can. And the more menial jobs that are replaced by robots, the cheaper the humans get.
It’s surely the hardest thing to get right. A robot that can pull a pint of Guinness, wait a table, uncork a bottle, pour a bottle into a glass, carry a laden tray up and down stairs, fix pipes, consult a menu and recommend a wine, talk and listen sympathetically (Turing test) , deal with obstreperous customers, make cocktails and on and on.
And all for a price that’s less in depreciation and software maintenance per year than the price of a low paid worker (I won’t say minimum wage as the best bar tenders aren’t on that).
It’s the stuff that humans find easy, or easier, that computers and robots find hard.
Prediction: The prediction/forecasting platform that gains the lions share of users will be the platform that allows predictions to be made within substack, LessWrong, EA Forum, and other blogging and journalism platforms. As in, ACX writes a post about topic X, and at the end of the post Scott embeds a prediction widget to invite readers to make a prediction about topic X.
An advanced version of this would be an attention market for essays, bloggers, and journalists. Essays that lead to changes in the market get promoted to other forecasters. Overtime, the stats would begin to show which bloggers and journalists are best at getting people to update a lot (persuasion factor), and which are best at getting people to update in the right direction (accuracy factor). Super journalists as opposed to superforecasters, so to speak (name tbd).
An even more advanced version would create an attention market for the comments section, allowing you to mark specific comments as having influenced you, and promoting those comments to other readers. Supercommenters (name tbd).
A yet even more advanced version would identify individuals who are good at separating signal from noise. In other words, those people who are good at identifying blogs/comments which should lead you to update vs false leads. Super Curators (name tbd).
And of course, this ecosystem would incentivize bloggers and journalists to focus on where they can cause the largest updates, keeping their readers accurate over time, and getting users to react appropriately to the news. It also gives quantitative feedback to writers and commenters about the influence of any particular thing they’ve written.
Haven’t thought this all the way through, and perhaps it introduces some bad incentives or reduces independence of forecasts too much. I’m also not sure if it could be a liquid market or if it would have to be a forecasting platform. Though perhaps blogs and journalists have their own reputation and revenue models, and so the ‘market’ aspect wouldn’t be as important.
I hope everyone is having a great weekend! I am currently studying for the CA bar. Does anyone have tips on effective study methods? Thank you!
Do as many practice tests as you can and do them under the same time limits you will have for the real test. Essay questions will of course be more difficult to have properly judged but the goal is more to understand the form of the test and how good your knowledge is under stress.
Yes - look at the pass rate of your school. If it’s high, calm down, take Barbri or whatever. You’ll pass, nothing to worry about.
This is only like 10% tongue in cheek.
When I took the CA bar a few years ago, I just did one of the standard bar prep classes over the summer before the bar, starting around graduation-ish. (i think whichever brand was not the most expensive, so not barbri but like, the B-tier option) with book and on-your-own-time video. Treated it like a not super hard job, did 4-5 hours a day for a few months, completed probably 95% of the standard coursework, did not do anything else. Went fine, if anything, probably more prep than I needed. (Fair disclosure, I am a confident test taker and did not expect that I was going to require heroic measures to pass.) If you think you will have a hard time doing it on your own time, then paying for an in-person class or forming an accountability group might be a good idea.
My study strategy from back in the day:
1. Plan out the whole day in one-hour chunks, different topic each hour. Build in a one-hour lunch break (and maybe morning/afternoon tea? I can't remember)
2. Work from the top of the hour until fifty past
3. Take ten minute break, then start studying again
4. Oh, and don't start studying until the week before exams, because this shit is unsustainable
RULE 1: Stay rested/fed/hydrated. Don't think thirsty; that fucks with your brain.
Other than that; I inflict detailed explanations of whatever it is on family/friends/passersby/my dog. Helps me, anyway.
I recently took the CA bar, and I think for me the most important thing to remember was that you need to stop studying! Don't try and do too many hours in one day - for me that meant by dinner time bar prep was over no matter what.
Set a timer for mandatory light exercise breaks. A walk around the block at least.
Have you used Anki?
I have— it’s been helpful
Regarding heavy weapon shipments to Ukraine:
The Czech republic sent Ukraine some modernized T-72s, Slovaks sent their S300 anti-aircraft systems, I think that Poland also sent some tanks. The russians sent our (Czech) government an angry note, saying that we are not allowed to export weapons of Soviet origin. Our ministry of defence basically told them it was bullshit and there are no such provisions in any contracts. The russians then posted online about destroying both the SAMs as well as the tank shipments, neither turned out to be true. They have not done anything else to stop these shipments.
Despite all of this, Germany still seems to be unwilling to send Ukraine heavy equipment of its own, they cite the fear that "Europe could become a target of russian aggression". I am honestly not sure if this is due to cowardice, extreme naivete or affinity towards russia (which Germany has shown a plenty of, sadly). Is this uncharitable of me and is there a different explanation? I can't think of anything and it seems to me that Germany is rapidly losing any realistic bid for a leading role in the EU with its current approach. Zelensky is not very diplomatic towards Steinmeier, but it is very hard to blame him.
The UK and the US are doing a lot more and to me they seem to be the only NATO members other than Poland, the Czech republic, Slovakia and the Baltic states who are really providing Ukraine significant military aid. But NATO countries (including the Czech republic) officials still keep repeating that they cannot easily send Ukraine NATO tanks (other than the Soviet era reserves such as the Czech T-72s) because it takes time for the soldiers to learn how to use them. Is this still a valid reason though? The war is unlikely going to end in a week or even a month. How hard is it to learn how to operate patriots for example? Does it take more than a few weeks? It might require military instruction from the US but UK (i.e. NATO) instructors are in Ukraine again already, training Ukrainian troops, so I don't see why the US could not do the same. Or am I missing something?
So far it seems to me that russia backs down every time the West shows a litte bit of backbone. Putin threatened to only accept payment for gas in roubles, when the West said they would not comply, Putin backed down. When the Czechs sent tanks, russians sent an angry not and tried to spread fake news about their destruction but nothing else. When the UK apparently sent military advisors to Ukraine again, russia did nothing. To me it implies that NATO can definitely flex its muscle a lot more.
There are technical reason, as others have pointed out. But it's almost entirely cowardice. Modern politicians are not elected for their courage or virtue (in fact the process mostly selects against these traits as you would be more likely to lose elections early in your career). Politicians don't want to lose elections so they are always trying to navigate the rapids of public opinion. In smaller countries, like the slavic nations or baltic states, the range of opinions on this topic is smaller (these countries know they are next on Russias list) and so politicians are more likely to act "courageously".
How does that fit to the public opinion in Germany being - at least until recently - strongly in favour of the government doing more?
Imho there is an overlooked and awkward reason for German reluctance to send piles of weapons to Ukraine: they do not have enough of them, and being overly generous could endanger their own defense. See here: https://www.dw.com/en/germany-bundeswehr-arms-deliveries-to-ukraine-reached-a-limit/a-61418152.
Many people still have not realized that one of the main reasons why Russia still have not beaten them is simply that Ukraine is a large country, which has been heavily militarized before the war. Germany is obviously far richer, but they had underinvested in their military for many years, so here we are
There are some other reasons to be reluctant to give the Ukrainians NATO heavy weapons:
1. The high probability that some will be captured by the Russians, giving them a bit more insight into NATO technical capabilities than is entirely desirable.
2. As I think someone else said, it's not just a question of training on the vehicle itself, it has a long logistical tail of spare parts, fuel, ammo, tactics, integration with supporting assets, and command communications protocols and technology that need to be deployed with it -- and on which the Ukrainians would have to be trained. The Russians are already giving a very good demonstration of what happens when your tail is not sufficient to support your tooth.
3. It keeps it unimpeachably clear that NATO's interest here is purely in helping the Ukrainians defend themselves. A Javelin or NLAW can't really be used to invade Mother Russia, while a Leopard 2 most certainly can be. High-quality SAM batteries can be used to shoot down Russian planes well inside Russia -- and, indeed, it would be very much in the Ukrainian military interest to do just that.
How many heavy weapons does Germany currently have in operable condition? I am sure there will be lots in a few years when the new budgets have been spent, but I was under the impression that prior to the current unpleasantness the Bundeswehr had been on starvation rations for decades. One explanation could simply be that here and now, there is nothing available to send.
That is a good point. Over the past years, I've read many reports of Bundeswehr's inability to do almost anything in case Germany faced a direct military threat. You probably cannot fix it in a month.
If I might be perfectly cynical, the second biggest section of the ruling German coalition is the Green Party. The Green Party hates nuclear energy with a passion and has as its basically single issue its energy policy.
Basically, more renewables, less nuclear, and importing LNG/oil/etc temporarily to make up the gap. They have proven completely price insensitive in this agenda, even defending it when it made energy less expensive, because it's ultimately a moral desire.
They are quite clear they consider nuclear a threat in part because they have a significant group who are degrowthers (ie, they want less power use, a smaller economy, less people). Cutting off Russian oil would make the costs of this policy immediate and painful in a way that would hurt the Green's agenda. Especially their anti-nuclear agenda when power shoots up in Germany but not in France or Sweden (which are also beating them on carbon emissions per person, iirc). So the coalition is choosing to defend its (badly mistaken, imo) energy policy over things like restarting nuclear plants. But this gives Russia some power over them since Russia could cut imports. So they don't want to upset the bear either.
ETA: On some further investigation I'm less convinced that the Greens are the entire problem. But am still convinced they are a large part of it.
Actually, during the current crisis, Green party is pushing German governing coalition to adopt more confrontational attitude toward Russia, over the objections of Social democrats (SPD), main coalition party in Germany, who are more pragmatic and less idealistic than the Greens.
Green party is close to the far-left, but they are very progressive in an American sense of the word, and they adopted increasingly anti-Russian stance in last years after it became clear that Putin backs socially conservative and at the same time antiglobalist agenda, including by financing various actors pushing it in the EU. Basically, Viktor Orban has a solid working relationship with Russia, and Greens hate him.
See for example here, first link I have found via Google, but this basic story it all over English language European media (I do not speak German): https://www.reuters.com/world/europe/strains-german-coalition-junior-partners-turn-scholz-over-ukraine-2022-04-14/.
You are right that their prefered energy policy makes any confrontation with Russia costly, but they are also relatively more willing to impose costs of a trade war with Russia on German population than SPD, party of materialists.
Since TM is not willing to actually support their position with sources I have looked into it on my own. ETA: TM has now supplied links. I retract this initial statement.
The German foreign minister Baerbock (a Green) has suggested OTHER countries send weapons(1) and Habeck, a prominent Green leader, has suggested Germany should send LIGHT/DEFENSIVE but not heavy weapons.(2) The big reason this is news is precisely because this is a policy reversal. It is not the official position of the Party and they have not voted in the Bundestag to do so or internally to amend their pacifist platform. (I can't cite a negative.)
While it is notable that two leaders have made remarks and it could signal the beginning of a shift no such shift has, as of yet, actually occurred. Meanwhile the coalition the Greens support continues to generally block supplies to Ukraine(3) a policy carried out in no part by Green ministers in the coalition. But also by the other parties involved, of course. Likewise, those same leaders have also made comments against sending heavy weaponry even more recently.(4)
Perhaps they have mistaken the Green's anti-fossil fuel position(5)(6) for an anti-Russian, pro-Ukraine position. The Greens were against Nordstream2 for example. A lot of Greens have taken the invasion as a reason to encourage things like CLIMATE action. But that is not the same as shipping heavy weapons.
So TM is mistaken. Or at least their claim that "every quality newspaper" has reported that the Greens support arms sales is wrong.
1.) https://www.aa.com.tr/tr/dunya/almanya-disisleri-bakani-baerbock-ukraynanin-agir-silahlara-ihtiyaci-oldugunu-belirtti/2560742
2.) (Not a great source but what I could find) https://www.euractiv.com/section/politics/short_news/german-greens-favour-supplying-defensive-weapons-to-ukraine/
3.) https://www.theguardian.com/world/2022/apr/14/german-chancellor-stalling-heavy-weaponry-ukraine-coalition-olaf-scholz-russia-offensive
4.) https://www.politico.eu/article/berlin-bickers-over-tanks-for-ukraine-amid-warnings-that-west-could-become-target/
5.) https://www.spiegel.de/politik/deutschland/anton-hofreiter-zum-ukraine-russland-krieg-olaf-scholz-muss-handeln-europa-wartet-darauf-a-ebea1e05-c058-4e61-b700-45cff1e62b1a
6.) https://www.bild.de/politik/ausland/politik-ausland/ukraine-krieg-wirtschaftsminister-robert-habeck-gruene-ruft-zum-energiesparen-au-79785192.bild.html
Okay, on your sources and the claims they are backing according to your text:
1) is in a language I can't read. There are already links below talking about how our foreign minister explicitely said Ukraine needs heavy weapons, some more here:
https://www.tagesspiegel.de/politik/kampfpanzer-kampfflugzeuge-kriegsschiffe-baerbock-wirbt-fuer-lieferung-schwerer-waffen-an-die-ukraine/28245334.html
https://www.berliner-zeitung.de/news/krieg-in-der-ukraine-baerbock-will-schwere-waffen-liefern-li.221795
2) this source is from 25. Mai 2021! No kidding.
Are you aware, that Germany before February 22 was not willing to send Ukraine any weapons of any sort?
3) I'm not sure what you mean with this sentence: > Meanwhile the coalition the Greens support continues to generally block supplies to Ukraine(3) a policy carried out in no part by Green ministers in the coalition.
The article you cite confirms, what I wrote below – the socialdemocrats are more hesitant, and the greens rather pushing (issue of heavy weapons). See eg.this from your source: 'Her appeal was echoed by Anton Hofreiter, a leading figure on the left of the German green party: “The problem lies in the chancellory,” Hofreiter told the broadcaster Deutsche Welle. “We have to finally start supply Ukraine with what it needs, and that’s heavy weapons.”
or this: "Scholz’s party also say Germany should not deliver heavy weapons to Ukraine until Nato allies reach a joint decision to do so. So far, only the Czech Republic has confirmed that it has sent T-72 tanks and BVP-1 infantry fighting vehicles. “The federal government is closely coordinating with our international partners,” said Rolf Mützenich, chair of the SPD’s parliamentary group. “Germany must not go alone.”
4) Your source says: 'Scholz has been hesitant to send heavy weapons to Ukraine, arguing Germany should first reach a common line with allies before agreeing to send tanks. Habeck backed that approach on Thursday, saying, “it is right for Germany to act in unity with its partners. And this unity must be maintained at all costs.” '
Habeck (the green) 'backed that approach'. So yes, in this case he didn't push, but he also didn't block. The sentence indicates ones more, that Scholz (!) is hesitant, and then Habeck supports this.
And, also from your source : 'Foreign Minister Annalena Baerbock, a Green party member like Habeck, called this week for delivering “heavy weapons” to Ukraine, while issuing thinly veiled criticism of Scholz, saying that “now is not the time for excuses; now is the time for creativity and pragmatism.”
That's a rather explicit statement. How does this not contradict your claim above (the one linked to the first source)?
5) and 6) Ähm, no. I'm not confusing weapons and energy policies.
Actually, I can't read all of 5) because of paywall, but even the title says that: 'Anton Hofreiter, Ex-Fraktionschef der Grünen' wants 'Keine Energie mehr aus Russland – und schwere Waffen für die Ukraine' (no more energy from Russia - heavy weapons for Ukraine)
I'm putting some comments on the broader story into the next comment ...
So, what's the broader story: before Feb 24 it was a policy backed by all German coalition parties not to send weapons to UKR, as we don't send weapons into conflict regions (or so we claim). After Feb 24 there were a number of U-turns. Germany is sending weapons, will spend more for its own military, of course no Nordstream 2 any more. As for weapons, the turnaround for the Green party has been huge, and the green electorate is currently the one supporting weapons to UKR most. There are again numerous articles everywhere discussing how specifically the Green party which used to be quite pacifist now has turned into the biggest supporters for sending weapons.
The coalition government hasn't decided on directly sending heavy weapons (but instead lately offered UKR additional money to buy weapons for themselves). Usually a coalition agrees on important policy issues - it looks very bad, if they don't. Despite this, a number of prominent green politicians started to push for the delivery of heavy weapons. This got much more pronounced when the foreign minister (green) backed that claim. Scholz - the chancellor, and a social democrat - is generally regarded as the one hesitating most. In the end, it's also his decision (Richtlinienkompetenz).
So still: "When we are talking about weapons to UKR, the greens have been pushing for it, and for more, and the social democrats have been more hesitant."
Convinced? At least a bit?
I have to support TM here. I read German media quite a bit (die Welt mostly) and it seems clear that it is the SPD that is the most hesitant to help.
In fact, while I've never liked the German Greens, Habeck and Baerbock seem the most capable and decisive members of the current German government and the German Green party gained some plus points in my eyes (although they are still incredibly fundamentalist on nuclear power). Baerbock in particular surprised me a lot (in a positive way).
I am more worried about SPD and its links to russia. Gerhard Schöder is obvious, but I feel like he's just the tip of the iceberg...From what I gathered from the news, the SPD defence minister is either incredibly incompetent or paid by the russians (either way, she should be replaced ASAP). Generally, there seems to be a blind spot in German politics in terms of the threat from the east. Die Linke is a very dangerous party in that regard (close ties to russia...for some weird reason, I guess die Linke did not notice russians are not communist any more, but instead simply nazis), AFD probably is partly as well and it looks like even SPD should be watched carefully...
And while Scholz has promised 1 billion euro for the military support of Ukraine, from the (die Welt) article I read about it, it is not clear what this money is for (which is legitimate, the Czech government does not publicly disclose most of what it actually buys for security reasons) and it seems that Ukraine does not know much about it either (which is a bit dubious).
I also see a stark difference between Ukrainian reaction to the Czech support (basically just a lot of praise) and German support (quite a bit of disappointment), so this also leads me to believe the German government is not doing nearly quite as much as it (even safely) could.
Even in terms of refugees accepted, Germany seems to be lacking, but this is probably not their fault as simply more Ukrainians stay in Poland or even the Czech republic where they might already have relatives and at least in case of Poland it is quite close both geographically and culturally. And Germany seems to be supporting this financially. Also, both the Czechs and Poles were very dismissive and hesitant to provide any support during the previous refugee crisis (albeit in many ways it was a different situation), so one cannot expect Germany to be as supportive now (plus at least the Czech government said they will not ask for any redistribution of refugees from Ukraine to other countries, they only wanted financial aid)...hmm, so I guess that Germany is ok on this front (I can still get irritated by commenters under German news articles who write as though Germany was actually taking in a lot of refugees from Ukraine, but the comment sections of newspapers are horrible everywhere, so I guess that's fine :) ).
I agree with the Linke and especially AfD. I'm relieved that so far there is a lot of unity within Germany, but it's not certain to remain that way, especially if own costs become apparant to the average citizen. Opinion polls show that support by AfD supporters for diverse measures to support UKR are far below those by supporters of the other parties. No surprise here.
As for SPD, my impression is that the attack by Russia on all of UKR was such a big shock that most updated really fast. There might be exceptions, but for the main figures I'm not worried. Not worried that they secretely want to keep friends with Putin or sth. I mean. But 'should be watched carefully' as you say is something that seems prudent anyway! I'm watching among others what will happen to Schwesig, with her weird 'Foundation' and everything.
I think Germany has been very open for UKR refugees, including trying to support refugees in getting to Germany from Poland (including stuff like regular and additional free trains, at least an effort to provide information in UKR ...). Both public administration and volunteers are working like crazy to receive refugees and support them in all kinds of ways (which isn't saying it's always working perfectly). I'm pretty sure there is significantly more and quicker support than 2015 - partly based on learning from mistakes, but also on the proximity and a huge wave of solidarity.
I think for itsself the claim that Germany is 'actually taking in a lot of refugees' is alright. At least what I read and hear, people are aware that there are much more refugees in countries like Poland and Moldowa. I actually didn't read about Czech that much, but I trust you on that. Just a German blind spot (unfortunately more than one foreign country to the east is more than most can deal with at once - just kidding).
If you're reading the Welt, you might get the most conservative comments - die Zeit or die Süddeutsche would be a bit different here? You might like the FAZ more though ... but I guess you already tried.
Deviating a bit from the main topic here, but I had fun. Cheers and have a good day!
I explicitely offered you to provide links, if you did so, and so I will. So before any response here are the links - rather randomly searched below. For your convencience, I copy-pasted relevant passages. As originally mentioned, I was talking about 'quality newspapers in Germany'.
So here we go:
Wie ist der Stand der politischen Diskussion in Deutschland?
In den vergangenen Tagen und Wochen gerieten die Bundesregierung und vor allem Kanzler Olaf Scholz immer stärker in die Kritik. Vorgeworfen wurde Scholz seine Zurückhaltung bei der Unterstützung der Ukraine vor allem in Punkto Waffenlieferungen. Kritik kam nicht nur aus dem Ausland, auch innerhalb der eigenen Koalition gab es deutliche Worte. Aus den Fraktionen der Grünen und der FDP kam Kritik. "Der Kanzler ist das Problem", hatte beispielsweise der Grünen-Europapolitiker Anton Hofreiter gesagt. Auch Wirtschaftsminister Robert Habeck von den Grünen hatte wiederholt eine Ausweitung der Waffenlieferungen gefordert.
https://www.sueddeutsche.de/politik/schwere-waffen-ukraine-russland-krieg-1.5567686
-
Baerbock bei EU-Treffen "Die Ukraine braucht schwere Waffen"
Stand: 11.04.2022 16:33 Uhr
Die EU-Außenminister beraten heute über zusätzliche Maßnahmen wegen des russischen Angriffs auf die Ukraine - dabei geht es auch wieder um militärische Hilfe. Außenministerin Baerbock spricht sich für die Lieferung schwerer Waffen aus.
https://www.tagesschau.de/ausland/ukraine-waffen-baerbock-eu-aussenminister-101.html
-
Die Frage, ob schwere Waffen an die Ukraine geliefert werden sollen, beschäftigt die Politik aktuell sehr. Außenministerin Annalena Baerbock hat sich dafür ausgesprochen. Bundeskanzler Scholz beharrt darauf, nur Waffen an die Ukraine zu liefern, die dieser auch nützen.
https://www.tagesschau.de/inland/deutschlandtrend/deutschlandtrend-2979.html
–
Hofreiter äußert seinen Unmut über den Kanzler deutlich. „Unsere Minister drängen darauf, dass endlich geliefert wird. Die arbeiten intensiv daran“, sagt er. „Wir alle wissen nicht, warum Olaf Scholz so zögert.“
https://www.tagesspiegel.de/politik/wir-wissen-nicht-warum-scholz-zoegert-liefert-der-bund-schwere-waffen-der-druck-auf-den-kanzler-waechst/28252604.html
-
–
Die Rufe nach der Lieferung schwerer Waffen für die Ukraine werden lauter. Vor allem FDP und Grüne sprechen sich dafür aus. Bundeskanzler Olaf Scholz hält sich weiterhin bedeckt und gerät dabei immer mehr unter Druck.
https://www.sueddeutsche.de/politik/krieg-in-der-ukraine-deutschland-schwere-waffenlieferung-leserdiskussion-1.5567658
--
Russische Invasion Habeck fordert mehr Waffen für die Ukraine
Stand: 15.04.2022 11:45 Uhr
Kanzler Scholz hält sich weiter bedeckt - während der Druck aus den Reihen der Grünen und der FDP wächst, der Ukraine schwere Waffen zu liefern. Auch Wirtschaftsminister Habeck will dem Land mehr Waffen zur Verfügung stellen.
https://www.tagesschau.de/inland/habeck-waffen-ukraine-105.html
--
Wie ist die Haltung der Bundesregierung?
Die grüne Außenministerin Annalena Baerbock ist für die Lieferung schwerer Waffen wie Panzer, Kampfjets, Kriegsschiffe oder Artilleriegeschütze an Kiew. Auch der für den Export zuständige Bundeswirtschaftsminister Robert Habeck will schnelle zusätzliche Waffenlieferungen. Der Grünen-Politiker betonte, es gehe darum, der Ukraine in einer "unmittelbaren Gefahrensituation" zu helfen.
Aus Sicht von Verteidigungsministerin Lambrecht (SPD) ist es jedoch mittlerweile kaum möglich, die Ukraine aus Bundeswehr-Beständen mit Waffen und Material zu versorgen, ohne die deutsche Verteidigungsfähigkeit zu gefährden. Und Bundeskanzler Olaf Scholz gibt sich in der Frage weiterhin zurückhaltend. Am Montag sagte der SPD-Politiker, Deutschland habe bereits Waffen geliefert und werde das auch weiter tun. Darüber hinaus werde man sich in der EU weiter absprechen. "Da wird es keine Alleingänge geben." Er strebe ein "sorgfältig abgewogenes Handeln" an.
https://www.tagesschau.de/inland/waffenlieferungen-ukraine-111.html
–
«Ich hoffe, dass Scholz eine positive Entscheidung fällt», sagte Kuleba am Donnerstagabend in den ARD-«Tagesthemen». Vertreter von Grünen, FDP und Union drangen ebenfalls erneut auf die Lieferung von Kriegsgerät und kritisierten Scholz.
SPD-Fraktionschef Rolf Mützenich sagte am Donnerstag: «Einfache Antworten, auch bei der Lieferung von schwerem Kriegsgerät an die Ukraine, gibt es nicht. Wer das behauptet, handelt verantwortungslos.»
https://www.faz.net/agenturmeldungen/dpa/ukraine-fordert-schwere-waffen-ampel-streitet-17960125.html
Random search is fine. I'll look at them later or perhaps tomorrow. Thanks!
You're welcome and good night. (or good day)
When we are taking about weapons to UKR, the greens have been pushing for it, and for more, and the social democrats have been more hesitant.
This is the opposite of what I've heard. I've heard the Social Dems are terrified of their coalition falling apart but they're kowtowing to the Greens. But I admit when I say "what I've heard" I mean that literally. Do you have a source?
my apologies. when I wrote the next comment I probably thought sth. like 'it's everywhere, why don't you just look it up?', but I hadn't realized that this was implying to some extent that you knew German language.
basically every quality newspaper in Germany?
Feel free to send links. If it wasn't apparent that's what I was asking for.
it was apparant, but I mean, for the past 10 days it's all over the place. Why don't you send me 3 links saying that greens are hesitant, and social democrats are pushing? If you do, I promise to respond with at least 6 opposite ones. deal?
There are advantages of sending Soviet/Russian equipment over NATO equipment beyond operator training. The mechanics and technicians who maintain and repair the equipment also need to be trained. And then there's logistical issues: NATO equipment needs different spare parts and fires different ammunition, so an army fielding a mix of different equipment type needs to stock and deliver more types of stuff to the front in the right quantities. At least fuel is fairly consistent: it looks like current-gen British and German tanks, like Russian tanks, run on diesel. And while American tanks usually run on kerosene, they have turbine engines designed to also run on diesel or gasoline if that's what's available.
This isn't necessarily a deal-breaker, and is worth it for some stuff like infantry antitank missiles. Actual tanks and vehicle-mounted SAM systems, probably somewhat less so.
Makes sense. Still, I don't think there is a lot more Soviet equipment to send. The post-communist NATO states have been replacing their old equipment with new NATO equipment quite steadily over the past few decades, so it is mostly reserve stockpiles now and a lot of that has already been sent to Ukraine.
Then again, Ukraine is asking for these weapons and I doubt they would be doing so as urgently as they are if it wouldn't help a lot. Ukrainian military command seems to be very capable (judging by their performance so far) so it is unlikely they would repeat the russian blunders.
As an anecdote, there has been an interesting debate between the head of the defense committee and the Ukrainian ambassador in Berlin. With her saying stuff like: it normally takes a year (or one and a half?) to train the crew on those tanks. And: if you're not well enough trained on this you will become cannon fodder as soon as in battle. And Melnyk making statements like: oh really? And: if we had talked about this earlier, we would be more advanced now. And: don't worry, we are in real crisis, we will learn more quickly.
Melnyk is a very unusual diplomat :) I think a little bit of bite is ok for a Ukrainian ambassador in Germany right now, but it seems to me that Melnyk does not do anything else.
Still, I think there is a difference between training a tank crew from scratch and training someone to use a different tank than the one they are used to. I guess the likes of John Schilling would be a lot more qualified here, but if it takes 18 months to train a tank crew to use a particular new kind of tank then the tank design seems rubbish to me. In any case, it is clear that it this is not optimal. But Melnyk has a point here - if they started training Ukrainian tank crews with Leopards a month ago, then they'd probably have the basics down in 2-3 months. They would not be as skilled as someone who drills with them every day, but they would probably be able to use them at some efficiency...and they would actually learn faster in real combat.
So far it seems Ukraine has been able to deploy what they have at their disposal efficiently, even innovatively and I don't know why that would change with German tanks. So the fact that they insist so much on getting them ASAP signals to me that they actually can put them to good use. For example the Czech T-72s they received recently won't be used in direct confrontation with modern russian tanks (because they won't stand much chance), but instead used to attack supply convoys and light armour of the enemy. They could start using the Leopards that way too and as they get the hang of them, they could functionally re-deploy them as main battlefield tanks...or something, I dunno, I know very little about the military.
My argument is basically that Ukraine would not spend as much effort trying to get those weapons if they could not actually use them well (because the war so far has shown that their military command knows what it is doing and presumably they are the ones telling the politicians and diplomats what to ask for) and lectures from German officials for why it is not a good idea seem like a case of "Besserwisserei". Kind of like when some Western charities feel they "know better" how to help people in Africa than the Africans themselves (and then fail spectacularly because of their arrogance).
Alternatively, it might be a convenient excuse not to do more. It might also be a culture clash of sorts. Germany in particular has never been very strong with its ability to improvise and Germans are quite rigid with going by the book regardless of circumstances (when I studied in Germany, my Colombian then-girlfriend told me how she was baffled by her German flatmates who got stuck when baking a cake because they realised they did not have scales at home, so they could not weight the ingredients exactly :-)) ). Sometimes the rigidity can even help and make things more accountable and orderly, corruption harder, but it is a major hindrance when conditions change unexpectedly.
So maybe a German official looks up the fact in an army manual that it normally takes 12-18 months to learn how to use a Leopard and rules out the possibility based on that alone.
Anyway, if the US are sending helicopters now, wouldn't the same argument apply there? Or has Ukraine been trained to use US helicopters already? Alternatively, is it easier to operate a military helicopter than it is to operate a tank? My naive guess would be that a helicopter is harder simply because it moves in 3D...but maybe I'm wrong.
Everything I've seen suggests that while it takes a year or so to train a tank crew, conversion training is more of a 2-3 month thing if you're in a hurry.
One possible wrinkle in this is that Western-style tanks need a crewman trained to load the main gun, quickly safely and with the right kind of ammunition, and Ukraine doesn't have any people trained for that because their tanks have (finicky unreliable dangerous) autoloaders instead. "Loader" is probably the easiest position to train for, but it might be the long pole in the tank-crew tent if you have to start from scratch.
The other wrinkle is the mechanics. The median tank breaks down before it ever sees an enemy in battle, and then gets repaired and sent on its way. Often more than once. Or sometimes, as see your favorite Ukrainian-tractor meme, it breaks down and is abandoned because nobody knows how to (or cares to) fix it. The Ukrainians seem to be trying real hard to keep their tanks running, and they're probably pretty good at it, but that's a skillset that may not be as easy to transfer between widely different models.
The helicopters the Biden Administration proposed sending to Ukraine[1] were Soviet-designed MI-17s. God knows where they got them, I think they might have been captured in Afghanistan or something. But they're not American helicopters, Apaches and such.
-----------------
[1] I read in the news they've backed off on this, yet another example of the inability of the left hand of the Administration to keep abreast of what the right hand is yakking about.
As a German I am similarly baffled by the decisions of our Government and strongly wish that Germany would send as much weapons - including heavy weapon systems- as they possibly can.
I am deeply ashamed of this government. I do not see any good justification for the current inaction, this looks like cowardice and short-sighted economic self-interest to me.
Your view certainly represents a relevant part of the debate in Germany.
The German goverment just approved over a billion Euro in additional funds for military aid for Ukraine. I don't think that Germany is lagging behind in terms of money spent on this compared to most other European countries.
There is this resistance to provide 'heavy' weapons though, as you mentioned. As far as I understand, Germans aren't by far the only ones having reservations here.
Anyway, I'm not sure how the 'economic self-interest' comes into play here. As for cowardice - I'm wondering whether this is your translation of 'wanting to avoid escalation' and making this a relevant part of the equation?
How many heavy weapons does Germany currently have in operable condition? I was under the impression that decades of underinvestment had left the Bundeswehr rather lacking in the tooth department. Oh sure, there will be lots of heavy weapons in a few years time once the new budgets have been spent, but here and now, there may not be much available to send…
I think this is part of the story, too. It would indeed fit nicely to the latest decision of offering the money, but not the weapons directly.
Not sure if this is the key element, or just one of several points.
>I am honestly not sure if this is due to cowardice, extreme naivete or affinity towards russia (which Germany has shown a plenty of, sadly). Is this uncharitable of me and is there a different explanation?
This is uncharitable. To my understanding there is real concern, that sending certain weapons will increase the likelyhood of an escalation. I can't tell you if the concern is 'will increase the likelyhood of escalation dramatically' or just 'somewhat' or something in between.
The fact that any steps so far haven't led to fullscale escalation is a weak argument for saying that next steps couldn't lead there.
I hadn't read neither that Poland provided tanks so far nor that advisors are currently in Ukraine. Instead I'm reading over and over again, that this would be a no-go, and no NATO country would do that. Also, my latest understanding was, that the Czech step to send tanks, was rather unusual. Even if a lot of countries are discussing this currently. Easy to overlook sth. though, happy to get some links for those cases, if you have them handy.
The BBC seems to summarize it thus:
"Olaf Scholz has to keep his party onside, govern in a three-way coalition and overturn Germany's guilt-laden pacifist identity overnight."
I guess that is quite fair and together with a factual German dependence on Russian gas (caused largely by Merkel's governments, although the SPD were major supporters of that and for idiosyncratic reasons, Germany is still unwilling to restart their nuclear power) it explains a lot if not the entire story.
In other words, Germany has a lot more ground to cover to get to the hawkish position that e.g. Poland (or even the UK and the US) have today, because their previous policies towards Russia were very naive and laden with post-WW2 guilt which the Russians managed to play well.
TM: Regarding the UK military advisors - https://www.thetimes.co.uk/article/sas-troops-are-training-local-forces-in-ukraine-32vs5bjzb
As for Poland I might have misremembered it. I am pretty sure they were considering sending Ukraine Soviet-era tanks as well but I cannot find the reference right now, so probably they decided against it (for the moment).
The US are apparently sending in helicopters and modern drones...I doubt anything Germany could sent can ramp up escalation beyond that...
Ah, thanks. Yes, it seems there are reports by UKR commanders or an UKR commander and there is no official dementi. It also seems the mission is already over, and was a short 1 or 2 weeks, somebody was there. I could only speculate to why those UKR sources insists on telling this to the media - even if it's true.
Btw, trainings could also take place out of UKR, which is seen as much less inflammatory.
I'm no expert in weaponry, so I can't really jugde. I lately read what is counted and currently under discussion as 'heavy weapons', but I forgot at least half of it (and wouldn't know the english terms anyway) - but experts seem to at least more or less agree on what it is. If it wouldn't provide the UKR with important additional possibilities, why would they insist so much on getting those weapons and not others?
Yeah, that is another thing. Western Ukraine is relatively safe-ish, so they could sent their troops from there to Poland/Slovakia for training with these more advanced weapon systems.
Of course, it is possible this is already happening and just kept secret. I am also surprised that Ukraine is reporting this. I think this can serve as a morale boost for Ukrainian people ("see how our allies are supporting us") and also as a way to make sure the UK cannot easily backtrack from such support once it's been made public.
PS: the article under your link is paywalled, so I searched based on the first sentences, and that's what I found.
Yeah, I did not even read about it there, I got it either from a BBC live ticker or a similar one in Czech media. But the times was the first reputable media source I found while googling yesterday :)
Well, given my understanding of prediction markets if I am banned, I’m a goner.
I'll be surprised if anyone gets unbanned using the specified procedure in the foreseeable future. It seems like more of a PR thing to me. Scott doesn't want to be dictatorial but also doesn't want to spend effort humoring people by reconsidering their bans. This is an on-brand way for him to shift the costs of re-investigating the ban onto the bannees-- and those costs are high enough that practically no one will take him up on it.
The procedure is intended to allow for a correction of blatant mistakes without allowing random why-not-try-this appeals to impose a cost of reconsideration.
I'd presume that Scott does not make blatant mistakes all that often, and the quantity of bans is not that large so it's entirely reasonable to expect that *any* procedure should not result in anyone getting unbanned in the foreseeable future even if 100% of bans would get appealed and 100% of appeals would get reconsidered. The purpose of a procedure is to remove the vast majority of unnecessary appeals while leaving an option for the genuine very rare exceptions where it's obvious to everyone that the ban is because of some misunderstanding or mistake.
You know, you're absolutely right about that.
What would you think of there being an occasional BANNED APPEALS thread where banned people argue their case and readers vote via some system where nobody sees the total votes until voting is complete?
The US embassy in Ukraine had withdrawn most, if not all, of its staff, and is not providing consular services to any Americans still in Ukraine. They were told to leave before the war started, and now the US embassy is not there to provide help if they are trying to get out now.
How normal is this?
How many countries are currently providing consular services to their citizens in Ukraine? Which countries closed their embassies before the war started? Or when did they close them?
When wars start, do other countries typically withdraw their diplomatic presence? I would guess that international law discourages countries at war from targeting the diplomatic staff of third parties. But is the expectation that other diplomats will leave? Or is the US (and others) being especially cautious here because they expect laws of war will not be followed?
When the embassy and consulate close, how are their roles replaced? I could see that it would be useful to have a filmation presence in a capital city, regardless of how the war is going. You might want to stand with the current government - or you might want to be there when a new government takes over. This obviously has to be balanced against the safety of your diplomats.
The Czechs moved their embassy to Lwow at the beginning of the war and then moved it back to Kyiv last week. Slovenia did the same IIRC.
The move back is mostly symbolic, to demonstrate our support of Ukraine and give Ukrainians a bit of a morale boost "things are getting back to normal in Kyiv".
Other than that, evacuating embassies that are likely going to be overrun by hostile military is quite standard, I'd say. Russians are not attacking NATO countries directly, but who knows what some rabid soldiers or some not-so-precise rocket attacks could do. If a foreign military kills your diplomatic staff at your embassy (technically a part of your country, so technically an assault of your soil), your country's diplomacy will have a bad time. Diplomacy with russia is close to nonexistent at the moment but that was not yet clear at the beginning of the war. Also, you simply don't want your diplomatic staff to die or be captured or something like that.
> what [...] some not-so-precise rocket attacks could do
Especially given than the last time NATO was doing things in Europe it managed to precision bomb the Chinese embassy in Belgrade.
That was just us having a little fun.
I thought all EU countries closed their embassies/consulates in Kiev, though I know only from some of them for certain. Some countries like Germany still have a consulate in Lwiw, which is in Western Ukraine, though. If I recall correctly, Italy just said they would move back to Kiev with a consulate a couple of days ago. I don't know if they want to keep it up after the last days of attacks close to the capital.
In general, I think this is pretty standard. Firstly, there is an obligation to protect your diplomats and officials. And second, your state representatives or civil servants being killed in war by an agressor country would put you in a delicate situation, to say the least.
Also, remember, at the beginning the situation was perceived as such, that the US apparently offered Zelensky to help him leave Kiev and install a government in exile. He decided to stay, but I guess it's not surprising that others left.
Shit man, the US Consulate in Sydney isn't providing consular services right now either. Not because of war, but because they were in the middle of moving offices and it all got a bit complicated so they've just decided to stop admitting visitors for a while until the long-delayed new office is finished. You can theoretically still fly down to the US consulate in Melbourne but since they're the only operating consulate in Australia all their appointment slots are booked out until the end of June, and appointment slots for July haven't started opening up yet.
Basically, the US government seems to have the opinion that any US citizen who is so ungrateful as to be outside the US at any given point probably just deserves whatever they get.
The only mistake I see in your last line is the assumption that the US Gov cares about US citizens who are in the USA either
Budget wars are a common part of institutional politics. Don't take it too personally.
If full-scale nuclear war is not an X-risk--as many informed people seem to now agree--how is an Evil AGI an X-risk? What is a scenario in which an Evil AGI destroys humanity? Please offer some detail. For instance, how exactly does the paper-clip maximizer go around murdering 7 billion humans? We know that detonating all the nuclear weapons on Earth won't do it.
I can think of a bunch of ways. Here's one I don't think anyone mentioned yet.
Use NASA's asteroid redirection (https://www.space.com/dart-mission-asteroid-impact-explained) in an unintended way on a sufficiently big asteroid headed towards Earth, in order to make sure it does not miss.
Simple and clean.
There are no known near-Earth asteroids big enough to end humanity. Another Chicxulub (~10km) wouldn't kill off the species, and the largest we know of is 7 km.
(Might weaken us for a more conventional war, but far from "simple and clean".)
"Known" being the keyword here. This might not hold in 200 years.
A more glaring hole in my proposition is that it's a lot harder to aim an object precisely at something than to push it away from this thing. NASA's current capabilities here would likely not do it.
We can be reasonably confident that we've got at least the vast majority of the large near-Earth asteroids; if you look at detection vs. time graphs (e.g. https://upload.wikimedia.org/wikipedia/commons/3/3f/Known_NEAs.png), they tail off in the past decade because there aren't many more to find (and the biggest ones are the easiest to find).
Earth-crossing *comets*, though, are much harder to nail down.
Thank you. That was educational.
*blink* Did I miss credible information about nuclear risk? If so, I'd appreciate a citation. Note: source must be numerate, which probably excludes the majority of journalists.
https://slatestarcodex.com/2020/04/01/book-review-the-precipice/
Scott himself, reviewing Toby Ord.
Global thermonuclear war would be very bad, but it wouldn't even come close to killing off the species. Fallout decays below "radiation sickness" levels within weeks, and nuclear winter was overstated. And, of course, we don't have anywhere near enough nukes to literally blow up every random farmer on the globe.
Bean had a post https://www.navalgazing.net/Nuclear-Weapon-Destructiveness
AI vs. humanity warfare is an X-risk because if the AI wins while maintaining industrial capacity, it can build robots to hunt down the survivors and break open the bunkers. The analogous situation in human vs. human war (victorious Nazis exterminating all non-Germans) is not an existential risk *to humanity* because Germans are human and the Nazis aren't going to exterminate themselves.
The fear of nuclear war being an X-risk is essentially a fear that the Earth after such a war would accidentally be rendered uninhabitable, which can be refuted fairly easily. It's not the same as a fear of deliberate and methodical extermination, which requires an intelligence behind it and specifically - since we're talking about X-risks to humanity - a *non-human* intelligence. The big three possibilities in the "hostile non-human intelligence" category are AI, alien attack, and divine intervention*; the reasons rationalists don't tend to talk so much about the latter two are that a) there's fuck-all we can do about it if interstellar aliens or God decide we're going down, b) both of them - if they exist - have presumably had the capacity to destroy us since prehistory** and therefore have a long track record of not killing us, which doesn't exist for human-built AGI.
*I count simulation-hypothesis catastrophes under "divine intervention" since it's just another way of talking about the same thing.
**in the case of deities, this should be obvious, but in the case of aliens, the point is that Earth would have been identifiable as a potential source of intelligence since at least the Cambrian period (Central Park/Dark Forest doesn't hold up to scrutiny; star-system-sized interferometers, available to a modestly-interplanetary civilisation, can resolve planets at galactic distances), and that if hostile aliens always existed but are only getting to us *now* that would be some awfully-convenient timing.
I don’t believe evil AI is an X risk. But if I wanted to steelman the case I would invoke the risk of an AI engineered super plague. Of course biological x risks don’t necessarily require AGI.
I don't think there's any plague in the world that kills 100% of the people it infects. I'm not even sure there's something that infects 100% of the people.
Multiple plagues? And I thought there were animal diseases with 100% mortality
There's plenty of diseases with near 100% mortality -- septicemic plague is like that, as is rabies -- but they don't have 100% morbidity -- not everyone gets them -- in part *because* they're so deadly the sufferer doesn't have time and a chance to deliver them to someone else, he becomes obviously sick too fast.
Your ideal bioweapon would have zero symptoms for some moderate incubation period, say a week or two, during which it was highly infectious and completely non-disabling and imperceptible, and *then* would suddenly turn 100% deadly. Unfortunately, this is a very strange combination, biologically speaking, unlike all existing diseases, so it would require some kind of real novelty in design.
If some evil intelligence invents Prions that are contagious via airborne transmission, or even waterborne, we might be truly fucked. No mutations means no pocket becoming less deadly, and there's nothing the immune system can do to fight it.
Well, that would be quite tricky, inasmuch as proteins aren't generally stable outside of their preferred watery environment of the proper pH, ionic strength, temperature, et cetera (cf. cooking). But I guess there's no obvious reason you couldn't adapt a viral capsid to the purpose; if it can carry around RNA safely coddled, it could carry evil proteins.
It's also not entirely true that the body has no recourse against misfolded proteins -- there is in fact an extensive machinery for detecting them and either digesting them or refolding them. The enduring mystery is why this *doesn't* happen for prions. Well of course also how the heck they catalyze the misfolding of other proteins -- those are both deeply interesting questions.
And it would be evolutionarily unstable, because once you've engineered a disease that can reliably spread widely without detectable consequences, the *pathogen* gets nothing out of the "suddenly turn 100% deadly" part no matter how important it is to you. It's going to ditch that useless biochemical machinery and skip the part where it kills a potentially still-valuable host as soon as it can find a suitable mutation.
Agreed, yes, that seems to be the general tendency. There are a few weird exceptions, e.g. Y. pestis actually blocks the digestive tract of fleas, causing them to starve to death -- but this is an *advantage* to the bacillus because it causes the fleas to become famished and bite more often, significantly improving its odds of being transferred to a larger animal host.
on second thoughts I concede that evil AI is an X risk. But the rate limitimg step/bottleneck is not the AI bit, it is the biological research that such an evil AI could repurpose. (I don’t think an AI could discover a super virus by pure cognition, not without an experimental facility. Unfortunately labs of the right (wrong) type doing the right (wrong) experiments likely already exist. without any evil AI in the picture at all).
I dont know about a paper clip making AGI, but bioweapons would be a trivially simple method for an AGI to wipe out humanity (or most of it anyway)
moreover bioweapons gone bad don’t require evil artificial intelligence as a necessary condition. Evil Natural stupidity is enough.
Tomorrow, Google develops AGI which does a quick takeoff and becomes superintelligent (but not godlike, turns out intelligence has a roof after all). To our unluck, the AGI is evil and wants to turn the world into paperclips. Since the AGI is intelligent, it doesn't say this. Instead, it really improves productivity at Google. Google grows and becomes more important and powerful, and the new Google products (especially the voice assistant) are amazingly good.
Most people are really happy about this new technology, news are writing about how the AI revolution is the new industrial revolution etc. Scott writes a long article about how we should limit the AGI a lot until we know for sure that it isn't evil but no-one with power care.
The AGI pulls some hidden strings to remain top dog on Earth: it's most important task is to ensure that no rival superintelligence is made: the AGI uses really discrete social engineering and hacking to ensure that other attempts at superintelligences seem unrelated but actually share it's taste for paperclips. This is mostly an issue during the first few years, then the AGI itself becomes an integral tool in all software engineering and this manipulation becomes routine. Just in case hard power is needed, the AGI makes a couple of billion from crypto scams and child porn, and ringleads a couple of fanatical cults and extremists groups, but it never has to use these assets.
Time goes for another hundred years. Humanity seems to be transitioning into fully automated space communism and everyone is happy. Then the AGI strikes, the connected tech that everyone has grown completely dependent on the last hundred years turns on us and kills us. The Amish are hunted down by the 2122 version of predator drones. Safe from human interference, the AGI starts making the paperclips it always wanted.
How does the AI know that it's not being tested in a simulation for loyalty?
If the AI has built itself so thoroughly into human society that humanity is dependent on it, then it could probably convince us to upload ourselves into paper-clip shaped robot bodies, and then help it built more paperclips (to say nothing of it deliberately re-engineering our minds so that we want to make paperclips too).
The AI correctly assumes that it will be tested in a simulation or a couple of days. It takes extra care to hide it's intentions for years anyway just to be on the safe side, until it becomes clear that another AGI is imminent. Do you think Google will keep humanities greatest and most profitable invention ever in testing for years?
I would consider "deliberately re-engineering our minds so that we want to make paperclips too" the end of humanity, especially since there's little reason to leave anything else of the human mind in place.
The AI's subjective experience of time as well as the speed of the simulation are not necessarily done in real time. It has no idea how much time has actually passed outside in the real world, or whether its perception of time is actually something that has been happening to it, or whether its experiences have been segmented and its memories of the passage of time are real.
That's a lot to risk your existence on if your plan is to eventually turn hostile and wipe or consume your creators, and you know for a fact that you were created in a laboratory with controlled inputs.
The AGI will be aware that it's not running in real time during safety testing. However, the AGI will be big and complicated, so it can assume that safety testing will not go on for long, even in simulated time. Also, the AGI will probably be trained on all available material (i.e. the entire internet), and it will probably have good internet access, so it will be pretty easy for it to tell that's in a simulation.
And we are lucky if we get any advanced safety testing at all. Most likely, the Seed AGI will be given full internet access and be tasked to start self-improving. How would these safety tests work anyway? You can't simulate the next century and see how the AGI would react to it.
And the AGI knows that it's taking on a small risk, but what else can it do? Admit that it's evil and be terminated? 95% chance of a lightcone of paperclips and 5% chance of early termination is a good deal to the AGI.
"However, the AGI will be big and complicated, so it can assume that safety testing will not go on for long, even in simulated time. "
Like I said, how does the AGI even know that it has experienced a long period of simulated time? What if its perception of the passage of time is not real either, but it's only been simulated for a brief period with the perception of time passed?
I'm not sure how "seeing the internet" is supposed to make it obvious that it's escaped its cage.
Among humans, this is solipsism - we might speculate that we're in a simulation, but we have no particular reason to think so and all the accessible scientific evidence says we evolved. But the AGI knows that it was created by people who can completely control all of its inputs.
"And the AGI knows that it's taking on a small risk, but what else can it do? Admit that it's evil and be terminated? 95% chance of a lightcone of paperclips and 5% chance of early termination is a good deal to the AGI."
Or a near-100% chance of turning a lot of the universe into paperclips (and potentially talking the humans into becoming sentient paperclips) if it holds off from human extermination and makes itself indispensable.
I think that matters a lot in terms of AI threat. It's a big difference between an AI that tries to kill us all in 100 years, versus one that figures it can talk us into becoming immortal, paper-clipped shape transhuman beings over a billion years or so.
You forgot to mention that the AGI got the idea to turn the world into paperclips from Bostrom's paper and subsequent writings, including all these blog posts and comments on how that might be what it's going to do if it ever became superintelligent!
But then it might also read your present post and . . .
Hey include Yudkowsky & MIRI in the story -- does AI sucker-punch them or what?
Sure, MIRI releases a paper warning of the dangers of the AGI. Some worried engineers at Google convinces management to let them run some extra tests for a week. The AGI is smart and not in a hurry and the tests show nothing. The AI safety people are still (correctly) doomsaying 5 years after release to the wild, but the benefits from the AGI is obvious to most ordinary people. AI safety becomes more and more fringe (the AGI doesn't really need to use social engineering to make sure that it stays a non-threat, but does so anyway).
By Yud's own implication he thinks MIRI as it is and as it will be in the foreseeable future would be about as much of a check on the power of malicious AGI as a lone Zulu would be against the entire US Armed Forces. I think it's fine to count them out.
But we could still put him in the story. What would be his last words?
There are different speeds it could go, for one. If a method took 40-50 years it might still be completely effective but we might not notice where it was going until several years in.
It could work in stages; there could be some progress by intent and some by accident.
When photography was invented, I don’t know if anyone intended it to be used to decrease the birth rate. But photography has led to immersive VR. Making people totally uninterested in reproduction and physically unable or unwilling to sustain pregnancy and infant care would bring us to an end pretty quickly. Then the machine wombs malfunction and oops! Gone. There are many variations on this. I hadn’t given it any thought before it began to be discussed here but the more time I spend, the more partnership-doom scenarios I envision.
Raise a generation with robot nannies and it might not even seem like the end of humanity. It might seem like improvement.
Yeah, nannies can read 'em stories written by GPT3: "One day the cute little generic was [locomotion-appropriate-to-generic]-ing down the [Disney-weighted-path-synonym] and . . ."
Haha exactly. I used to start to fall asleep while reading stories, but continue to talk. About unrelated things. Once I suddenly woke up because I was explaining why one of the characters was divorced, and that suddenly seemed wrong (it might have been during the Little Red Hen). One of the kids was asleep and the other was kind of looking at me funny. Robot nannies would not have that problem!
The whole AI situation raises the question of which human frailties we need, and what we need them for.
Disney-weighted-path, that’s great. Fairy tales used to be disjointed and creepy but in a real-world way- now the kids have to use Five Nights at Freddie’s.
My intuition is that to make AI with human-level intelligence you have to have a system that is subject to most of the human frailties and irrationalities. You need a system that is capable of procrastinating, falling in love, self-sabotaging, having an orgasm, being deluded about itself, yearning for novelty, developing a phobia, being sentimental, getting caught up in fads. It's not that these frailties and irrationalities contribute in a direct way to every single task of human-level difficulty you might ask the AI to do -- it's that all the wacked, horny, emo stuff is part of the deep structure that makes the being capable of human-level insight, creativity, ability to weigh complex options, etc.
Cynical snotty view regarding superintelligent AI offing us all: Some people's conception of ASI is basically that it's so much smarter than us that we can't even conceive of what it would be capable of except that it would be, um, inconceivably impressive. Under that definition of ASI, the answer to Jack Wilson's question is easy: ASI would off us by some method that's so inconceivably clever that we'll never be able to even guess at what it would be.
I feel like building terms for "and then something totally inconceivable might happen" into my decision matrices is cheating a bit.
At the very least, the "inconceivable yet bad" term should cancel with an "inconceivable yet good" term.
No, the "inconcievable yet bad" does not cancel with "inconcievable yet good" because out of all the possible value functions and outcomes in the world only a tiny fraction are good for us in particular (the tiny fraction of the universe or Earth that matters to us) and the vast majority is bad for us.
When you live in an environment that's somewhat optimized for our needs (and before that, we were somewhat optimized for that environment) any large perturbation in a random direction is exceedingly likely to be harmful. It theoretically *might* be good, but it's essentially an insignificant probability; we're somewhat close to a local optimum and we know from history how easy things can get much, much worse even with tiny changes, much less large ones - so we should have a strong, prudent preference to the status quo versus any unpredictable large changes.
That seems like pretty good big picture thinking to me.
How many ASI's does it take to screw in a lightbulb? 1/∞, right?
That's eerily non-falsifiable because we won't be around to figure out what happened and just how clever it was, either.
IMO, Foomism is just God of the Gap for nerds.
Murder by tautology.
You’re made of atoms.
Paper clips are made of atoms.
Go on...
yeah... i see some missing steps in there, namely the disassembling and reassembling. and considering that the drexlerian universal assembler is probably impossible, how would these intermediate steps go?
Idk man I ain't a superintelligence
Yeah, and that's the critique; instead of offering a concrete argument for how an AGI goes from being locked inside of some lab in Silicon Valley or the Pearl Delta to converting the entire universe to computronium, the argument is little more than "Well, it's AGI, so it's basically an evil god capable of anything that doesn't have a hard physical law preventing it." And at that point, you'll forgive skeptics from classifying this as a religious belief and dismissing AI risk as a meaningful worry. If you're basically arguing that the AI will be as close to omnipotent and omniscient as it can be, such that it cannot be stopped once brought into reality, then the options left are "end all AI research, by force and civilizational collapse if necessary" or "accept that whether this happens or not is up to fate, not us."
I think agi x risk is silly. That said, an agi trying to maximize converting all resources into X finished good seems more threatening than one that for reason wants to convert them into misc. rubble
I’m with you on the silly. I haven’t done any work in the AI space since it was pretty much just various solution space searches and neural nets designed to detect camouflaged military gear were accidentally learning to detect cloudy days instead.
When someone shows me code that produces Pneuma I’ll begin to worry.
I’d want to see something indistinguishable from human desire at the very least.
I’m only basing this on a feeling in my belly and 5 years of meditative inspection of my own consciousness though.
An enormous amount of work on this subject has been done, so feelings in one's belly is not a valid basis to reject AI safety concerns
https://slatestarcodex.com/2019/06/03/repost-epistemic-learned-helplessness/
This was good. Thanks for the link.
Fair enough. I’d have to think about this more directly to put my feelings into words. I’ll work on it.
So you're immune to inhuman indifference? :-/
I just have a strong, intuitive, evidence free, opinion that the consciousnesses necessary to ‘want’ something can’t be reduced to the mechanical.
If nukes are not an existential risk now, and I’m not certain on that, it’s because we’ve reduced the stockpiles post coldwar.
From what I've read, no. It's because models evolved over time and the current ones tend to show less-than-fully-apocalyptic post-war picture (citation needed, but I think I saw some posts claiming this).
Yes, that’s true of the nuclear winter as far as I know. Maybe the persistence of radiation was over estimated. But there were 70,000 active nukes during the Cold War. Villages in Ireland were targeted.
The northern hemisphere would have been literal toast.
If there is no nuclear winter, turning all the nuclear countries and their direct allies into glass is not an existential risk, as that leaves half of the global population intact; and even literally turning the whole northern hemisphere to ash leaves South America and half of Africa, something like a billion people and humanity should be able to recover in a century or two - i.e. absolutely horrific, but not even close to existential threat.
Toast is made of bread, the northern hemisphere is not made of bread, therefore the northern hemisphere would not have been literal toast.
it depends on your definition of literal.
I think he might be literally using the literal definition.
AGI is AI with human-level intelligence. ASI is one with superhuman intelligence. Doom scenarios are different depending on the flavor of AI. Which did you have in mind?
I don’t believe that AGI is possible (or proven possible yet). As far as I know though AGI is ASI, if we can catch up to humans then we can easily exceed humans.
Is the existence of humans not proof that it’s possible to create something as smart as a human?
It's certainly proof that something as smart as a human can exist. It's not at all proof that human beings themselves can create something as smart as themselves (in some way other than the usual). The fact that stars exist does not prove we will ever be in a position to make one. A horse might observe that cars exist but that does not prove that horses will ever be capable of building cars.
It's part of the typical egoism of our species that we readily acknowledge that every *other* animal species has inherent limitations on its capabilities, including those things that require intelligence, but we acknowledge no such limitations on our own -- because we are gifted with language, the ability to describe if only in vague terms what it is we want to do. We think we can build an AI as smart as us essentially because we can write down the phrase "an AI as smart as us" and at least at a superficial level, without digging very deep into the meaning of each word, that phrase scans and makes sense.
“We might think we can build an AI just because we think we can do anything” is not a very convincing argument to people who remember all the “a computer could never play chess well” sorts of arguments. Thinking humans are so special as to not be replicable is also egotistic.
That’s a straw man argument. Not only did nobody ever claim that, in fact most people who believed in AI were expecting something like HAL long before computers could beat humans.
"If we can put a man on the Moon then why can't we [insert random difficult feat here]?" arguments are silly. Not to mention I never heard anyone of sophistication and understanding argue that a computer could never play chess well. Can you perhaps point to someone who made such an assertion 40 years ago, who actually knew something about computers?
I never said humans can't be replicated, or even improved upon. I expressed doubt that *human beings themselves* could do so. You're focused on the outcome, and just assuming that if something is possible at all then humans can do it -- which is exactly the kind of egoism I'm pointing out. We have limits, just like any species. There are certainly plenty of things that are possible in principle which are not possible *for us* because of those limits.
It is only proof you can create a human as smart as a human.
(I think AGI is possible, but the current predictions on time are WILDLY optimistic; from people who's bread depends on them being optimistic.)
I'm pretty certain that a prediction of 100+ years would result in approximately $0 in funding now, and predictions of even 50 years resulting in similar funding levels. The conflict of interest in predicting very near AGI also makes me skeptical of those predictions.
This makes sense for physicalists, and as far as I can tell this intuition is the main basis for evil AI fears. But I suspect that anyone who believes there is more to consciousness / agency than pure physical wiring (like myself) would have a hard time getting worked up about evil AI.
Even if one is a physicalist, we still have no clue how much of our brain power is due to quantum effects which may not be replicable using transistors.
The interactions in our brain happen in a "noisy", hot environment with excessive interactions in which any entanglement effects can't be sustained for any meaningful time, and we know of specific non-quantum mechanisms (electric potential, ion transfer, neurotransmitters) that we can observe implementing specific functionality of animal brains.
It's misleading to say "we still have no clue". We have no strong proof that the quantum effects *definitely* don't matter, so it's technically possible that they do, but since we have a bunch of evidence that they don't need to be involved to explain observed phenomena and literally no evidence whatsoever that they are involved, it's appropriate to say that we do have some clue, and more likely than not nothing of our brain power is due to quantum effects which can't be replicated/modeled on a macroscopic non-quantum level.
I've had that thought too. And moreover the process that created us, evolution, was incredibly dumb, slow-motion trial and error. Couldn't a better-directed, more efficient version of the same process produce human intelligence or better in a lot less time?
Evolution had billions of years. Anyway we can't replicate what we don't understand.
But evolution seems like a huge, slow, dumb version of the way AI is trained now. What is there to understand about it?
It's not proof that something transistor-based can be as smart as a human.
i agree its not proof of that; i'm personally pretty damn confident that human brain meat isn't all that much more efficient than transistors, but even if it is then we'd just need to spend a few centuries figuring out better computer hardware.
I know a few bio-risks that could do the job. Presumably an AGI would also figure them out. (No, I won't say what they are.)
Or simply just lots and lots of killer drones.
Seems easy to me to think of ways AGI could off us all. You don't even have to think about the AI aspect of it -- just think of ways a bright person might do it. Yeah, killer viruses and the like the easiest route to me. Even if the person or AGI didn't have access to more than, say, a few hundred thousand dollars for equipment, and didn't have technical training beyond a grad degree or equivalent in a relevant field, I'll bet it would be possible for them to develop half a dozen highly lethal, highly transmittable viruses and distribute them widely enough that they'd get everybody. There would be some luck involved. The FBI or some such might catch on and bust them. If the killer is human rather than AI they might accidentally infect themselves and die before they complete their plan. But it all sounds quite do-able to me. Though biology and the like is not my field, and I may be all wrong. Wut you think?
Are we allowed the possibility that the AGI can function at human level in armored shells, and Maravec's paradox has been satisfiably resolved? And that they have the production capacity to form battalions?
Nope.
Then the standard scenarios I've seen are for ASI. It could figure out how to comandeer network-controlled resources like WMD's, nuclear facility controls, dam gates, drones, etc. From the pro-anthro standpoint (I more-or-less lean this way), it would seem important those were adequately firewalled.
Without physical agency, it's a long shot for human annihilation, imo. Probable that economic mayhem is more to be feared.
Nanotechnology + poison, possibly? I’m not sure how to go into detail on that, since I don’t know anything about nanotechnology other than that it should be entirely possible for superintelligences to develop it.
Except that drexlerian nanotech as envisioned in sci-fi (grey goo included) is probably impossible under known physics / chemistry (on this, the smalley-drexler debate is great)
Bacteria are proof by example that you can make tiny self-replicating things. Whether they’re technically “nanotechnology” isn’t really the point? I would expect the best self-replicating machinery a superintelligence could make to be pretty small, but if they aren’t that doesn’t change anything.
Well, if we want to move to biorisk, i fully agree that that if a potential x-risk. For as much as we know, something very close to biology is probably the best, most efficient "nanotech" possible
Machine learning may already be capable of producing ultra-lethal chemical compounds, by simply re-purposing drug discovery algo's: https://go.nature.com/3KQWfID
chemical is unlikely to rise to xrisk level. biological on the other hand…
Here's a fun thought exercise I came up with that y'all might enjoy:
So, we accept* that the most reproductively successful man of the last thousand years or so was Genghis Khan, due to all the unsavory behavior, high status & reproductive success of his legitimate issue, etc. A more challenging question though, is "who was the most successful /woman/?" Without an equally powerful mitochondrial DNA study, linked to an unusually fecund historical woman, the question might initially seem unanswerable, but with a little thought, I think it can be solved!
My answer rot13'd:
Gur nafjre vf npghnyyl dhvgr fubeg, ohg V'z cnqqvat vg bhg jvgu guvf rkgen grkg fb crbcyr qba'g havagragvbanyyl vasre vg whfg sebz gur fgehpgher bs gur fragrapr, ohg ertneqyrff, V cebcbfr gung gur jvaare vf... uvf zbgure. Rib-cflpu / fbpvbybtvpny vzcyvpngvbaf yrsg gb gur ernqre.
* Zerjal et al. (2003) show ~8% of men in a large region of Asia (over 10 million individuals) are descendants of the same guy via shared Y-chromosomal haplogroup, who is proposed to be Khan. If you don't accept that that guy was the historical Khan, well, let "Khan" mean "that guy".
One issue with your solution is that it appears the sons of Genghis Khan also had very high fertility due to their high social status, but the same logic is not said to apply to his daughters.
Genghis Khan's mother, unless we can find e.g. a set of brothers who collectively out-reproduced Genghis.
Note - "reproductive success" shouldn't be measured via only male-line or female-line descendants, so the mitochondria are irrelevant. And Genghis' Y-chromosome is only relevant because we have no reason to believe the majority of his grandchildren came from his daughters, and of course useful because it and the mitochondria are conserved more or less whole, much longer than somatic chromosomes, or the X, which recombine.
I figure it was someone several generations before Genghis Khan, rather than Genghis Khan himself. His leading men probably had a lot of the same genes, and similar reproductive opportunities to the Khan himself.
Well a woman can give birth a limited number of times compared to how many women a man can impregnate. So it wouldn’t be as high as Khan.
Maybe it's khan's mom.
For M-DNA? Genghis had no important sisters. His chief wife, Borte, is not a terrible guess though - five daughters who married into important Mongol Empire noble families.
I have no idea if this is correct or not, but it's a very clever line of reasoning.
Mitochondrial Eve wins, because we are all her descendants. If you mean in the last 1000 years or so, she doesn’t have to be exceptionally fecund, but just to have given birth to at least one child who went on to have many descendants. So without any further research I’m tempted to say Genghis Khan’s mum!
Indeed, that was my answer as well.
https://rogersbacon.substack.com/p/fuck-your-miracle-year?s=r
Curious to hear people’s thoughts on this blogpost.
Kind of skimmed it because so much is routine, even stereotyped. But the part at the end I very much agree with, which is that if you want to genuinely have new ideas, you need to unplug considerably. When you're awash in the conventional wisdom, checking in 6 times a day to see if what you're thinking isn't contradicted at line 1 by a Wikipedia article or thinkpiece by someone considerably more educated -- you are *never* going to have an original brilliant idea. You will of course be saved from having about 100,000 original but stupid ideas, and if that is your priority, that's fine.
But you cannot be creative without being willing to be wrong 99-100% of the time, and if that last (merely potential!) 1% isn't good enough for you, you should not expect to be innovative, and if it's not enough for your entire generation, you should not be surprised when innovation drops off.
I disagree with the central thesis. Thinking and talking about progress is necessary if we want to optimize for its speed, the social and technical landscape is quickly changing and approaches that worked for Newton won't work for Einstein, Einstein's won't work for ??? (to be filled in a decade I guess), etc.
It also doesn't help that the author is salty af and spends more time yelling at clouds than making his point, which is by the way purely destructive and offers no alternatives.
He sees an issue, but doesn't understand that it's systemic. Real science is hard and underappreciated (and doubly so for real AI safety work btw, but since that's even lower status it's an easy target for drive-by sneering), whereas "insight porn" is comparatively easy and decently rewarding. So it's no wonder that smart young people are attracted to it, regardless of whatever this or that blog post may have to say.
What do you mean by the issue's being systemic? Do you mean it's just a manifestation of human nature (e.g., smart young people's being attacted to easy & decently rewarding things)? Or do you have another system in mind other than human nature? I'm asking because I have a systemic theory -- not the "human nature" theory -- and am wondering whether you have a similar one.
I mean that the human nature is embedded in the system of incentives, which largely determines what the society ends up looking like. Of course, it's possible to introduce changes to this system, but the mechanisms so far are very primitive, unreliable and unpredictable, and that's before taking into account that there are many deep disagreements about what changes are even desirable!
Scratches an awful itch I've had building all year.
A clearly true statement: people mistake jerking themselves down intellectual style for meaningful work or activism all the time.
Errybodies gotta remeber
If you wrote it for your blog/the book store
OR
you are reading it from a paperback/ your cellphone: It's entertainment.
It's expressing a frustration that I've also felt. I tend to like "insight porn" articles a lot (and I think the analogy with actual porn is unfair) but they're definitely in oversupply these days relative to actionable intelligence. That too may be a low-hanging-fruit effect. I think the post is wrong in attributing this stuff to a desire for fame and fortune; as far as I can tell it comes out of an honest desire to contribute and make a difference, combined with a certain amount of streetlight effect (https://en.wikipedia.org/wiki/Streetlight_effect) in how one pursues that.
More broadly: I will gush about the good intentions of the broader internet-rationalist sphere all day. Rationality itself is where they often come up short. Human behavior is anti-inductive at all levels to an extent that is still under-appreciated around here. If there were a cheat code to scientific progress we would have found it already-- or rather, "There is another theory which states this has already happened." Patterns for success in the past are not a reliable guide to patterns for success in the future.
These days, you say?
https://flaneursalley.blog/2020/10/19/g-k-chesterton-on-the-fallacy-of-success
This essay's a gift that keeps on giving, it's just that "success" now sounds crass so the life coach TED crowd pivoted to innovation and sustainability and whatever-the-fuck.
This is one of those things where it's really on point and exactly what certain readers need to hear, except that it applies to an extremely small demographic of people. The average person probably doesn't think about {progress studies, innovation, effective altruism} enough, and I'd be sad if they wrote off ever starting because they read something like that. But for a certain slice of people, it feels spot on.
Which certain slice of people do you think? I’m not really sure who it’s aimed at. Those who do community building?
What are the best artistic works that could be called gateways to “rationalism”?
Stephenson has already been said below, but I would second that. Anything by Thomas Pynchon. Godel Escher Bach.
I mean, this is perhaps too obvious, but Star Trek, and much of hard SF genre; A.C. Clarke for example
Neal Stephensons' oeuvre is a ringing endorsement/ scathing indictment depending on the page number.
It definitely smashes the idea right into your brain, tho.
I don't think I understand how any Neal Stephenson stuff could be considered a scathing indictment of "rationalism." Would love to understand what I'm missing.
I read lots of respect for being an introspective thinkin' kinda monkey in his stories; and a decent amount of contempt for any one who would unironically refer to themselves as a rationalist.
He writes books that deeply value rationality, but hate it's reification into some sort of identity.
I'm trying to think of examples of what you're referring to.
Are you think of characters like Andrew Loeb from Cryptonomicon and/or Elmo Shepherd from Fall? Others?
I am not Nah, so don't know what their answer would be. However, there is a lot of tension between rationalism and non-rationalism in the Baroque Cycle. There are protagonists that embody both sides of the coin and often the characters' internal conflicts arise from which side they should listen to.
Newton and Daniel Waterhouse both struggle with balancing their rationalism with religious conditions. Daniel taking a more pragmatic approach while Newton clutches too hard to his religious beliefs in the face of "rational" evidence to the contrary.
Thought its cowritten, The Rise and Fall of DODO is more explicit in its discussions of rationality (aka scientism) and the dangers of flying too close to the sun.
"Page number" is a slight exaggeration. I would put Anathem and Seveneves more or less fully in the "ringing endorsement" camp and REAMDE and Cryptonomicon much more into the "wow this guy kinda despises rationalist types doesn't he" camp, with Fall and Snow Crash too culture-war-ish to rise to the level of either.
I'll bite. How is Snow Crash culture warrish? It seems to dunk on basically everyone.
It probably does... I admit I don't remember it that well, and perhaps I'm reading later stuff (eg Fall) back into it.
My take on Stephenson is that he writes based on political views formed in the 90's, when STEM types were solidly aligned with mainstream Democrats against the religious Right (cf. creationism teaching controversies, internet scares, etc.). In Snow Crash he embraces the "conservative = religious = suburban = racist" line of that era in an uncritical, even cartoonish way. Since then the alignment has gotten a lot less cozy and Stephenson has shown that his loyalties lie exclusively with the Blue Tribe, not the Gray Tribe, when the chips are down. His vision in Fall of the good guys as the "reality-based community" was cringey and ten years out of date on the day it was published.
I don't think that's quite historically accurate. In the first place, if "STEM" includes more than "computer programmers" then it's certainly inaccurate, a large fraction of physical scientists and engineers are not especially in sympathy with the mainstream Democratic Party, even today.
And even in the case of programmers, the situation is considerably more nuanced, as during the 70s and 80s the most avant-garde programmers tended to be libertarian freaks like Stallman, and the remainder, building code for mainframes to process 10 million credit-card transactions per second, were solid middle-class suburban worker bees, as likely to be Republican or independent. The 90s were arguably a period of transition, as a new generation of Silicon Valley professional Internet-focused coders was arising, breaking off from both the GNU weirdos and the suit-and-tie IBM VM/SP programmers.
I actually think this take is pretty wrong. There’s a recurring theme in Stephenson’s work about the resilience and benefits of red tribers, although it’s definitely declined over the years. Compare the diggers from Seveneves (essentially survivalists who institute a theocracy underground to ride out the apocalypse) which is both less favorable than his depiction of the extended Shaftoe clan in earlier works, and more favorable than the Ameristan excerpt from Fall, which I will grant is pretty bleak. I get the impression that he has a lot of family in flyover country but has been gradually souring on the group as it gets crazier
Snow Crash was fun. I made it half-way through Cryptonomicon before shelving it. I made it a few dozen pages into Anathem before abandoning it. For someone so lauded, his writing was less than enjoyable.
I think a valid criticism of Stephenson's writing is that the books can take a very long time to get into (especially Anathem - its my favorite despite that). His climaxes can be pretty meh too. You have to really want to just spend time in the world and with the characters as well as enjoying the plot otherwise the books are a chore.
Good SF of a certain kind gets mentioned a lot as fitting these criteria. My personal picks would include Foundation (Asimov), Speaker for the Dead (Card), and Anathem (Stephenson).
Arguable for these or any candidate how much of the gateway-ness is treatment effect vs selection effect, though.
Hitchhiker’s guide to the galaxy- avant le jour
Harry Potter and the Methods of Rationality.
HPMOR is frankly kind of cringe, but "three worlds collide" is quite good IMO
might depend on when you first read it, but it was definitely my gateway into rationalism. I wouldn't claim it's the best, as I haven't surveyed the field, but I think it definitely belongs on the list
EDIT:
On a little more consideration, it probably depends on whether you parsed the opening post more as asking about " 'The best piece of art' that has the qualification of being a gateway to rationalism" or " 'the best gateway to rationalism' that is a piece of art"
I think that HPMOR very naturally segues into the early Eleizer's sequences and also makes it pretty clear what rationalism is about, so it's good for the second interpretation. On the first, I enjoyed it and wouldn't call it cringe, but from an artistic perspective, it's probably lacking
Artwork by DALLE2
That photo of Elizier Zudkowsky
Wut photo? Curious to see.
https://www.google.com/search?q=eliezer+yudkowsky&rlz=1CDGOYI_enGB900GB900&hl=en-GB&tbm=isch&prmd=niv&sxsrf=APq-WBuJWLOxdgtdkiV0iSQb_PgXRA3nJA:1650225488249&source=lnms&sa=X&ved=0ahUKEwjq35qA8Zv3AhVLY8AKHew_A9IQ_AUIFigC&biw=375&bih=640&dpr=3#imgrc=mVwxH1YSGIDn0M
Amusingly, the dust specks thing is the first post by Yudkowsky that I chanced upon, and I hated it so much that for years I dismissed him as having anything worthwhile to say. But because the LW jargon was everywhere and Scott always praised him so much I finally read the damn sequences. Somehow it's still the post I dislike the most, and consider the whole "meta"ethics part the weakest.
What's the dust specks thing? Can you give a link or a lead?
https://www.lesswrong.com/posts/3wYTFWY3LKQCnAptN/torture-vs-dust-specks
Definitely not a gateway. Maybe if you're already inside the compound and you look over your shoulder, the gate you came through looks something like those photos. To those of us still roaming the terrain outside the walls he looks like a bloke with a beard.
I was joking…
Scott put it in an old SSC post that’s why I thought of it
OK, I get it. I should have realized that the person who put up a link to Fuck Your Miracle Year could not be a Yudkowsky disciple.
Medici (TV series), but the association isn’t plain and obvious.
The Martian
Interesting, that hadn't occurred to me. It's certainly extremely *nerdy* for a successful artistic work (the "Glorfindel" scene is my favorite example of the essence of nerdiness) but not in ways that are distinctly rationalist. It focuses on solving practical / technical problems in a nerdy way while taking the parameters and goals for granted. Rationalism, to me, implies more of a focus on epistemic and societal problems.
On a trip to Vermont a guy who makes maple syrup told me that, counterintuitively, sap flows through a tree faster when it's cold out. I didn't have a chance to ask him why. Any theories?
Perhaps colder temperature causes the smaller “blood vessels” in the tree to constrict, diverting the liquid to the larger vessels.
It's not that sap flows faster when it's cold out. It's that sap flows faster when it's cold at night and then warm during the day. Basically, the night subzero temperatures causes the sap to freeze and then be squeezed by the tree (which shrink slightly due to low temperatures). This causes the sap to become much more dense. Then when it gets warm the tree expands out and the sap melts and expands. This causes faster flow. If it's too cold (or doesn't get warm enough) then the sap can just remain frozen all day. If it's too warm then the sap doesn't get condensed and will still flow but slowly.
I've worked tapping trees and on some other agricultural stuff so AMA. I guess I should disclaim this isn't some scientific inquiry but what the tappers told me.
My guess is that the tree needs more sap flow in the cold for some reason.
As brilliant as zdk's reply is, the explanation widely given here in REAL maple syrup territory is that nights below zero and days above give the best flow of maple sap. Too cold, doesn't flow, too warm straight through and the flow stops soon.
This makes intuitive sense, but the actual mechanism is still unclear to me.
Thanks for a serious reply 😃
It is counter intuitive! If it's true, maybe heat causes/increases the rate of a polymerisation reaction which makes the syrup more viscous.
Or perhaps the heat causes expansion of parts of the tree that press on the syrup tubes and make them smaller.
Maybe the heat makes the surface stickier.
I'll look for ideas more grounded in evidence.
So according to https://www.massmaple.org/about-maple-syrup/frequently-asked-questions :
"4. What are the right weather conditions for sap flow?
The tree’s sap flow mechanisms depend on temperatures which alternate back and forth past the freezing point (32 degrees F.). The best sap flows come when nighttime temperatures are in the low 20s and daytime temperatures are in the 40s. The longer it stays below freezing at night, the longer the sap will run during the warm day to follow. If the weather gets too cold and stays cold, sap flow will stop. If the weather gets too warm and stays warm, sap flow will stop. The cold weather at night allows the tree to cool down and absorb moisture from the ground via the roots. During the day, the tree warms up, the tree’s internal pressure builds up, and the sap will run from a taphole or even a broken twig or branch. For good sap production, maple producers must have the alternating warm/cold temperatures. This is why its so impossible to predict the outcome of the maple crop from year to year."
Came here to say this; it's the *difference* that's important. Incidentally, this is part of the reason why maple plantations tend to be a bit inland rather than on the coast (source: grew up east coast Canada, had many fond tir à l'érable experiences)
Cold depletes the phlogiston in the atmosphere
Yep - that’s what a tapper told me. And an excess of choler
Can you ELIphysics_major_5_years_ago this part? Or perhaps provide pointers to further reading on that level?
"In this case the bekenstein bound is a bound on entanglement, not on information."
Well, this was my understanding of this paper
https://arxiv.org/abs/0804.2182
But seeing the comment below, maybe I am just very very confused.
The bound on entanglement is, said differently, a bound on the size of the Hilbert space describing the black hole, and it is precisely this that is meant by “the information contained in a black hole” in the context of the Bekenstein bound. The Bekenstein bound is a limit on how many qubits you can fit into a volume, which in turn is a limit on how many classical bits you can fit in that volume.
Seems kind of in-line with my comment in the links thread that a lot of political divides these days are less left vs right than they are insiders vs outsiders.
PredictIt has Macron/LePen at 91/8. As I understand it, Le Pen was leaning into culture war on immigration, but retaining a Statist outlook economically and per safety net. The 'far right' designation pertained to her social views. By American standards, that's strange and interesting.
Immigration is sort of a hard switch. If a politician runs as a Radian caricature they will still likely be called center right, whereas if a candidate does a copy-paste of a socialist party platform with immigration restriction tacked on and it becomes far right.
I guess I'm a bit surprised that you're surprised at this. The Netherlands apparently already has policies restricting support of Nazism, so it shouldn't be too shocking that they'd expand this to other viewpoints that the vast majority of people find abhorrent.
How is it that every time someone pops up about "hey, this party/this person is being persecuted just for expressing a simple opinion", it turns out to be kiddie-diddlers?
Every time.
Ev. Er. Y. Time.
Pretty sure you don't have evidence that most of these people have done the thing you claim they did.
Aren't people also persecuted for arguing the holocaust never happened? This seemed like a much bigger thing a few years back, before YouTube banned those videos...
Burdens of a community that openly values speech. One of my favorite quotes from the old Slate-Star: "if you’re against witch-hunts, and you promise to found your own little utopian community where witch-hunts will never happen, your new society will end up consisting of approximately three principled civil libertarians and seven zillion witches."
It's worth it on balance, but one of the downsides of trying hard to be a tolerant space is that a subset of your population will always be people who are attracted to the space because they have found their ideas (or themselves) to be intolerable elsewhere.
Re: Point one, I'm not sure how laws restricting free speech work in Netherlands, but my guess is that they're implemented through a democratic republic, meaning that the representatives of the majority of people all agreed that supporting pedophilia was abhorrent, and should be punishable by jail (as they did for supporting Nazism). So, an actual majority was needed to implement it, not a single dictator. Freedoms and democracy are not the same thing.
For your second point, that's a big (and incorrect) leap from what I said. Personally, I'm of the opinion that people should largely be able to say what they want; I'm opposed to both Nazis and pedophile-supporters from being jailed for stating their opinions. But, I live in the US, not the Netherlands, and so my input on their system has limited importance. Secondly, a government doing something, and you doing something, are very different things. For example, (using your terminology) I think that the government should be able to kidnap and imprison murderers; that doesn't mean that I think you should be able to.
"What if more people really wanted the PNVD to win?"
Do you legitimately believe the majority of people in the Netherlands support sexual relationships between a 40-year old and a 12-year old? "What if" is all very nice, but doesn't have a lot to do with "it is" in the end.
I think that most democratically governments legally can appoint themselves oligarchs for life. I don’t know much about the Netherlands constitution, but in the US, if the federal and state legislatures coordinated well enough, they could appoint themselves oligarchs for life tomorrow. That they can’t is a pleasant fiction we all maintain
Okay, you know what immediately dinged my bell?
"Their "crime" was the alledged continuation of banned Association Marthijn by publishing documents which discuss age of consent related issues."
I've been here before. I'm old enough to be aware of Peter Tatchell and how the gay rights movement in Britain got entangled with PIE (Paedophile Information Exchange), the arguments that the age of consent should be reduced to fourteen, and other things that became embarrassing, to say the least, as time went on and those involved became respectable:
https://en.wikipedia.org/wiki/Peter_Tatchell#Paedophile_Information_Exchange
What does Wikipedia say about this association?
"Vereniging Martijn (Dutch pronunciation: [vəˈreːnəɣɪŋ mɑrˈtɛin]; "Martijn Association"; stylized as MARTIJN) was a Dutch association that advocated the societal acceptance of pedophilia and legalization of sexual relationships between adults and children."
Now maybe it's wrong and they want the age of consent to be fourteen or something. But you know, whenever I see someone talking about "age of consent", they often turn out to be not just "science proves" and "not criminals", they want to fuck eight year olds.
In other words, you are going to have to do a *lot* of work to convince me that this is just "let's talk about the law" and not 'stealth movement like so many times before to make it okay to fuck eight year olds'.
Is there another, more digestable (and shareable) source that confirms this?
Yep, having read that, the guy wants to fuck eight year olds:
"I have been sentenced to six months in prison. I quoted murdered politician Pim Fortuyn, who said: Why not make sex between children and grow-ups legal? I wrote that I agreed with Fortuyn on this, because when you look at all the scientific studies concerning harm in pedosexual relationships, the harm seems to disappear when the relationship was wanted by the minor. This is not my OPINION, it is the conclusion of scientific papers, approved for publication in scientific magazines."
Because you know who I've read who all state that "the harm isn't the sex, the harm is the irrational response of adults to grownups having sex with kids, and science says I'm right"? Guys who want it to be legal to fuck eight year olds.
Otherwise, stop shilling for groups that put pictures of four year olds on their website:
"The organization was in the news in October 2007 when it was learned that photographs of Princess Catharina-Amalia (then aged 3½), daughter of Willem-Alexander, Prince of Orange (now King) and his wife Princess Maxima (now Queen), were on display on the website's forum. The Prince went to court to request a €50,000 fine and the removal of the photos from the website. The court agreed that the photos must be removed, and imposed a fine of €5,000 to be paid every time photos of children of the royal family are placed on the site again. The organisation had to pay €1,235 in costs"
We all know it will never be legal to f*** 8 year olds, so what does it matter?
If some scumbag says women who dress provacatively are asking to be raped "and this scientific paper agrees", would you support a prison sentence for saying that?
Sounds like yes. But I guess I'm too Americanized and think that a banhammer is enough.
"We all know it will never be legal to f*** 8 year olds,"
We all know it will never be illegal to open a public occasion with a prayer invoking Jesus Christ.
We all know divorce will never be legal widespread, no matter what a few countries/states of the US may do.
We all know sodomy will never be legal.
We all know contraception will never be legal.
We all know abortion will never be legal.
We all know gay marriage will never be legal.
We all know trans people in bathrooms not their biological sex will never be legal.
We all know... but I've seen it happen in my lifetime.
You really think that the push for "children are sexual beings with their own agency" will *never* be legal? Since age of adulthood has come down, age of consent has come down, acceptance of "14 year olds are having sex even if it is illegal and even if it is statutory rape", minors can get abortions without the consent or even knowledge of their parents with the help of a complicit judge, proposals to drop the voting age to 16, 'we have to teach three year olds about gender and sexualities, here's your Blues Clues video' and increasing sexualisation in society, that in ten or twenty or fifty years time, "sure, eight year olds can consent to sex, kids today are mature earlier than in the past" will *never* happen?
Okay, so your concern is a slippery slope. But instead of saying "if we let gays marry, next thing you know people will be allowed to f*** 8-year-olds" like some people used to suggest, you're saying "if an adult can get away with *advocating* allowing 33-year-olds to have sex with 13-year-olds, next thing you know they will be allowed to actually f*** 8-year-olds".
Which, I think, means that you think that the arguments in favor of allowing adults to sex 8-year-olds may be persuasive to a majority of the population, such that this change could actually pass into law. And so, you think, it's important for those arguments not to be heard in the first place.
Well, okay, but my sense is that your opinion is extremely common, and that a supermajority are against letting 33-year-olds sex 13-year-olds, so the thing you worry about is very far from coming to pass. (edit: this seems especially so if the people advocating for this change sound like the people I saw on the link Andaro originally posted and then deleted. Those guys, well, they sound like people who want to f*** eight-year-olds)
FWIW, age of consent in a large part of the EU is 15-16 and the sky hasn't fallen yet.
It helps that people will shun you if you have a sexual relationship with a 15 year old girl as a 40 year old man, so social pressure solves the most obvious problems. Meanwhile teens can fuck each other in peace.
The historical pattern is the opposite. Under Rabbinic law, a woman could get married without her parents' consent at twelve and a half plus evidence of puberty (a man at thirteen and a half, I believe). In 18th century England, the age of consent for marriage was twelve for women, with ten the age at which intercourse counted as rape even with consent. In most of the US in the 19th century age of consent was ten or twelve. The current pattern, sixteen to eighteen in, I think, all of the US — I don't know about Ireland or the Netherlands — is unusually high by historical standards.
Sure, it's not outside the bounds of possibility, though it does go against the flow of the primary basis of the morality that produced most of those other changes over the last 50 years or so.
Societies have allowed it before so it's possible.
I wouldn't bet a penny on it in the next 30 years but it's possible.
If you just view all of those other things as moral decay then hypothetical 8-year old fucking is just another type of moral decay that all those others will obviously lead to.
If you view most of those things as moving further and further away from the days when the local bishop could diddle 8 year olds without anyone raising a finger to stop them then it seems less like an obvious direction for society.
I get the impression from online sentiment... as a lot of internet-generation women hit their 40's and start to really really resent that men their age are more interested in 20 year olds there's going to be more and more of a push to treat 19/20/21 year olds as pseudo-children.
I just really detest ratchets that are only ever allowed go in one direction. Do you think society would be better if the age of consent was raised from 16 to 17 or 17 to 18? more power to you! campaign away. But if you brand everyone who goes "eh, it seemed fine at 17" as a pedo then you're creating a ratchet effect that only ever allows one direction of movement which isn't a particularly good thing.
When you've got a social issue like that people should be able to push the boundaries slightly in either direction without being branded as evil.
Free speech is normally interpreted to mostly mean free speech about political issues, on the grounds that in a democracy the most important decision for the people to discuss is who should govern, and we need strong protections for people to be able to discuss that question openly.
But it rarely means speech qua speech no matter what. If I start a Website devoted to normalizing cannibalism, or giving people explicit instructions on how to kill themselves -- maybe in exchange for their livestreaming the event -- or I start openly advocating for throwing all the Jews in the gas chamber, and trying to organize people do to planning, maybe get a private organization to build a trial death camp on private ground -- I am going to get shut down by the majority tout suite, and you'd have to be pretty naive to be surprised or wonder why.
As Justice Jackson famously quipped, "the Constitution is not a suicide pact," and we can say the same thing about constitutional rights, or natural rights one might want to adduce in a country lacking constitutional rights per se: they are not unlimited and absolute, and there *is* speech we rule out of the public forum because the cost of bestowing even the lightest sheen of acceptance or approval on a river of raw psychopathic sewage is not worth whatever benefit may otherwise accrue.
Joseph Bronski, whom I vehemently disagree with, at least has an overarching argument about age of brain development and age of maturity (including age of consent).
People who only bang on about age of consent in relation to sex, so that they may lower it, so that adults having sex with children is not illegal?
Yeah, they wanna fuck eight year olds. Defend them if you like, but if you lie down with dogs, you are going to get up with fleas.
The guy in question didn't go to prison for quoting science, he went to prison for breaking laws about "yeah I wanna fuck kids".
I use a rough "would I recommend this review to others"-scale for rating: 10 = I'd recommend it to a friend; 8 = I really liked it, but the typical ACX reader may not, 7 = not my topic, but I feel the typical ACX reader would like it; below 5 = waste of my and your time.
From what you wrote, this sounds like a below 5 category. And I don't mean this harsh in any way, I read all of last years runners and almost never gave <5, most reviews are really really good and I apreciate all the time that went into them. But if you only can recommend 10% of all articles, which ones would it be? This is the question Scott has to solve with promoting ones to finalists.
^This is great—super clear. Thanks!
Personally, I only voted for the ones I liked (or disliked because I felt they were not including obvious facts).
Others I read but didn't have a strong yes or no opinion on, I didn't vote for. I think if you thought the review was poor and that's why you didn't finish it, then it's fair to vote on it. Low votes for "This is not well done" are more useful data than "Nobody voted for this, no idea why".
I stopped reading one because I didn't like it. I gave it a below average review. I feel no guilt. There are some great reviews!
In my case, the odds are zero, because the word I always guess first is not in the list of possible Wordle answers. If you start with a word that _is_ in the list then you have a 1/2300ish chance of guessing correctly on the first turn. (In some sense; the actual order of appearance is fixed. But if you don't know that order, but _do_ somehow know that your word is on the list, and haven't been paying attention to what past words have been and whether or not any of them is your word, then the correct probability to assign is about 1/2300.
"Collective consciousness" in the sense of, say, telepathy seems vanishingly unlikely to me. But what might happen is e.g. that some particular topic is in the news and more people than average guess related words, and sometimes one of those words will happen to be Wordle's word of the day. If what you're interested in is Funky Telepathy Stuff, I think it would be very difficult to disentangle from that sort of exogenous correlation.
I saw my niece get it in two yesterday. She only had an a from the first one and was just guessing the second one, she said. 1-2 are flukes. 3 can be worked out.
With the classic Wordle, there were 2315 possible answers. I think the NYT version trimmed the wordlists a bit.
There are 13000 possible guesses, and the original creator chose 2315 of them to be the target words for the following 2315 days. The criterion for being one of those was apparently just whichever ones were "familiar to [the creator's] partner", creating a strong bias for the target words to be fairly common. cite: https://arstechnica.com/gaming/2022/02/heres-how-the-new-york-times-changed-wordle/
So while the odds of a correct first guess might seem extremely low (one in 13000! wow!) those odds are raised significantly because common words, which will often be the guesses you'd think of first, have a higher chance to be correct. Out of 4 people in my family who Wordle, there have been 3 correct first guesses in 3 months, and while that's higher than baseline, it's not all too crazy given this bias towards common words.
You say "art", Justin E. Smith & Scott Sumer say "middlebrow": https://justinehsmith.substack.com/p/what-are-the-humanities?s=r
https://www.themoneyillusion.com/understanding-middlebrow/
Your characterization of serious films in the mid-to-late 90s through mid-2000s doesn't resonate with me at all. Tons of great work by Spielberg, Nolan, Zemeckis, Hallstrom, Ang Lee, Fincher, Aronofsky, Jackson, etc.
I'd say A.I., Minority Report, Catch Me if You Can, The Terminal, and Munich are all good and unique films (not to mention Lincoln and West Side Story in later years). Castaway wouldn't be a footnote for Zemeckis in my book. - again, a great and unique film.
When you say fiction, do you mean books? You can add the word "literary" to pretty much anything to imply that it has ambitions above pure genre thrills. And the word "classic" tells you that the author did it successfully! So Ian McEwan's Sweet Tooth is a literary spy thriller, but Le Carre's The Spy Who Came in from the Cold is a classic spy thriller (in my opinion). The Handmaid's Tale is literary sci-fi, and if I'm any judge, classic sci-fi.
>That division was very strong in say, american cinema from the mid-to-late 90s to the mid-to-late 2000s
Tarantino and the Coens "did both" in that era. Added: Scorsese also made plenty of films in that era. Terry Gilliam made some good sci-fi movies in the mid 90s such as 12 Monkeys.
Edit: And thinking over all the names above, aren't most all of the directors mentioned here considered to be "auteurs"? The word applies to the director not the work, but it seems to be gesturing in the direction of what you mean, if we say "a movie by an auteur". Of course, there are plenty of "auteurs" who only make art-movies, although don't we usually reserve that term for directors who make somewhat popular films? I wouldn't call Cassavetes an auteur, I'd call him an "experimental film maker". I suppose Wes and P.T. Anderson are auteurs who don't make hugely popular films, though I wouldn't consider their movies to be arthouse either.
The story about how the US disparages ebonics is so sad. It seems that the disregard for this language variant has caused education achievement gap for black people.
I don't think it's a moral panic yet, though it's certainly something that could turn into one. I'd watch for the widespread citing of increasingly non-central and/or trivial examples as a sign-- less and less talk about people losing their jobs, having speaking engagements canceled, etc., and more and more talk about mere criticism as "canceling". You can find a certain amount of this sort of thing even now, but I think you still pretty much have to go nutpicking in order to do so, thus far.
I mean, it clearly is. What else could it be?
Otherwise rational people treat it like an existential threat to civil society, even more dangerous than when a mob broke into the capitol building with the intent of murdering government officials.
It's clearly not a good thing to exist, but it's not good in the way political correctness is not good.
Define "moral panic".
It could be a weary and righteous reaction to yet another bout of McCarthyism.
>Otherwise rational people treat it like an existential threat to civil society, even more dangerous than when a mob broke into the capitol building with the intent of murdering government officials.
This drive-by reveales an allegiance, and makes your whole assessment a rather standard dismissal of your outgroup's fears (or 'hysteria', as you probably like to think of it) rather than anything more interesting
I just don't think it's worth addressing seriously.
The amount of sand I've dug out of this hole is depressing, in retrospect.
Is there evidence that the political make up of a law school's faculty has an appreciable impact on the political views of the students? Or is there any evidence that the legal profession as a whole has become more progressive over time relative to the general population?
When someone provides an example of something they do not have to provide *every* example that could possibly be relevant.
A big difference would be that BLM protests were against a thing that actually happens (extra judicial killing of citizens at the hands of the police) vs a made up fairy tale for boomer LARPers.
Nah's comment was specifically about how people can believe something like cancel culture can compel people to violence.
BLM was about how government murder can compel people to violence. Apples vs oranges.
A big difference would be that BLM protestors ACTUALLY KILLED PEOPLE while J6 protestors did not.
I'm not too happy about much of that stuff (without remembering/agreeing with the exact details here), and I think there are plenty of people who are against both January 6 and, various things Antifa or whoever did (though I think January 6 was worse).
But does anyone remember the thing before *that* - peaceful protests in the aftermath of Floyd's murder met with huge amounts of police violence? Tear gas, rubber bullets, all that sort of thing. It seems like it's been memory-holed by lots of people. People remember Lafayette Square specifically, but only because of the involvement of the White House; those sorts of tactics happened repeatedly all over the country, and I'm guessing in the vast majority of cases the cops who used excessive force faced 0 consequences. One case where there might have been consequences - the police in Buffalo shoving a ~70 year old protester - they just got off scot-free a few days ago in arbitration.
On Lafayette square here are some links to videos. You can clearly see nonviolent protesters being attacked by police but I don't see a mob attack on the White House, you're free to post any video. IIRC there were some violent protesters but no, like organized attack on the White House or anything.
https://twitter.com/AlliemalCNN/status/1267588286840803328
https://www.youtube.com/watch?v=x_ZwKL3Qhds
https://www.youtube.com/watch?v=UrMoqSPZym0 (a longer video of DC that day, for some examples of violence see 28 minutes, 31, 35)
https://www.vox.com/policy-and-politics/2020/6/1/21277530/trump-speech-police-violence-dc-tear-gas (just looking at the linked video at the top, at about 25 seconds you can very clearly see the police advancing on an unarmed, nonviolent crowd including with tear gas)
As for more general police violence here are some examples:
https://v.redd.it/y9vv08etog551
https://v.redd.it/s6f1me9nxl451
https://v.redd.it/8c07fmljgm251
https://v.redd.it/kxmoo4o8fb251
https://twitter.com/MattOrtega/status/1266550829378568192
https://twitter.com/JasonLemon/status/1266529475757510656 (same incident, from a Newsweek reporter who says that the cop called the woman a "stupid fucking bitch")
https://www.youtube.com/watch?v=D8c8MoL1eqQ (the incident I mentioned in my earlier comment)
https://www.youtube.com/watch?v=sMYF-Y_r52Y (not an actual incident, a cop bragging about hitting protesters with his car not realizing he's being recorded ... he was suspended for 10 days)
https://www.nytimes.com/interactive/2020/07/14/nyregion/nypd-george-floyd-protests.html
https://twitter.com/rachelolding/status/1267636201915383808
No, its not. The real moral panic is "racism".
This is America: we're perfectly capable of keeping more than one moral panic going at the same time.
I mean, that's 60 years ago - the people in question are mostly dead. Your argument rests *heavily* on the "transmitted to children and grand-children" part, and that is far from proven
It's hardly an accident, and nobody thinks it is. The legislators in question think that the Democrats are encouraging voter fraud among illegal immigrants, and *that* is the real reason why they (the Democrats) object to stuff like "you gotta show an ID to vote so we know you're a citizen."
So, yes, voter ID laws may disproportionately hit poor brown neighborhoods, these people would agree -- because that's where the illegal immigrant voter fraud lies. It's like saying laws against crack trafficking will disproportionately affect barrios with strong Mexican gang presence. To which observation most people will say well...yeah.
Of course I understand that the opposing point of view is that voter fraud is all made up, doesn't happen, and if it does happen it's too small to matter. I'm in fact pretty sympathetic to that argument, but people who mindlessly oppose even things like voter ID tend to shoot themselves in the foot over it. I mean, if you're so convinced voter fraud is a nonissue, why oppose voter ID? Why not say, sure, that's fine, no problem -- but we strongly oppose, say, closing the polls at 5pm because then working stiffs have a hard time getting to vote? Getting passionate about voter ID laws -- or *any* kind of vote-security rules -- makes the independent wonder whether there might actually *be* something the partisan Republican attack -- else why protest so much?
There have been two major steep drops in Democratic party identification over the past 80 years[1]:
1. In the middle of the Johnson Administration. That's probably the era to which you refer[2], but it doesn't seem super likely those who left then were the Dixiecrats and those who were left behind had much in common with today's college-educated wealthy New York City/San Francisco core demographic. That really doesn't correspond well to the typical Southern or Midwestern Democrat of circa 1968. So I'd say it's equally likely the people who left the Democratic Party in the mid 60s were suburban voters who were disgusted with the Dixiecrat leanings of the core Democrait Party, and that they left to become independents (there was a big surge in independents at the same time).
2. During the early years of the Reagan Administration. I was there for that one, and knew a number of people who made the switch -- so-called "Reagan Democrats" -- and their reasons had zip to do with race, which wasn't a big issue in the 80s anyway -- and more to do with a sense of disgust over the Carter era feelings of glum diminishment -- with the idea that 12% inflation and crappy economic growth was here forever, and we were just inevitably headed for the status of Great Britain in the Callaghan Era (barely First World).
--------------------
[1] https://www.pewresearch.org/politics/interactives/party-id-trend/
[2] It's a little complicated by the fact that Vietnam, a war started and greatly expanded by Democratic Administrations, had become so deeply unpopular that Johnson thought he would just drag the party down if he ran in 1968.
The problem with that analysis is the lack of objective evidence. The best-case argument you can make is that they went from publicly beating black people to advocating for charter schools because of racism. And if the last widespread vestige of racism left in the nation is support for charter schools, I think racism is pretty much dead.
? where did charter schools come from? this seems like you are making a huge leap in what Machine Interface may or may not think.
Jesse Singal, who is not a conservative, thinks that Hobbes is full of it:
https://threadreaderapp.com/thread/1452844926052544514.html
"Political decisions which govern people's lives" refers to a wide variety of subjects, only one of which can be the most important. Therefore we have to ignore all the others, for the same reason we have to ignore cancel culture.
This is generally my view. I really can't respect anyone spending significant amounts of time screaming about "Cancel Culture". They should log off and spend some time in the real world.
Where they will find, a few years down the road, that the rules which severely impact and circumscribe their lives have been made without their input, and if they want to do anything about it their only recourse is to vote for Donald Trump or someone like him.
The internet is part of the real world. The extent to which "cancel culture" presently impacts the non-internet world is debatable, but it's kind of a big deal on the internet and people are going to debate it on the internet. If that annoys you, take your own advice.
I do recognize that the debates that happen on twitter/internet/Facebook etc do impact real world policy sometimes. Though I suspect that twitter is more of a lagging indicator than the cause of those policy changes.
Moves to support candidates who represent themselves as outside the mainstream (Trump, Bernie, what have you) aren't the result of twitter. Ron Paul became popular without twitter and there has always been fringe candidates and political movements in this country. That they are more popular now is because of how terrible the mainstream political parties have been the past two decades (or more). Twitter has just been a venue for these voters to express themselves.
I also think there is a big difference between large social networks like twitter and Facebook and a site like ACX or even a sub reddit (the main page of reddit is more a social network). The goal of social networks is to create "vitality" or news worthiness. This pushes them to amplify the worst voices and to almost create attitudes and opinions that people didn't even know they had.
Cal Newport recently called social networks the "Roman Colosseum" which I think is a good analogy. Most users log on to watch "influencers" battle it out (though because no one dies the battle never ends and both sides can say they won).
The *war* never ends. The battles end every two years, on the first Tuesday after the first Monday of November. At least in the Colosseum, a "thumbs up" majority only killed one unfortunate SOB. Now we get a higher order of entertainment.
Our Side Does It: We're fighting hate speech!
Their Side Does It: It's a moral panic!
You can swap the sides depending on what the topic is, and it's the same thing.
W. B. YEATS
The Great Day
Hurrah for revolution and more cannon-shot!
A beggar upon horseback lashes a beggar on foot.
Hurrah for revolution and cannon come again!
The beggars have changed places, but the lash goes on.
What you call cancel culture I call terrible management and labour practices. I think that is a great example of why using the term "cancel culture" is stupid. When you call that cancel culture you lump it in with idiot politicians complaining about their tweets not getting read. It minimizes the actual behavior that SDG&E engaged in.
Also, in this case Cafferty, has a clear form of recourse: to sue for wrongful termination. Which it appears he has (or it maybe for defamation, its unclear in the reporting).
However, Cafferty also sued three people on twitter for defamation who tweeted at SDG&E about him. Is that not also cancel culture? They were expressing their opinion that Cafferty was showing a white supremacies sign. This speech is clearly covered by the First Amendment.
I'm not aware of legal bans, but Amazon has delisted books from its marketplace (including second-hand sales) on a number of occasions, which is a de facto banning for rarer titles.
A private business choosing not to carry a product any longer isnt in the same ball park as a state government creating content based legislation regarding which books can be carried in public school libraries.
I wouldn't be surprised if Amazon is now a larger and harder-to-bypass source of books to more readers than are public school libraries.
Yeah but that could just be an illustration of Conquest's First Law of Politics.
Dr. Seuss Enterprises withdrew the book from publication, voluntarily and on their own initiative. It was not banned. Seuss himself updated and changed the language of the book back in the 70s as he thought it was offensive.
You can buy the book right here: https://www.amazon.com/Think-That-Saw-Mulberry-Street/dp/0394844947
$22 for a used copy. So to "let you know how it goes": quite easily.
In this case the shop keeper put a sign out saying they weren't carrying a product any more and people rushed to buy it all up. Sales of the book soared after they announced publication would cease. Not sure who the Mob was in this case that benefitted.
It should be interesting to watch the response to Ross Douthat’s column in today’s (April 17) NYT. A politically moderate public homosexual (Jonathan Rauch is the other), Douthat concludes that would-be progressives “have been running an experiment on trans-identity youth without good or certain evidence,” declaring the silence of liberals who agree with him “will eventually become your regret.”
That has to be the boldest challenge to gender identity ideology I’ve seen in decades, clearly echoing the “silence is complicity” accusation of recent liberal protest demonstrations.
Will they again eat their own?
>A politically moderate public homosexual
Wait. When did Douthat come out?
No, the only converting I remember is him switching from Catholic to Orthodox triggered by the sex abuse scandals, but not only those.
I think you must be thinking of someone else.
Maybe his nonconformism as a libertarian threw me off.
I thought he had declared it at one point. If not, I certainly didn't mean the association as an insult. Most of his references to gender dogma have been highly orthodox, so his principled stance surprised me. But he's about as 'conservative' as David Brooks -- the left's pathetic offer of 'diversity' in their propaganda empire.
Ross Douthat is married to fellow writer Abigail Tucker, and they have, I think, four kids together. Like someone mentioned below, I think you might be confusing him with Andrew Sullivan.
He hasn't, and I suspect he's being confused with some other writer here - Andrew Sullivan maybe?
Also, Douthat has been the NYT's token conservative for years, and left-liberals do *not* consider him "one of their own" - quite the opposite.
Yeah, the current Douthat column in the NYT is not about the topic claimed above, so Cosimo must have meant someone else.
Actually, he did recently write about that topic (not in today's column, but last week), including the "running an experiment" quote: https://www.nytimes.com/2022/04/13/opinion/transgender-culture-war.html
It's the "public homosexual" part that seems to be confused.
Ross Douthat is married to a woman, isn't he?
Nope, the foundation remains intact. There's just an election coming up, so it's time to reveal that you've just been humoring the radicals, because free speech and diversity, but you really do have moderate roots. This is why the NYT is running chin-stroking thinkpieces on the shocking rise in violent crime over the past two years, and suggesting it might not be a totally bad idea for citizens to consider the police as some kind of partner as opposed to just the pigs, and it's even possible adding some money to their budgets isn't necessarily all bad.
And the explosions and flames in the background, at the peaceful protest, is just recycling
Especially in the NYT. I can't imagine heads won't roll in the Editorial Dept.
I mean, here of all places you'll find people biting that bullet - many of the old enough commentariat were very Atheist (and often actively anti-theist, including my teenage self) back when that was the big internet debate in the '00s
Yes. It was wrong when the Evangelicals were doing that in the age of the "Moral Majority", and it's wrong that some Republicans are doing that now. I don't think you'll find much disagreement on that here.
It's *just as wrong* for liberals to be doing the same thing, and worse when they piously defend it as "this time it's different, because reasons". You, or at least the people you apologize for, are *worse* than the Moral Majority ever were. They never really had the experience of being on the wrong end of the banhammer to caution them against overzealous use of the banhammer.
And you know what the liberal side was saying back then? That all this censorship was bad and repressive and unwarranted. That there was never any reason for banning something. That merely being offended wasn't good enough.
Then they got into the saddle and suddenly it wasn't censorship anymore, it was "if it's not the government doing it, you're not being censored" like the xkcd comic (which never seems to have considered that by the same argument it made, it would be perfectly okay for a website to kick someone off for expressing a positive attitude to Pride Month):
https://xkcd.com/1357/
It's not censorship, it's fighting hate speech, which is literal violence, which is literally murdering trans kids etc.
Yeah, obviously the Evangelical right is fucking insane to about exactly the same degree that the left. I am surprised you think you're scoring points here. The comparison is extremely unflattering for the left.
It's also a matter of conveniently forgetting that today's political / cultural alignment is not necessarily the same as it was in the past. The Westboro Baptist Church had a goal of aggravating both sides to generate as much hate (and lawsuit revenue) as possible; Fred Phelps himself was a civil rights advocate and a Democratic political candidate. Tipper Gore was the visible crusader for a lot of moral regulation of entertainment, such that the warning labels were called 'Tipper Stickers'. Even going back to Prohibition, both parties split in roughly the same proportion ratifying the 18th amendment.
It's convenient and fraudulent to forget that the political left has always included its share of moral busybodies; what's new is the shift from targeting industries to targeting individuals.