I'm trapped in a room with a small child. The child asks me why the lights work. I explain that the light switch turns them off and on. The child asks "why?" and I explain about the electrical circuits in the walls. The child asks "what are circuits?" and I say they're loops of wire. The child asks "what is electricity?" and I explain about electrons.
The child is following a simple rule where it just asks "why?" about each new unrecognised noun in the explanations. After electricity, it will ask "what are electrons?" and then "what are quarks?" and so on until we reach the bottom of the universe and the questioning stops. At which point the small child will probably opine: "That's silly!"
The child has drilled a straight borehole down through concept-space, missing untold wealths of both theoretical and practical knowledge branching off on all sides. If the child had asked, "but why do the electrons make the light shine?" for example, we could have careened off down one of them. I am ready to tell the child all about filament wires and LEDs, about Kirchoff's Laws and resistance, inductance, capacitance, about power generation and the Grid, about static electricity and lightning and thunderstorms. If only the child wanted to ask me. I am some kind of contrived entity who cannot take initiative and can only answer the questions I was asked.
What better set of rules could have prompted the child to come up with the question, "but why do the electrons make the light shine?"
What is the best algorithm for asking iterated "why?" type questions, that leads to maximum exploration of the concept space?
Can the questioning algorithm be tuned to promote "focussed" vs "unfocussed" exploration? ie, "focussed" exploration might be asking questions that result in learning one field "well", and getting a bunch of concepts that reaffirm each other. "Unfocussed" exploration might be broadly drifting across hundreds of fields and bumping into many many disconnected concepts.
Can the questioning algorithm be tuned to promote obtaining theoretical knowledge or practical knowledge specifically?
Focussed/unfocussed and theoretical/practical are two axes, are there others we should be caring about as well?
And how do you begin deciding what the questioning rules should be?
If you don't think Africa is going to flourish, then you don't think the human race is going to flourish."
Why do people take this guy seriously? Africa's GDP per capita has gone sideways over the past decade, and that's WITH massive amounts of foreign financial and food aid, and foreign investment in and operation of natural resource projects etc. There's already diminishing returns and that's with everything being done for them by foreigners.
On what planet does ANYTHING about Africa indicate that things are going to radically change? Economists are a joke.
You say they do, but do they really? It seems like almost any other issue is capable of holding left wingers’ attention better than climate change. No climate protest in the US has ever come close to the size of the BLM protests/riots. That is, a single black man dying was to the left a more serious issue than something we’re told will lead to human extinction. And it’s literally more acceptable for a politician to say they don’t believe in climate change than to say anything that contradicts the race orthodoxy in America.
Look at LGBTQI stuff – this is unambiguously more of an animating force for the american left. I see LGBT stuff on a daily basis – T-shirts, buttons, stickers, twitter emojis, ‘inclusive/safe space’ signs in stores, health facilities etc. but it’s very rare to see anyone with climate change merchandise or signage. Online leftist ‘activism’ has to be many times more concerned with transgender issues than it is climate change. A few hundred ‘transwomen’ being killed this year (mostly by black men, not government policies) is somehow a bigger problem than the destruction of the world’s ecosystems and earth being left ‘uninhabitable’ for future generations. You can care about multiple issues, but the proportion of concern is wildly out of whack with the supposed seriousness.
Or look at immigration. And in the US, Trump is currently a betting favorite and is otherwise considered to have a high likelihood of being able to win the presidential election. And a huge amount of his support is due to his views on immigration. The Biden government is doing very little to stop even illegal immigration and are punishing southern states for trying to protect the border. There are also a number of recent examples of left-wing parties in Europe losing elections or losing parliamentary seats to anti-immigration parties. The upshot of all of this being that left-wing parties are essentially refusing to compromise on immigration and so are effectively willing to risk losing elections in order to not support stricter immigration policy – this also means that they’re willing to risk right wing parties winning and repealing climate policies in order to save their pro-immigration policies. How does nobody see this?
If climate change is a serious as we’re told, why would you not being willing to compromise on almost anything else in order to gain/stay in power to ensure climate friendly policy is pursued? I’m sure there’s some dumb cope argument about mass immigration is actually necessary to fight climate change or something, but this is obvious nonsense, especially since immigration from developing countries to the US substantially increases people’s carbon footprint.
The one thing that *would* make kind of make sense is that this immigration is being used to increase the number of future left-wing voters, and this increase in electoral power will give the Democrats (and whoever in Europe) the freedom to aggressively pursue climate policy, though even if this were true I hardly expect anyone to admit it.
I had an insight the other day about the phenomenon of "left-wing billionaires". Why do so many ultra-rich people find common cause with the ideology that promotes stealing from the rich, instead of supporting the free market that made them wealthy?
The insight came while dealing with my small children (2 and 4). They were arguing over some toy, and I told them to just share it. This was, I realised almost immediately, not fair; the stupid plastic toy belongs to one of the kids (the older one in this case) and if she doesn't want to let the other one play with it right now then that's her prerogative, but nonetheless my instinct was to just tell them to share it so that hopefully they'd shut up a bit. I incorrectly believed that since I don't really give a shit about this cheap plastic toy then nobody else should either.
Then I realised - this must be how billionaires see the world too -- with themselves as adults, the middle class as four-year-olds and the lower classes as two-year-olds. They don't care about the rights and wrongs, they're so far above the fray that they can't see anything wrong with just letting the lower classes take the middle classes' pathetic six- and seven-figure net worths if it will just calm them down a bit.
> Table 33: Do you support the solution of establishing one state or two states in the following formats:
> One-State Solution for Two Peoples: 5.4%
> A Palestinian state from the river to the sea: 74.7%
They were explicitly given an option for this supposed utopian democratic peaceful integrated state where everyone has rights... and they overwhelmingly denied it in favor of an Arab Muslim ethnostate resulting from the ethnic cleansing and genocide of all Israeli Jews.
Some other findings from the poll:
> Table 27: How much do you support the military operation carried out by the Palestinian resistance led by Hamas on October 7th?
> Extremely Support or Somewhat Support: 75%
> Table 29: How do you view the role of the following parties?
> Al Aqsa Brigade - Very Positive or Somewhat Positive: 79.8%
Held off on this one because it really is Not Safe For Work or indeed Not Safe For Life.
Tasting History, and making real garum. If you're ever wondering "How in the name of God did anyone think this was a good idea?" I can't answer that for you, but it will certainly make you ask that question! And yet, the end product is so very different:
I believe it was Scott who first tipped me off to Neom and The Line. Well, learning about linear cities has had a major effect on how I spend my time. I've put at least a thousand hours into this proposal:
LineLoop
(A Linear City + Vacuum Transit)
What happens when you take two crazy ideas (Hyperloop and The Line) and combine them with lots of common sense? Good things.
The Solow Model concludes that a nation can become more prosperous by building more capital until the burden of maintaining existing capital slows the growth down.
This seems reasonable to me. It's a very simplified model, but it's pretty predictive.
I'm interested in whether I'm missing something, but let me know if you've read the article, or if you're just addressing my simplification of the simplification of the real world.
I have a minor nitpick-- sometimes you get a technology like stainless steel, that slows depreciation (the need for maintenance), but that just delays the eventual slowdown.
There's an old saying that a tree doesn't grow to the sky, but no one won a Nobel prize for it.
Seeing Scott refer to his own kids as "surprisal-minimization engines" really rubbed me the wrong way, I suppose because of the reductionism of it, but also because it's actually so very highly speculative. This particular theory of the brain that Scott has covered before wasn't even covered in The World Behind the World, neuroscientist Erik Hoel's book on just how in the dark we are on the brain. In particular, he said we don't really have a theory on how the brain works: there are multiple candidates, and it doesn't seem like there are any that can win out, as they all have severe issues. So I wonder what Scott (or anyone who agrees with that characterization of the brain as a "surprisal-minimization engine") would make of that book. It would be great if Scott did a book review of it. Think the thought that hurts the most right?
Let me share with you a little nugget of history which I just learned today.
To set the background: there's a new French version of "The Three Musketeers", split into two movies. A review of the first by an online acquaintance says it's the Muddy Middle Ages setting, even though it's the 17th century. Nobody in France is able to wash their faces because this must be before soap was invented.
I was pretty sure Castile soap was around in the 17th century and looking it up, discovered: "What unlikely thing do you think there will be fighting over?" and the answer is "Popish soap":
"In the 17th century, the soap caused controversy in England, since it supplanted the unnamed local soap after the Spanish Catholic manufacturers purchased the monopoly on the soap from the cash-strapped Carolinian government. Its ties to Catholicism caused a public-relations campaign to be established, featuring washerwomen showing how much more effective local soaps were than Castile soap. The sale of a monopoly in Protestant England to a Catholic company caused a great uproar, ending with the Castile soap company eventually being stripped of the monopoly."
"Popish soap was a derisive name applied to soap manufactured under a patent granted by Charles I. Because the board of the manufacturing company included Catholics, the term Popish Soap (after The Pope) was applied to this monopoly commodity. It was said by anti-Catholics to be particularly harmful to linen and washerwomen's hands.
During the personal rule of the English King Charles I (1629–1640), one of the ways in which he attempted to raise money was through the granting of patents. This came about as a result of a loophole in the statute forbidding such action.
One such patent was granted to a soap corporation.
The soap industry was overseen by Lord Treasurer Portland and his friends, all of whom displayed Catholic character. When Portland died, Laud and Cottington contended over the company, which increased annual profits to the crown to nearly 33,000 pounds by the end of the 1630s.
It was alleged that popish soap scarred the soul as well as skin and fabric."
There'll be no Catholic soap touching our good Protestant undergarments, thank you very much! 😁
Visual art such as paintings. I don't get it. I get music, novels, movies, even cinematography, but i am a Philistine when it comes to great paintings. I don't get them. I mean I can enjoy spending an hour in an art museum looking quickly at a bunch of paintings. But I don't feel emotional when I view a painting, whereas I do when listening to music, reading novels or watching movies.
Is this just some innate thing one can't fix? Or can I learn to appreciate visual art at a deeper level? What's your experience?
so honest question: if I just get my recommendations from GiveWell anyway, why wouldn't i just give to GiveWell and let them distribute it? Like, clearly I *already* trust them.
i definitely get those people exist, but its possible to not like catholicism from a conservative perspective too. The people you mean dislike it more from a political
perspective, and see any church as an extension of progressive leftism; its theology is that ideally. You have the same with the whole "liberation theology" mess.
a conservative would see ritual as an impediment, and tradition as a barrier to actually experiencing Jesus. The old cliche "a personal relationship" is meaningful here; they'd agree on the core aspects of the christian story in terms of the new testament but "holy tradition" no. it may be an american thing, but the anger against that other cliche "organized religion" has a point. too many "high" churches were operating by rote at a time when people needed much more. Ritual can be beautiful but also suffocating.
My girlfriend and I are having a discussion: does it make any kind of sense to teach our child to read/write Shavian (or some other artificial phonemic alphabet - ie, where one letter maps to one sound) before trying to tackle English?
Points against: it's twice as much to learn, and useless in real life. At best the child is wasting its time; at worst it's actively being confused by the two different alphabets, and could end up retarded relative to other kids.
Potential points for: I'm wondering if it wouldn't actually make English easier to learn, by virtue of splitting one compound lesson into two simpler ones. By learning Shavian first, the child only has to understand the basic concept of "sound out shapes" --> "put them together to make words". So by the time it comes to dealing with all the irregularities and special cases of English, it already has a love of reading and a fair amount of preexisting experience it can use to make sense of the new system.
Main points against: it's not what other people do, which means there may be pitfalls that haven't been discovered; it's not the default and therefore requires extra effort on our part.
According to the same link, recurrent depression (or depression that started in teenage years) is 4-5 times more likely to occur in people who have parents (or siblings) with recurrent depression.
My question: is anyone here aware of any research suggesting how this changes depending on whether one or both parents have depression? For example, if having a single parent with recurrent depression makes it 4-5x more likely that a child has depression; does having two parents with recurrent depression make no difference compared to having a single parent, some kind of "additive" difference (e.g., 8-10x more likely that a child has depression) or a "multiplicative" difference of some sort (e.g., 15x, 20x more likely)?
My intuition is that "no difference" is nonsense, "additive" would work for traits that are monogenic, "multiplicative" would be more for polygenic traits. Since depression is clearly (very?) polygenic (and likely actually covers multiple different etiologies), then I would expect it to be definitively "multiplicative", but I have no idea if my line or reasoning makes sense and I have no intuition "how much" polygenic it would be?
I am aware this is all extremely unclear, we don't know anything scientifically with high confidence, we're just talking probabilities. Still, I'd like to know what the "state of evidence" is nevertheless.
I'm a ACX/EA kind of guy born in Mumbai, grew up in Singapore (and organized a ACX meetup there once), and currently study CS and write science fiction in America. I'm in Mumbai 8 Jan - 24 Jan. Anybody would like to meet up / know how I can meet some cool people there?
Forgive me if I’ve missed this, but is there a way to see our prediction answers to the 2023 prediction form we did? I checked the old form link, but just says the form is closed. Want to go back and check my answers to see what I was right about.
Christmas is the king of holidays and the most dramatic of them all, The Most. The darkest and coldest time of the year becomes the best time of the year, the warmest and fullest of light.
Christmas is right around the winter solstice, the longest night of the year. Leading up to it, the days get shorter and the nights get longer, it gets worse and worse, and when it reaches its peak and it seems that the light and warmth might just die away entirely, we find that we can create our own light and warmth. And then we find that the light and warmth are coming back on their own, anyway!
Of course, Christmas is not actually on the solstice. It's 3 days after the darkest point. Hence the relevance to the Jesus Christ story. Christ is the exemplar, the perfect person, who shows us that if we are what we could be, we can handle poverty and hunger, persecution and prosecution, torture and death, and still experience ourselves as within the perfect peace of the Kingdom of Heaven each step along the way. At least, up until things reach the solstice, the darkest moment, and no matter how good you are, you cry out "My God, my God, why hast thou forsaken me?"
But Christ is the example of maintaining faith even when you can't remember why it makes sense, so he dies shortly thereafter, still expressing faith, even as he descends into hell. For 3 days. And 3 days after the darkest point is Christmas, the brightest light out of the darkest place.
Christmas is The Most, technically speaking. It's the holiday of all the colors, and also the holiday of white light, the colors in harmony. It's the brightest, warmest, most colorful, most fun, most sacred holiday, so of course it has to be placed at the worst time of the year, so that it can redeem it, and by so doing, make it into the best time of the year, and by so doing, prove that redemption is possible for any time of year or season of life. Heaven out of Hell, and even in the midst of it, is the reason for hope.
Talking about domestic violence below reminds me-- are the good theories about why significant number of people imprint on people who don't want them. You'd think, or at least I'd think, that the lack of reciprocation would would cause the obsession to fade and end. Wouldn't looking for a new partner be more efficient? Why do domestic abusers put so much work into keeping a partner who doesn't want them?
I'll grant that in the ancestral environment, there wouldn't be a large choice of potential partners, but there would still be some.
A General Theory of Love by Lewis and Amini starts from the fact that humans are the only animals that die from lack of contact in infancy, rather than just being damaged by it. (CW: some very unpleasant experiments.)
It's so common for people to suffer from problems with eating and sleeping after a loss of a partner that it may be hard to think it needs an explanation, but the book gets into how intimacy regulates metabolisms. There was even work on *which* parts of a baby rats metabolism were regulated by which sorts of contact with a mother.
When I read the book, it didn't cover the question of how it happens that A is looking for regulation from B and B just wants to be somewhere else. It's plausible to me that imagination plays a large part for humans, that it's possible to imagine a partner who will be what one needs and imagination causes a sort of imprinting.
I have a notion that a lot of people and a lot of government policy is mercantilist with regard to immigrants. I'm taking mercantilism to be a belief that money is the true measure of prosperity.
So, an immigrant who shows up mostly with a willingness to work is a drain. They're competing with the local workforce and/or they want government money.
The only immigrants who are welcome are wealthy retirees.
Did you know for example that Jews regard Christmas as an ANTI-holiday when it's forbidden to study Torah? Or that if the Jewish Bible *does* manage to magically predict Jesus that the most likely such prophetic verse makes a mockery of modern Christianity? Or that some of the greatest (Pharisee) Rabbis of his day did not oppose him?
Lots more but the lecture holds deeper civilizational points than the large collection of arcana that come up.
Due to my complete lack of success at gaining the attention of the "viral makers" (or even not to be shadowbanned for who knows what reasons), if you know how to get this video out there I think it's likely to do a lot of good.
No doubt the fools will nitpick as they are wont, but if enough good and wise people watch it, whether of faith traditions or not, we might succeed in getting our act together so that we can enjoy the superlative writings of the nascent Kai and Lyra rather than living Lehrer's song when, in no time at all (2027 by Scott's optimistic calculation) we are on schedule to "all go together when we go".
Pretty funny since anyone on the right who says that this is what the left do and believe are called conspiracy theorists, even on here. (I remember one clown on here saying that Scott's claim (on the fertility rate post) that making Korea have Kenyan (or something to that effect) would fundamentally change Korea into a different country and not be the same as Koreans having kids again as an example of "white replacement theory"). Well, here it is, straight from the horse's mouth.
Of course, the left have been saying this stuff for decades and salivating at the thought of white americans becoming a minority. But here it is, on tape, coming from the president himself. I don't care if you think it's a good thing this is happening, that's entirely not the point. The point is that this is referred to as a "conspiracy theory", which is categorically false.
Has anyone in the Rationalist community written on the subject of generalization w/r/t biology, economics, or other similarly complex systems? I'm pessimistic about the idea we can even generalize in these disciplines, because any proposed universal laws are constantly shattered, and each event must be handled on a near case-by-case basis. The web of causality here becomes so complex it feels nigh impossible for the human brain to keep up, even presuming you have an enormous amount of knowledge on the topic.
I saw the thing about the genetically engineered plaque, but I was trying to minimize my internet presence for Advent. So this is a little late, but I just want to point out that, as a respiratory therapist, one of my regular duties is brushing teeth. Every patient on the ventilator is supposed to get their teeth brushed every four hours. We use disposable toothbrushes and sponges with antimicrobial rinses, not toothpaste, but the general idea is the same: to kill off as many germs as possible.
We don't do this because we care about patients' teeth (though they can develop dental problems). We do it as part of a standard pack of precautions against ventilator-associated pneumonia, or VAP. VAP is insanely common; basically if you've been on the vent long enough, your lungs will inevitably start producing enormous quantities of horrible gunk. It's just a matter of when. Intubation bypasses your normal airway defenses--you can't cough, you can't swallow, you just sit there with your oral secretions dribbling down your throat and pooling in your lungs. Ideal growth conditions.
Basic oral care isn't all that hard and doesn't take a lot of time, but it still gets skipped a lot. In most hospitals, AFAIK, it's a shared responsibility with nurses, alternating. If you have a really conscientious nurse who isn't too busy, she'll probably do it a lot. If your nurse doesn't understand the importance, or is getting hammered with other duties, or if your RT is getting called on to transport a bunch of people to MRI ... well, it's one of the first corners to cut. A couple minutes apiece, saved on six vented patients, adds up to a fair amount of time. Not proud of it, but sometimes something has to go.
You probably see where this is going: a genetically engineered benign bacterium playing watchdog sounds like it has enormous potential to fight against one of the most common scourges of the ICU. The current project has limitations, and mostly what it means is that I would expect anybody who goes into the ICU with the new benign plaque already on to have slightly better outcomes on the vent, if only because it's tricky to really cover all surfaces of the teeth when there's a tube in the way (and sometimes patients clench down so you practically need a crowbar to get in there). Actually applying it as described to already intubated patients would be tricky for the same reasons--I don't think you could remove all the old plaque thoroughly enough around the tube, though it might be possible with tracheostomy patients.
I'm mostly interested in how this would open the door or build goodwill for something more ambitious. If it would be possible to get something that spreads itself around the oral cavity more aggressively--well, I'm not a microbiologist, I don't know what's possible, but it's a very attractive thought. The mouth really is just about the filthiest part of the human body. I wouldn't be surprised if the benign plaques of tomorrow actually decreased disease rates in the general population. A lot of elderly people are just chronically weak and can't protect their airway worth a damn. Could you get benign bacteria growing in the nose too, murdering flu bugs? I don't know, but it'd be cool.
So I think the general opinion (including my own) is that the chatGPTs have basically blown right through the OG Turing test. But has this actually been empirically tested?
What I'm thinking is that it would be fun and instructive to put together a pool of humans and LLMs and set up anonymous 3-way 15 minute text chats (or audio/visual zooms with the appropriate filters) comprising one LLM and 2 randomly selected humans with the explicit task that each human has to identify which of their interlocutors is a person and which is a machine. Completely open format and lying is permitted, obviously. Maybe this would turn out to be easy -- or maybe it wouldn't! Plus of course the LLMs would be learning and improving all the time...
And you could have leader boards for the humans. And rankings like ELO for chess. People like it when you gamify stuff like this, yes?
For advanced play, you could also have incognito LLMs lurking in the human pool.
TBH I'm not sure why this hasn't already been done. Or maybe it has -- in which case, please advise with a link!
I've just had fun doing the Finders course (now 45 days to awakening course), an online meditation course. Can anyone recommend any other online meditation courses?
I don't do Christmas. I'm an atheist, and after my parents died I lost hold of doing it because it's a family thing. I'm enjoying sharing thoughts here tonight with other people who apparently are not in the throes of Christmas. So you other guys, why aren't you?
Sorry I didn't think of this during the discussions of Scott's kidney, but better late then never. You can be an effective altruist (lower case) without giving up a body part by being a "human guinea pig" - volunteering to take part in a clinical trial or research study. Likely only most convenient in major metro areas with the right institutions, they are easily found by researching on-line. In the medical mecca where I live they even advertise on the subway. Some typical ones I did were diet studies where they give you a supplement to eat which may or may not contain e.g. strawberries, and you go in for regular visits where they do a blood draw and do some tests like bone density or muscle strength. There is a stipend which is not a lot but enough to be useful spending money for someone like me who's retired. And you get to help advance medical science. I also got an entertaining story to tell at parties about "The effects of testosterone supplementation in older men..."
I have a collection of stamps from the space race that my grandfather left me. He says it's every stamp emitted by every country to celebrate every major launch or accomplishment. They are in boxes and folders, not particularly organized. What should I do with it?
My dear friends, I would like to talk to you about the police, because in my mind they are utterly useless-to-harmful, but I know that real responsible adults out there in the world think they're a net good, and I'd like to hear why.
My position has always been "why does anyone on earth want the police" (in the US, as they currently exist)? If you're upper or middle class, they're spending 80% of their manpower literally farming you for ticket revenue rather than actually solving crime or doing anything useful. If you're lower class or a minority, they spend their time murdering and/or oppressing you for kicks.
Who, besides 80 year old Karens who feel "safe" because police are out there harassing skating teenagers, poor people, and minorities, actually wants the police as they currently stand?
But wait! If we defunded the police, crime would be rampant! *Gestures broadly at rampant and public homelessness, drug abuse, systematic retail theft, etc in every major city, that the cops do literally nothing about.*
Have you ever had your car vandalized or broken into? Your house burgled? Your self mugged? Were the police any help AT ALL?
My street car got broken into 4 times last year, including two broken windows. In fact, the block it was parked on one of the nights it had a window broken I later found out *routinely* has a line of 15+ cars get their windows broken, on something like a semi-weekly to monthly basis. That's conservatively $10k a month in direct financial damages just for the window replacement, much less anything actually stolen. But if you report this to the police? They basically laugh at you. If you really persist, they'll make you fill out a form and immediately throw it away. Do they ever actually solve a crime, or step up enforcement? No, of course not, that's not their job.
I mean forget property crime, less than half of *MURDERS* are solved per year! What about the 25k rape kit backlog they're sitting on and going nothing about? And the cops are using their significant manpower and budgets to what? 80% of it to give you traffic tickets over minor speed and parking infractions.
But wait, what about speeding in school zones, and drunk or reckless driving, and <insert other legitimate social function>? Sure, I'll grant you that - so the police need 10% of their current manpopwer and budget to do that stuff, because they aren't using their current size and manpower to actually solve any real crimes. A third of it goes to deterrence and enforcement of those kinds of things, and 2/3 goes to solving real crimes.
But society will go to hell without the deterrent effect of police! *Gestures again at rampant homelessness, theft, drug abuse etc*. We're ALREADY THERE. The cops do literally nothing about any actual crime. When you need help urgently the police are just 30 minutes away, and even if they arrive they'll do nothing.
Basically, I've pretty much always perceived that we're being soaked for taxes to support a bunch of murderous thugs who spend most of their time farming us for revenue and harassing teens and minorities, and the recent social decline and significant increases in crime post-COVID have only made that evaluation firmer and more salient.
So seriously, what is the case for the police as they currently stand? Why shouldn't we defund them to 10-20% of their current level and write some laws saying they need to spend those resources on fighting actual crimes instead of minor moving and parking violations?
Who here is actually happy with the law enforcement status quo, and what specifically makes you happy about it?
I've been thinking, as one does, about qualia and emotions in the context of the current AI hype.
Let's consider emotions as a generic standin for qualia, with pain need a specific example.
Suppose we have a Tesla (as a stand-in for an AI). Does the Tesla feel qualia or pain or whatever? There seems absolutely no reason to think this. So why not?
Answer A: emotions ARE essentially motivations. Our Tesla can be "intelligent" without being motivated; its motivations are outsourced to us. So it can indicate, with varying degrees of urgency, that it needs more electricity or pumped up tires, but it doesn't need to actually CARE about these, we do that. We are the ones who will decide that, sure it may destroy the car, but we need to get to the hospital right away, and we will do that by running the car into the ground.
But that's not quite right.
Suppose we breed feral Teslas that need to find their own electricity every night, they can't just sit at home waiting for the owner to make decisions for them. We can run our genetic algorithms to reprogram them to calculate the distance to the nearest charging station, to make money from carrying passengers, to calculate how to spend money on tires vs electricity and balance passenger carrying time vs charging time, etc.
But none of this seems like it has to involved emotions and qualia. It's basically a question of either Lagrange multipliers and math or, if you don't know the exact numbers, run a bunch of simulations and see the range of outcomes.
So let's consider
Answer B: Think of a situation where you are like the feral Tesla, so you are putting together your budget. You may want a big screen TV, but you know rationally that rent and food come first, then utilities, then transport, etc. You may be disappointed that you can't buy your big screen TV, but you don't feel PAIN or any of the qualia of pain.
But you do feel (a version of) pain when you are hungry. So what's the difference?
Hypothesis is that emotions and pain are how you get a chemical system to ACT as a computer. When you're an animal hiding in your cave, of course you want to stay there and sleep. Outside it's cold and wet and dark and the other animals are very scary. If you had a computer brain you could run the simulations and decide whether the optimal strategy tonight is to go out and find food, or stay home and be safe. No qualia required.
But you don't have a computer brain, just a vat of chemicals, and the best you can do in terms of simulation is something like release one set of chemicals that vote for staying home (it's warm, outside is wet, cold, dark, dangerous), a different set of chemicals that vote for going out (hungry!!!) and see which set overpowers the other set. Qualia are the epiphenomena of simulating the world (roughly) via chemicals that have side-effects on the rest of the body so they're also triggering sensors and muscles god knows what else as they're doing their sloppy pseudo-calculation.
This doesn't exactly get at "why does it feel like it feels" but I think it does get at "why do animals feel and computers [or for that matter Chinese Rooms] don't".
Which implies that when you have a computer computer, whether it's running a standard program or a neural net, it doesn't need emotions-as-qualia, it can achieve whatever goals it is "programmed" to achieve via bloodless calculation.
This doesn't answer other issues like "consciousness" or even "the AI's will kill us all", but it does answer the question I found most interesting over the past month or so of hubbub.
What do you think? Makes sense? Or you can easily find some situations where it doesn't work and we have to keep searching?
Around the Winter Solstice (for those in the Northern Hemisphere), I blogged last year about Vincent van Gogh and a famous quote of his: "in spite of us, and without our permission, there comes at last an end to the bitter frosts."
Near the end of this post, there are links to several extraordinary, van Gogh-themed works. I'd LOVE to see all of these find a wider audience – please view and share any or all, if you're so inclined!
* A trailer for, and a short "making of" video for, "Loving Vincent," "the world’s first fully painted feature film. … Every one of the 65,000 [or more] frames of the film is an oil-painting ..."
* A stunning performance of "Vincent" by Don McLean (of "American Pie" fame), on a German TV show in 1972.
* A segment from the "Vincent and the Doctor" episode of the long-running British science fiction TV series, "Doctor Who." An alternate, time-travel reality for a day in Vincent van Gogh's life.
What's the end game for the migrant crisis in NYC?
Are the migrants just going to stop ending up in NYC? How?
If they don't stop, where are they going to live?
If your answer is YIMBYism, who's going to pay for the new buildings? Because the government will have to either fund them themselves or force developers to lease the apartments to migrants with no money instead of many thousands of middle class workers who would live in the city if it were a bit cheaper or easier to find an apartment.
People believe, or pretend to believe, that the main problem with AI is this: someone, after many attempts and using sophisticated techniques, manages to get AI to print an offensive or sexually explicit message, or draw a picture that contains nudity, and it hurts their feelings. You act like this is a huge problem.
In fact, the problem with AI is that some dictator will get strong AI before democracies get it and use it to destroy democratic countries.
Democratic countries should develop AI as fast as they can.
A big economic problem right now is that legible qualifications on a resume -- you worked at X for Y years -- don't provide enough information. You could have been an A+ worker at X, or a D worker at X. Any set of legible qualifications on a resume could correspond to an extremely wide range of actual productivity. So employers in tech give an all day on-site interview to test skills de novo and that's pretty expensive and time consuming and demoralizing to fly people out there and reject most of them. So they compensate by rejecting 99% at the top of the funnel and getting tons of false negatives. There are probably a lot of people languishing in jobs that are below their abilities just because there wasn't any way to make their abilities legible enough on LinkedIn. I would not be surprised if improving the legibility of labor productivity boosted GDP by trillions. We're already paying 700B/year on postsecondary education for a half-assed signal of labor productivity.
So Iranian/Yemeni militias have decided to just shut down all trade through the red sea and AFAICT the rest of the world is just... letting them? The US doesn't care enough to really get involved since it doesn't affect US trade, And Europe and China (whose trade it does affect) don't want to get involved because the Yemenis blame Israel and they don't want to seem like they're taking Israel's side. This seems... bad, in multiple ways.
I'm a pescatarian, because I draw the line of "enough probable consciousness to have rights" somewhere between fish and tetrapods - or at least the kind of tetrapods that people eat. Obviously this is completely arbitrary, it's not like I have an in depth knowledge of any of this. But given that I have chosen to draw the line here, can I eat cephalopods? I know they are considered honorary vertebrates, but do they have enough mentality to be considered honorary tetrapods?
There will be human activists asserting AI rights by December 2025.
Consequential prediction: the AI riskers will need to beef up their security arrangements. If you start believing in AI sentience their endeavours look like research into how to keep your field hands submissive and productive in the antebellum Deep South.
This is probably the last public Open Thread, or the one before the last if I'm unlucky and Scott decided to make another one before 11:59:59 PM 31st of Dec in any of the regions of the world where ACX commenters vibe.
Whatever it may be, I wish to all who read this a Merry Christmas, a happy new year, an excellent whatever-you-want-to-call-it holiday. Pray, wish well, or give blessing to people going through horrors in Gaza, Sudan, Ukraine and all the countless named and unnamed places where Man's beastly nature has trampled over its angelic counterpart and compelled him against his better knowledge to wreak havoc on his human brethrens.
Veritasium, in what's very likely his last 2023 video as well, has discussed the Prisoner's Dilemma[1]. The prisoner's Dilemma is one of those things where you can say its name followed by the "All Else Is Commentary" declaration. In one English paragraph and a 2x2 table, you can summarize War, Peace, Nuclear Profileration, Sexual Competition, Social Media Comment Sections, Suicide Bombers and Altrusim.
Prior to watching the video, I knew that Tit-For-Tat is the "best" strategy, but I didn't know just how strong the basis for that is. It turns out the basis is surprisingly robust and impressive[2]. Robert Axelrod, an American political scientist who wrote a 1984 book on The Evolution of Cooperation, asked people from all around Academia to submit programs that will play the iterated Prisoner's Dilemma. He got around 15 entries, to which he added the entry "Random'', which simply and obliviously picks cooperation or defection 50% of the time independently of past behaviour by both the player and the adversary.
Tit-For-Tat won.
In a later rematch, Axelrod got more than 60 entries. Moreover, Axelrod published the results of the first contest, he published all programs and the leaderboard complete with statistics about how often each program won. People developed all sorts of "Nasty" (more on that shortly) programs that will try to exploit the goodwill and naivety of "Nice" (more on that shortly) programs, such as a program called "Tester'', which tries to detect if it's playing against the "Generous Tit-For-Tat" strategy, a strategy that allows its opponents two defections before it defects. Tester will test the water, defecting once and seeing if the adversary still cooperates after that and if so, defecting every other turn to take maximum advantage of Generous Tit-For-Tat's forgiveness. Despite all this, Tit-For-Tat still won the contest, again, against 60+ other contestants.
In the first contest, all top-scoring strategies were "Nice". In the second contest, all but one were. What's a "Nice" strategy ? Wikipedia describes it thus :
> [A Nice strategy] will not defect before its opponent does (this is sometimes referred to as an "optimistic" algorithm) [...] A purely selfish strategy will not "cheat" on its opponent for purely self-interested reasons first.
Along with that trait, top scoring strategies shared others :
> [2] Retaliating: The strategy must sometimes retaliate. An example of a non-retaliating strategy is Always Cooperate, a very bad choice that will frequently be exploited by "nasty" strategies.
> [3] Forgiving: Successful strategies must be forgiving. Though players will retaliate, they will cooperate again if the opponent does not continue to defect. This can stop long runs of revenge and counter-revenge, maximizing points.
> [4] Non-envious: The strategy must not strive to score more than the opponent.
The video doesn't mention [4] as it is, instead replacing it with "Clear", meaning the strategy has to be as unambiguous and clear in its pattern of behaviour as possible, or otherwise its adversary will simply find it easier to assume the worst than to understand it and behave accordingly.
So, Niceness wins right ? It seems so. Goodness pays. In an ecological simulation - playing strategies against multiple copies of each other and other strategies as if they're organisms - every single Nasty strategy went extinct or lost most of their numbers while the population of Nice strategies skyrocketed. The one Nasty player that thrived the longest was Harrington, the sole nasty top player in the second contest, and it was apparently only thriving because its method of play was suitable against other Nasty players, once all the other Nasties went extinct or nearly-extinct, it too diminished greatly in numbers against the thriving Nice players.
=======================================
There is one massive assumption you need to make, though, so that this analysis makes sense : Non-Noisy actions. Non-Noisy actions means that when a player makes an action, it's (A) Conveyed to the environment exactly as intended by the player (B) Perceived by the adversary exactly as it happened. Basically, a flavor of Perfect Information (just after the fact).
To appreciate how disastrous noise could be, the video asks you to consider 2 copies of Tit-For-Tat playing each other. In a non-noisy environment, that's boring, the first player cooperates, its adversary replies with cooperation, and the two live happily ever after from here in a skyrocketing, ever-higher mountain of trust.
But consider what happens if a player, meaning cooperation, "mistakenly" commits defection. You can imagine this as possibly the highly complex and unpredictable environment twisting the player's good-intentioned action, like someone who wants to carry groceries for you unintentionally stepping on your foot, or maybe it's just the player's propagandized and twitchy adversary perceiving a cooperation as a defection, like the soldiers of a certain army firing on unarmed people despite them waving white flags and screaming "help" in the soldiers' native language.
Whatever you imagine it to be, the impossible happens : A Tit-For-Tat strategy commits defection first. The other, naturally, commits defection as well. But then the first, unaware of the other's reasoning because in its eyes its last decision was cooperation, also commits defection. Ergo, a cycle of violence. The video discusses a solution for this : Tit-For-Tat should probabilistically retaliate, forgiving a defection 10% or so of the time. It's implied that this doesn't reduce Tit-For-Tat's fitness as a player. But this is just one strategy and one scenario involving it, the general problem is how devastating Noise could be.
I was planning to connect the Prisoner's Dilemma with the one conflict that has occupied an unrivaled percentage of my brain power and mental health for the past ~3 months, but this wall of text has already gotten too long for anybody to enjoy reading. So maybe next time.
“More light entering the pupil causes it to constrict. But [for normal people] simply imagining something bright like the Sun also causes (a smaller, but still measurable) constriction. Aphantasics show perfectly typical pupillary responses to actual changes in light. However, their pupils do not change to imagined light.”
Here are a few videos made by a guy who does "vigilante justice" to thieves -- nothing too serious, just a fart bomb, lots of color, bomb-like countdown, and recording their reactions.
On one level, it is fun. On another level, what the fuck is the police doing, so that this guy is left to fight crime during his free time and he seems to do a much better job.
(Especially, towards the end of the last video, he mentions a place where thieves sell stolen stuff. Everyone knows where the place is, because it was on a TV a year ago. And still, it is safe for the thieves to go there and sell stolen stuff, because... apparently, no one cares. What the fuck?)
Since I'm doing the preparations including the dinner for a diminished family, it really does make a difference when it's only yourself and/or one other, and the entire family. Whether or not Scott and family celebrate the Solstice, Hanukah or Christmas, this will be the first one as a family and subsequent years it really will be different now he has children.
Congratulations once again, and season's greetings to all on here!
I’m looking for medical blog recommendations with a similar style to Scott’s but focused almost entirely on doctoring and medicine. Ideally one with thoughtful takes on both the day-to-day of medicine and interesting topics in medicine as a whole. As medical student, I’ve searched around for one but never found one anywhere near the quality of Scott’s
What's the research on the most effective way to get students to want to learn? I don't mean inspiring interest in any particular subject. I mean inducing a generalized desire in a student to seek knowledge.
Plato thought that goal of education should be exactly that. Even if you don't believe in the importance of training philosopher-kings it's easy to see the appeal, as it would make the job of any teacher much easier and presumably produce a much brighter populace for any society. But in a world of generative AI I find myself thinking that it's becoming ever more important that we figure out how to actually achieve it. My own experience with ChatGPT has been immensely positive. I use it as an expert tutor and it's help me make a lot of progress as a programmer and even been useful for brainstorming ideas for work. But the stories and anecdotes I hear from teachers about how it's being used by students are far more depressing (i.e. mass cheating).
Given that the generative genie is out of the box, there's no hope in simply telling students not to use it and likely no hope in devising ways to detect it's use that students won't be able to work around. Outright banning it would be simply impossible and incredibly foolish, given the very real benefits it can provide as a teaching tool. The ideal course of action then would seem to be encouraging in students a desire to learn for it's own sake, thus ensuring (or at least predisposing) them to utilize the tech in such a way as to enhance learning rather than avoid it. So, to return to my question at the beginning, what if any research is there on the best way to accomplish that?
I'm trying to find a story (possibly a Reddit post?) that Gwern referenced somewhere (now I can't even find the reference), where a character is stuck in a time-loop until they can beat Stockfish (the chess engine).
I don't think I'm crazy but my Googling, ChatGPT queries and god help me Binging have all come up empty, appreciate anyone who can help out
While reading Varieties of Religious Experience by William James I came across something called “mind-cure” which turned out to be an early 19th century movement around the idea that the mind has actual healing powers over the body. Digging deeper into the rabbit hole led me to a person called Phineas Quimby, who, after being treated for tuberculosis with a tooth rotting crystal, came up with his own system of healing, which feels somewhat like early psychology. On top of that, one of his patients, Marg Eddy founded the Church of Christ, Scientist, giving a start to Christian Science. Fascinating stories, both of them. I wrote a somewhat longish piece about them ( https://open.substack.com/pub/valentinsocial/p/200-years-of-positive-thinking ). Does anyone know more about these people and their lives or any fun facts about the movements that they started?
I'm trying to figure out how US infrastructure spending (in dollar terms, total or per Capita) compares with other countries (specifically France, but any other countries would be good).
For the US, I found this table which implies that it's about 8% of 7 trillion total government spending, or about 500 billion a year:
I haven't managed to find much other data - google and gpt just give me things like "France plans to spend 40 billion euros by 2040", but I think that's under a specific program, not total infrastructure spending.
(Overall I suspect the US spends more total or per Capita on infrastructure than other countries with much better infra, but I'm having trouble finding good data to prove it).
Let's re-read Scott's ancient posts! "The place we lived when we were young" (https://pastebin.com/RLRWsach) is not at all about rationality, and all about Scott's childhood and melancholy.
When doing probability calibration puzzles\predictionbook.com stuff, how do you handle updating your prior predictions with regards to new information?
Sticking to your prior commitment and not changing anything throws away the new data
Simply changing the probability estimate throws away the old data from your original estimate
Simply making a new estimate alongside the old one double-counts the data that went into the original estimate
The mathematically rigorous way would be to introduce a complexity penalty to all your predictions, but that it a whole other can of worms.
For better or worse, I've decided to start writing about psychiatry as a way to look more critically at the things I think I've learned over the course of my training. I just finished writing and publishing my first essay on the commonly used sleep drug trazodone, and (spoiler alert) how bad the evidence is for its use, as well as some interesting things I learned along the way about how we study sleep, the mental pain model of depression, and some long dead psychiatric drugs.
I have been hugely inspired by Scott, so I think it's only appropriate that I post this here.
Future topics in the works include a plea for people to please stop panicking about QTc intervals all the time, why it's important to have consult rejections be part of a consult service in a teaching hospital, and an analysis of the side-effects of ECT and whether or not we should think about using it much sooner than we currently do.
P.S. If this stuff is supposed to be relegated to the classifieds threads, just lmk and I'll delete it!
Are people actually surprise-minimizers? It seems to me that people are surprise-optimizers, perhaps with a setting of keeping surprise fairly low. However, to various degrees, people get bored, take up dangerous sports, marry dangerous people, follow sports, and like actual surprises in fiction.
There are definitely surprises people don't want, like a body part failing to do what one intends. Well, usually. A lot of people like getting drunk, and a smaller number like doing extreme sports to the point of exhaustion.
Estimates for AI-Extinction of >90%¹ are unreasonable because it would mean that every dependency's complement is <10%. If one were to take aggressive, conservative priors for the path to PCM (paperclip maximizer), at least a few of them would be greater than 10% on their own², or at the very least, some non-intersecting combo of them would be as well.
I. Precondition failure
1. Moore's Law plateaus
2. the scaling hypothesis fails
3. appropriate warning shots arrive
4. controls arrive without warning shots
5. AI-safety theorists succeed
II. Known doubts about PCMs come true
6. instrumental convergence doesn't work as planned
7. the orthogonality hypothesis doesn't pan out
8. (I'm not an expert on PCMs, but those who are can fill in these blanks)
9. ...
10. ...
III. And of course, there are several unknown unknowns about PCMs, given how theoretical they are
11. ...
12. ...
13. ...
14. ...
15. ...
I forget who, but some professor who is a hero of this community, went back and calculated that estimates for atmospheric ignition pre-Trinity Test didn't appropriately factor in the probability of getting the maths/physics wrong (i.e., unknown unknowns) and that estimates should have been much higher. Meanwhile, the PCM story has less empirical support than atmospheric ignition, so the same "unknown unknowns" concern should also apply.
How is Eliezer's forecast for PCMs not an example of bad faith or a failure of imagination³?
[2]: As an example, even a reasonable, well-informed Moore's Law hawk would concede a ~10% chance that it stops short somewhere. And by well-informed, I mean someone knowledgeable about Rock's Law, basic physics, etc.
On last weeks Open Thread I asked for advice regarding a wart treatment approach (see https://www.astralcodexten.com/p/open-thread-307/comment/45638770). Somehow this formed into a small study to determine if the treatment works. I don't have any website or so prepared yet, so here are the basics:
We compare three different methods (one was previously studied by professionals, one is a placebo, one is my new method). If a participant has more than one wart, it's ok to try several methods. Participants have to buy the material themselves, expect about 20$. Trial duration is 2 weeks, primary outcome that is measured is "is the wart gone after 2 weeks", but we try to also measure size to see if it shrinks.
Once I have results of about 10 participants (or if we don't find 10 participants then as soon as the ones we did find finished their 2 weeks) I will post the results (including the info what was the method under study and what the placebo) as a reply to this comment.
Send a mail to Sunnyafternoon123@proton.me to participate, tell me how many warts you have so I can assign the right number of methods.
I wrote about how rituals help us to find meaning at the end of life.
> But there is a conflict between following rituals and making the choice to live one’s life freely. I expect Sartre would say that rituals are good when they are freely chosen.
I think I recall there being a reason for not allowing “top” sorting for comments here, but I can’t remember/find it. What was the reason?
I’m assuming it’s to avoid mainstream/hivemind topics or to allow for unbiased coverage, but I think there has to be a better method/algo. Or maybe allow readers to make the decision themselves?
Good morning, or maybe I should say Sad morning. My Youtube channel was cancelled the other day for "inappropriate content," even though it was mainly about woodworking. This was a real blow because I invested some precious cash in camera equipment. You can read more about this nonsense here: https://falsechoices.substack.com/p/modern-woodworking. Merry Christmas and Happy New Year's to you all.
Would you be able to clarify the resolution criteria of this market? The correct probability is heavily dependent on what sort of errors you're going to count as "bizarre hacking-like tricks" vs. legitimate mistakes, and without further guidance we're not really able to place meaningful bets.
A problem to think about:
I'm trapped in a room with a small child. The child asks me why the lights work. I explain that the light switch turns them off and on. The child asks "why?" and I explain about the electrical circuits in the walls. The child asks "what are circuits?" and I say they're loops of wire. The child asks "what is electricity?" and I explain about electrons.
The child is following a simple rule where it just asks "why?" about each new unrecognised noun in the explanations. After electricity, it will ask "what are electrons?" and then "what are quarks?" and so on until we reach the bottom of the universe and the questioning stops. At which point the small child will probably opine: "That's silly!"
The child has drilled a straight borehole down through concept-space, missing untold wealths of both theoretical and practical knowledge branching off on all sides. If the child had asked, "but why do the electrons make the light shine?" for example, we could have careened off down one of them. I am ready to tell the child all about filament wires and LEDs, about Kirchoff's Laws and resistance, inductance, capacitance, about power generation and the Grid, about static electricity and lightning and thunderstorms. If only the child wanted to ask me. I am some kind of contrived entity who cannot take initiative and can only answer the questions I was asked.
What better set of rules could have prompted the child to come up with the question, "but why do the electrons make the light shine?"
What is the best algorithm for asking iterated "why?" type questions, that leads to maximum exploration of the concept space?
Can the questioning algorithm be tuned to promote "focussed" vs "unfocussed" exploration? ie, "focussed" exploration might be asking questions that result in learning one field "well", and getting a bunch of concepts that reaffirm each other. "Unfocussed" exploration might be broadly drifting across hundreds of fields and bumping into many many disconnected concepts.
Can the questioning algorithm be tuned to promote obtaining theoretical knowledge or practical knowledge specifically?
Focussed/unfocussed and theoretical/practical are two axes, are there others we should be caring about as well?
And how do you begin deciding what the questioning rules should be?
Noah Smith says:
https://twitter.com/Noahpinion/status/1740922602183639451
"Afropessimism is just pessimism, period.
If you don't think Africa is going to flourish, then you don't think the human race is going to flourish."
Why do people take this guy seriously? Africa's GDP per capita has gone sideways over the past decade, and that's WITH massive amounts of foreign financial and food aid, and foreign investment in and operation of natural resource projects etc. There's already diminishing returns and that's with everything being done for them by foreigners.
On what planet does ANYTHING about Africa indicate that things are going to radically change? Economists are a joke.
Why don’t the left care about climate change?
You say they do, but do they really? It seems like almost any other issue is capable of holding left wingers’ attention better than climate change. No climate protest in the US has ever come close to the size of the BLM protests/riots. That is, a single black man dying was to the left a more serious issue than something we’re told will lead to human extinction. And it’s literally more acceptable for a politician to say they don’t believe in climate change than to say anything that contradicts the race orthodoxy in America.
Look at LGBTQI stuff – this is unambiguously more of an animating force for the american left. I see LGBT stuff on a daily basis – T-shirts, buttons, stickers, twitter emojis, ‘inclusive/safe space’ signs in stores, health facilities etc. but it’s very rare to see anyone with climate change merchandise or signage. Online leftist ‘activism’ has to be many times more concerned with transgender issues than it is climate change. A few hundred ‘transwomen’ being killed this year (mostly by black men, not government policies) is somehow a bigger problem than the destruction of the world’s ecosystems and earth being left ‘uninhabitable’ for future generations. You can care about multiple issues, but the proportion of concern is wildly out of whack with the supposed seriousness.
Or look at immigration. And in the US, Trump is currently a betting favorite and is otherwise considered to have a high likelihood of being able to win the presidential election. And a huge amount of his support is due to his views on immigration. The Biden government is doing very little to stop even illegal immigration and are punishing southern states for trying to protect the border. There are also a number of recent examples of left-wing parties in Europe losing elections or losing parliamentary seats to anti-immigration parties. The upshot of all of this being that left-wing parties are essentially refusing to compromise on immigration and so are effectively willing to risk losing elections in order to not support stricter immigration policy – this also means that they’re willing to risk right wing parties winning and repealing climate policies in order to save their pro-immigration policies. How does nobody see this?
If climate change is a serious as we’re told, why would you not being willing to compromise on almost anything else in order to gain/stay in power to ensure climate friendly policy is pursued? I’m sure there’s some dumb cope argument about mass immigration is actually necessary to fight climate change or something, but this is obvious nonsense, especially since immigration from developing countries to the US substantially increases people’s carbon footprint.
The one thing that *would* make kind of make sense is that this immigration is being used to increase the number of future left-wing voters, and this increase in electoral power will give the Democrats (and whoever in Europe) the freedom to aggressively pursue climate policy, though even if this were true I hardly expect anyone to admit it.
90 millionth hole to emerge in israeli rape accusations
https://twitter.com/MaxBlumenthal/status/1740647278455525377
I had an insight the other day about the phenomenon of "left-wing billionaires". Why do so many ultra-rich people find common cause with the ideology that promotes stealing from the rich, instead of supporting the free market that made them wealthy?
The insight came while dealing with my small children (2 and 4). They were arguing over some toy, and I told them to just share it. This was, I realised almost immediately, not fair; the stupid plastic toy belongs to one of the kids (the older one in this case) and if she doesn't want to let the other one play with it right now then that's her prerogative, but nonetheless my instinct was to just tell them to share it so that hopefully they'd shut up a bit. I incorrectly believed that since I don't really give a shit about this cheap plastic toy then nobody else should either.
Then I realised - this must be how billionaires see the world too -- with themselves as adults, the middle class as four-year-olds and the lower classes as two-year-olds. They don't care about the rights and wrongs, they're so far above the fray that they can't see anything wrong with just letting the lower classes take the middle classes' pathetic six- and seven-figure net worths if it will just calm them down a bit.
Let’s see what Palestinians mean by “from the river to the sea”:
https://www.awrad.org/files/server/polls/polls2023/Public%20Opinion%20Poll%20-%20Gaza%20War%202023%20-%20Tables%20of%20Results.pdf
> Table 33: Do you support the solution of establishing one state or two states in the following formats:
> One-State Solution for Two Peoples: 5.4%
> A Palestinian state from the river to the sea: 74.7%
They were explicitly given an option for this supposed utopian democratic peaceful integrated state where everyone has rights... and they overwhelmingly denied it in favor of an Arab Muslim ethnostate resulting from the ethnic cleansing and genocide of all Israeli Jews.
Some other findings from the poll:
> Table 27: How much do you support the military operation carried out by the Palestinian resistance led by Hamas on October 7th?
> Extremely Support or Somewhat Support: 75%
> Table 29: How do you view the role of the following parties?
> Al Aqsa Brigade - Very Positive or Somewhat Positive: 79.8%
Do all human names have meaning? That's mostly the case in my native Arabic. Is it the case in English? Other languages? Why/why not?
Held off on this one because it really is Not Safe For Work or indeed Not Safe For Life.
Tasting History, and making real garum. If you're ever wondering "How in the name of God did anyone think this was a good idea?" I can't answer that for you, but it will certainly make you ask that question! And yet, the end product is so very different:
https://www.youtube.com/watch?v=ICZww0DtQKk
I believe it was Scott who first tipped me off to Neom and The Line. Well, learning about linear cities has had a major effect on how I spend my time. I've put at least a thousand hours into this proposal:
LineLoop
(A Linear City + Vacuum Transit)
What happens when you take two crazy ideas (Hyperloop and The Line) and combine them with lots of common sense? Good things.
Rodes.pub/LineLoop
https://www.noahpinion.blog/p/what-the-solow-model-can-teach-us/comments
The Solow Model concludes that a nation can become more prosperous by building more capital until the burden of maintaining existing capital slows the growth down.
This seems reasonable to me. It's a very simplified model, but it's pretty predictive.
I'm interested in whether I'm missing something, but let me know if you've read the article, or if you're just addressing my simplification of the simplification of the real world.
I have a minor nitpick-- sometimes you get a technology like stainless steel, that slows depreciation (the need for maintenance), but that just delays the eventual slowdown.
There's an old saying that a tree doesn't grow to the sky, but no one won a Nobel prize for it.
Gaston Glock has died, which means it's time to reach out to antagonistic activist organizations to put in his obituary:
https://www.upi.com/Top_News/World-News/2023/12/28/austria-Gaston-Glock-dies/7191703741657/
I'm sure the prewritten Bibi obit includes extensive quotations from CAIR.
Seeing Scott refer to his own kids as "surprisal-minimization engines" really rubbed me the wrong way, I suppose because of the reductionism of it, but also because it's actually so very highly speculative. This particular theory of the brain that Scott has covered before wasn't even covered in The World Behind the World, neuroscientist Erik Hoel's book on just how in the dark we are on the brain. In particular, he said we don't really have a theory on how the brain works: there are multiple candidates, and it doesn't seem like there are any that can win out, as they all have severe issues. So I wonder what Scott (or anyone who agrees with that characterization of the brain as a "surprisal-minimization engine") would make of that book. It would be great if Scott did a book review of it. Think the thought that hurts the most right?
Let me share with you a little nugget of history which I just learned today.
To set the background: there's a new French version of "The Three Musketeers", split into two movies. A review of the first by an online acquaintance says it's the Muddy Middle Ages setting, even though it's the 17th century. Nobody in France is able to wash their faces because this must be before soap was invented.
I was pretty sure Castile soap was around in the 17th century and looking it up, discovered: "What unlikely thing do you think there will be fighting over?" and the answer is "Popish soap":
https://en.wikipedia.org/wiki/Castile_soap
"In the 17th century, the soap caused controversy in England, since it supplanted the unnamed local soap after the Spanish Catholic manufacturers purchased the monopoly on the soap from the cash-strapped Carolinian government. Its ties to Catholicism caused a public-relations campaign to be established, featuring washerwomen showing how much more effective local soaps were than Castile soap. The sale of a monopoly in Protestant England to a Catholic company caused a great uproar, ending with the Castile soap company eventually being stripped of the monopoly."
https://en.wikipedia.org/wiki/Popish_soap
"Popish soap was a derisive name applied to soap manufactured under a patent granted by Charles I. Because the board of the manufacturing company included Catholics, the term Popish Soap (after The Pope) was applied to this monopoly commodity. It was said by anti-Catholics to be particularly harmful to linen and washerwomen's hands.
During the personal rule of the English King Charles I (1629–1640), one of the ways in which he attempted to raise money was through the granting of patents. This came about as a result of a loophole in the statute forbidding such action.
One such patent was granted to a soap corporation.
The soap industry was overseen by Lord Treasurer Portland and his friends, all of whom displayed Catholic character. When Portland died, Laud and Cottington contended over the company, which increased annual profits to the crown to nearly 33,000 pounds by the end of the 1630s.
It was alleged that popish soap scarred the soul as well as skin and fabric."
There'll be no Catholic soap touching our good Protestant undergarments, thank you very much! 😁
Visual art such as paintings. I don't get it. I get music, novels, movies, even cinematography, but i am a Philistine when it comes to great paintings. I don't get them. I mean I can enjoy spending an hour in an art museum looking quickly at a bunch of paintings. But I don't feel emotional when I view a painting, whereas I do when listening to music, reading novels or watching movies.
Is this just some innate thing one can't fix? Or can I learn to appreciate visual art at a deeper level? What's your experience?
so honest question: if I just get my recommendations from GiveWell anyway, why wouldn't i just give to GiveWell and let them distribute it? Like, clearly I *already* trust them.
i definitely get those people exist, but its possible to not like catholicism from a conservative perspective too. The people you mean dislike it more from a political
perspective, and see any church as an extension of progressive leftism; its theology is that ideally. You have the same with the whole "liberation theology" mess.
a conservative would see ritual as an impediment, and tradition as a barrier to actually experiencing Jesus. The old cliche "a personal relationship" is meaningful here; they'd agree on the core aspects of the christian story in terms of the new testament but "holy tradition" no. it may be an american thing, but the anger against that other cliche "organized religion" has a point. too many "high" churches were operating by rote at a time when people needed much more. Ritual can be beautiful but also suffocating.
i just wanted to mention that perspective.
My girlfriend and I are having a discussion: does it make any kind of sense to teach our child to read/write Shavian (or some other artificial phonemic alphabet - ie, where one letter maps to one sound) before trying to tackle English?
Points against: it's twice as much to learn, and useless in real life. At best the child is wasting its time; at worst it's actively being confused by the two different alphabets, and could end up retarded relative to other kids.
Potential points for: I'm wondering if it wouldn't actually make English easier to learn, by virtue of splitting one compound lesson into two simpler ones. By learning Shavian first, the child only has to understand the basic concept of "sound out shapes" --> "put them together to make words". So by the time it comes to dealing with all the irregularities and special cases of English, it already has a love of reading and a fair amount of preexisting experience it can use to make sense of the new system.
Main points against: it's not what other people do, which means there may be pitfalls that haven't been discovered; it's not the default and therefore requires extra effort on our part.
The NYT has sued OpenAI and Msoft for copyright infringement:
https://apnews.com/article/nyt-new-york-times-openai-microsoft-6ea53a8ad3efa06ee4643b697df0ba57
Hello!
Here's a horribly simplifying question.
According to https://med.stanford.edu/depressiongenetics/mddandgenes.html?tab=proxy 50% of depression is caused by genes (source: identical twin studies).
According to the same link, recurrent depression (or depression that started in teenage years) is 4-5 times more likely to occur in people who have parents (or siblings) with recurrent depression.
My question: is anyone here aware of any research suggesting how this changes depending on whether one or both parents have depression? For example, if having a single parent with recurrent depression makes it 4-5x more likely that a child has depression; does having two parents with recurrent depression make no difference compared to having a single parent, some kind of "additive" difference (e.g., 8-10x more likely that a child has depression) or a "multiplicative" difference of some sort (e.g., 15x, 20x more likely)?
My intuition is that "no difference" is nonsense, "additive" would work for traits that are monogenic, "multiplicative" would be more for polygenic traits. Since depression is clearly (very?) polygenic (and likely actually covers multiple different etiologies), then I would expect it to be definitively "multiplicative", but I have no idea if my line or reasoning makes sense and I have no intuition "how much" polygenic it would be?
I am aware this is all extremely unclear, we don't know anything scientifically with high confidence, we're just talking probabilities. Still, I'd like to know what the "state of evidence" is nevertheless.
I'm a ACX/EA kind of guy born in Mumbai, grew up in Singapore (and organized a ACX meetup there once), and currently study CS and write science fiction in America. I'm in Mumbai 8 Jan - 24 Jan. Anybody would like to meet up / know how I can meet some cool people there?
Please email me at omnileshpune@gmail.com
Just ran across a great present for somebody: ROKR 3D Puzzle Marble Run
You put it together, then it does the great stuff shown below. Buyers are writing that it is well-engineered.
https://www.youtube.com/watch?v=iw8wyxBultc
I remember Scott once reviewing something by Herzl but now I cant find it anywhere. Can someone point me to where I can find it?
Forgive me if I’ve missed this, but is there a way to see our prediction answers to the 2023 prediction form we did? I checked the old form link, but just says the form is closed. Want to go back and check my answers to see what I was right about.
Christmas is the king of holidays and the most dramatic of them all, The Most. The darkest and coldest time of the year becomes the best time of the year, the warmest and fullest of light.
Christmas is right around the winter solstice, the longest night of the year. Leading up to it, the days get shorter and the nights get longer, it gets worse and worse, and when it reaches its peak and it seems that the light and warmth might just die away entirely, we find that we can create our own light and warmth. And then we find that the light and warmth are coming back on their own, anyway!
Of course, Christmas is not actually on the solstice. It's 3 days after the darkest point. Hence the relevance to the Jesus Christ story. Christ is the exemplar, the perfect person, who shows us that if we are what we could be, we can handle poverty and hunger, persecution and prosecution, torture and death, and still experience ourselves as within the perfect peace of the Kingdom of Heaven each step along the way. At least, up until things reach the solstice, the darkest moment, and no matter how good you are, you cry out "My God, my God, why hast thou forsaken me?"
But Christ is the example of maintaining faith even when you can't remember why it makes sense, so he dies shortly thereafter, still expressing faith, even as he descends into hell. For 3 days. And 3 days after the darkest point is Christmas, the brightest light out of the darkest place.
Christmas is The Most, technically speaking. It's the holiday of all the colors, and also the holiday of white light, the colors in harmony. It's the brightest, warmest, most colorful, most fun, most sacred holiday, so of course it has to be placed at the worst time of the year, so that it can redeem it, and by so doing, make it into the best time of the year, and by so doing, prove that redemption is possible for any time of year or season of life. Heaven out of Hell, and even in the midst of it, is the reason for hope.
Talking about domestic violence below reminds me-- are the good theories about why significant number of people imprint on people who don't want them. You'd think, or at least I'd think, that the lack of reciprocation would would cause the obsession to fade and end. Wouldn't looking for a new partner be more efficient? Why do domestic abusers put so much work into keeping a partner who doesn't want them?
I'll grant that in the ancestral environment, there wouldn't be a large choice of potential partners, but there would still be some.
https://www.amazon.com/General-Theory-Love-Thomas-Lewis/dp/0375709223
A General Theory of Love by Lewis and Amini starts from the fact that humans are the only animals that die from lack of contact in infancy, rather than just being damaged by it. (CW: some very unpleasant experiments.)
It's so common for people to suffer from problems with eating and sleeping after a loss of a partner that it may be hard to think it needs an explanation, but the book gets into how intimacy regulates metabolisms. There was even work on *which* parts of a baby rats metabolism were regulated by which sorts of contact with a mother.
When I read the book, it didn't cover the question of how it happens that A is looking for regulation from B and B just wants to be somewhere else. It's plausible to me that imagination plays a large part for humans, that it's possible to imagine a partner who will be what one needs and imagination causes a sort of imprinting.
I have a notion that a lot of people and a lot of government policy is mercantilist with regard to immigrants. I'm taking mercantilism to be a belief that money is the true measure of prosperity.
So, an immigrant who shows up mostly with a willingness to work is a drain. They're competing with the local workforce and/or they want government money.
The only immigrants who are welcome are wealthy retirees.
I'm an (agnostic) Orthodox Rabbi who knows a lot about books but nothing about the modern internet.
In honor of Christmas I recorded and posted the following video regarding the Jewish perception of Jesus.
https://youtu.be/vKyqMk1tnVg?feature=shared
Did you know for example that Jews regard Christmas as an ANTI-holiday when it's forbidden to study Torah? Or that if the Jewish Bible *does* manage to magically predict Jesus that the most likely such prophetic verse makes a mockery of modern Christianity? Or that some of the greatest (Pharisee) Rabbis of his day did not oppose him?
Lots more but the lecture holds deeper civilizational points than the large collection of arcana that come up.
Due to my complete lack of success at gaining the attention of the "viral makers" (or even not to be shadowbanned for who knows what reasons), if you know how to get this video out there I think it's likely to do a lot of good.
No doubt the fools will nitpick as they are wont, but if enough good and wise people watch it, whether of faith traditions or not, we might succeed in getting our act together so that we can enjoy the superlative writings of the nascent Kai and Lyra rather than living Lehrer's song when, in no time at all (2027 by Scott's optimistic calculation) we are on schedule to "all go together when we go".
https://youtu.be/vKyqMk1tnVg?feature=shared
Joe Biden: Non-white immigration will never stop, and white Americans becoming a minority "is the source of our strength": https://twitter.com/VivekGRamaswamy/status/1731888060848132097
Pretty funny since anyone on the right who says that this is what the left do and believe are called conspiracy theorists, even on here. (I remember one clown on here saying that Scott's claim (on the fertility rate post) that making Korea have Kenyan (or something to that effect) would fundamentally change Korea into a different country and not be the same as Koreans having kids again as an example of "white replacement theory"). Well, here it is, straight from the horse's mouth.
Of course, the left have been saying this stuff for decades and salivating at the thought of white americans becoming a minority. But here it is, on tape, coming from the president himself. I don't care if you think it's a good thing this is happening, that's entirely not the point. The point is that this is referred to as a "conspiracy theory", which is categorically false.
ACX grants: I applied with a fundamental physics project. Here is an explanation: https://mariopasquato.substack.com/p/testing-universal-free-fall-with
Any thoughts?
Has anyone in the Rationalist community written on the subject of generalization w/r/t biology, economics, or other similarly complex systems? I'm pessimistic about the idea we can even generalize in these disciplines, because any proposed universal laws are constantly shattered, and each event must be handled on a near case-by-case basis. The web of causality here becomes so complex it feels nigh impossible for the human brain to keep up, even presuming you have an enormous amount of knowledge on the topic.
I saw the thing about the genetically engineered plaque, but I was trying to minimize my internet presence for Advent. So this is a little late, but I just want to point out that, as a respiratory therapist, one of my regular duties is brushing teeth. Every patient on the ventilator is supposed to get their teeth brushed every four hours. We use disposable toothbrushes and sponges with antimicrobial rinses, not toothpaste, but the general idea is the same: to kill off as many germs as possible.
We don't do this because we care about patients' teeth (though they can develop dental problems). We do it as part of a standard pack of precautions against ventilator-associated pneumonia, or VAP. VAP is insanely common; basically if you've been on the vent long enough, your lungs will inevitably start producing enormous quantities of horrible gunk. It's just a matter of when. Intubation bypasses your normal airway defenses--you can't cough, you can't swallow, you just sit there with your oral secretions dribbling down your throat and pooling in your lungs. Ideal growth conditions.
Basic oral care isn't all that hard and doesn't take a lot of time, but it still gets skipped a lot. In most hospitals, AFAIK, it's a shared responsibility with nurses, alternating. If you have a really conscientious nurse who isn't too busy, she'll probably do it a lot. If your nurse doesn't understand the importance, or is getting hammered with other duties, or if your RT is getting called on to transport a bunch of people to MRI ... well, it's one of the first corners to cut. A couple minutes apiece, saved on six vented patients, adds up to a fair amount of time. Not proud of it, but sometimes something has to go.
You probably see where this is going: a genetically engineered benign bacterium playing watchdog sounds like it has enormous potential to fight against one of the most common scourges of the ICU. The current project has limitations, and mostly what it means is that I would expect anybody who goes into the ICU with the new benign plaque already on to have slightly better outcomes on the vent, if only because it's tricky to really cover all surfaces of the teeth when there's a tube in the way (and sometimes patients clench down so you practically need a crowbar to get in there). Actually applying it as described to already intubated patients would be tricky for the same reasons--I don't think you could remove all the old plaque thoroughly enough around the tube, though it might be possible with tracheostomy patients.
I'm mostly interested in how this would open the door or build goodwill for something more ambitious. If it would be possible to get something that spreads itself around the oral cavity more aggressively--well, I'm not a microbiologist, I don't know what's possible, but it's a very attractive thought. The mouth really is just about the filthiest part of the human body. I wouldn't be surprised if the benign plaques of tomorrow actually decreased disease rates in the general population. A lot of elderly people are just chronically weak and can't protect their airway worth a damn. Could you get benign bacteria growing in the nose too, murdering flu bugs? I don't know, but it'd be cool.
So I think the general opinion (including my own) is that the chatGPTs have basically blown right through the OG Turing test. But has this actually been empirically tested?
What I'm thinking is that it would be fun and instructive to put together a pool of humans and LLMs and set up anonymous 3-way 15 minute text chats (or audio/visual zooms with the appropriate filters) comprising one LLM and 2 randomly selected humans with the explicit task that each human has to identify which of their interlocutors is a person and which is a machine. Completely open format and lying is permitted, obviously. Maybe this would turn out to be easy -- or maybe it wouldn't! Plus of course the LLMs would be learning and improving all the time...
And you could have leader boards for the humans. And rankings like ELO for chess. People like it when you gamify stuff like this, yes?
For advanced play, you could also have incognito LLMs lurking in the human pool.
TBH I'm not sure why this hasn't already been done. Or maybe it has -- in which case, please advise with a link!
I've just had fun doing the Finders course (now 45 days to awakening course), an online meditation course. Can anyone recommend any other online meditation courses?
I don't do Christmas. I'm an atheist, and after my parents died I lost hold of doing it because it's a family thing. I'm enjoying sharing thoughts here tonight with other people who apparently are not in the throes of Christmas. So you other guys, why aren't you?
Sorry I didn't think of this during the discussions of Scott's kidney, but better late then never. You can be an effective altruist (lower case) without giving up a body part by being a "human guinea pig" - volunteering to take part in a clinical trial or research study. Likely only most convenient in major metro areas with the right institutions, they are easily found by researching on-line. In the medical mecca where I live they even advertise on the subway. Some typical ones I did were diet studies where they give you a supplement to eat which may or may not contain e.g. strawberries, and you go in for regular visits where they do a blood draw and do some tests like bone density or muscle strength. There is a stipend which is not a lot but enough to be useful spending money for someone like me who's retired. And you get to help advance medical science. I also got an entertaining story to tell at parties about "The effects of testosterone supplementation in older men..."
I have a collection of stamps from the space race that my grandfather left me. He says it's every stamp emitted by every country to celebrate every major launch or accomplishment. They are in boxes and folders, not particularly organized. What should I do with it?
My dear friends, I would like to talk to you about the police, because in my mind they are utterly useless-to-harmful, but I know that real responsible adults out there in the world think they're a net good, and I'd like to hear why.
My position has always been "why does anyone on earth want the police" (in the US, as they currently exist)? If you're upper or middle class, they're spending 80% of their manpower literally farming you for ticket revenue rather than actually solving crime or doing anything useful. If you're lower class or a minority, they spend their time murdering and/or oppressing you for kicks.
Who, besides 80 year old Karens who feel "safe" because police are out there harassing skating teenagers, poor people, and minorities, actually wants the police as they currently stand?
But wait! If we defunded the police, crime would be rampant! *Gestures broadly at rampant and public homelessness, drug abuse, systematic retail theft, etc in every major city, that the cops do literally nothing about.*
Have you ever had your car vandalized or broken into? Your house burgled? Your self mugged? Were the police any help AT ALL?
My street car got broken into 4 times last year, including two broken windows. In fact, the block it was parked on one of the nights it had a window broken I later found out *routinely* has a line of 15+ cars get their windows broken, on something like a semi-weekly to monthly basis. That's conservatively $10k a month in direct financial damages just for the window replacement, much less anything actually stolen. But if you report this to the police? They basically laugh at you. If you really persist, they'll make you fill out a form and immediately throw it away. Do they ever actually solve a crime, or step up enforcement? No, of course not, that's not their job.
I mean forget property crime, less than half of *MURDERS* are solved per year! What about the 25k rape kit backlog they're sitting on and going nothing about? And the cops are using their significant manpower and budgets to what? 80% of it to give you traffic tickets over minor speed and parking infractions.
But wait, what about speeding in school zones, and drunk or reckless driving, and <insert other legitimate social function>? Sure, I'll grant you that - so the police need 10% of their current manpopwer and budget to do that stuff, because they aren't using their current size and manpower to actually solve any real crimes. A third of it goes to deterrence and enforcement of those kinds of things, and 2/3 goes to solving real crimes.
But society will go to hell without the deterrent effect of police! *Gestures again at rampant homelessness, theft, drug abuse etc*. We're ALREADY THERE. The cops do literally nothing about any actual crime. When you need help urgently the police are just 30 minutes away, and even if they arrive they'll do nothing.
Basically, I've pretty much always perceived that we're being soaked for taxes to support a bunch of murderous thugs who spend most of their time farming us for revenue and harassing teens and minorities, and the recent social decline and significant increases in crime post-COVID have only made that evaluation firmer and more salient.
So seriously, what is the case for the police as they currently stand? Why shouldn't we defund them to 10-20% of their current level and write some laws saying they need to spend those resources on fighting actual crimes instead of minor moving and parking violations?
Who here is actually happy with the law enforcement status quo, and what specifically makes you happy about it?
This is something I wrote in email for my brother, but I think it's an interesting enough idea that it deserves wider discussion:
----------------------------------------------------------------
I've been thinking, as one does, about qualia and emotions in the context of the current AI hype.
Let's consider emotions as a generic standin for qualia, with pain need a specific example.
Suppose we have a Tesla (as a stand-in for an AI). Does the Tesla feel qualia or pain or whatever? There seems absolutely no reason to think this. So why not?
Answer A: emotions ARE essentially motivations. Our Tesla can be "intelligent" without being motivated; its motivations are outsourced to us. So it can indicate, with varying degrees of urgency, that it needs more electricity or pumped up tires, but it doesn't need to actually CARE about these, we do that. We are the ones who will decide that, sure it may destroy the car, but we need to get to the hospital right away, and we will do that by running the car into the ground.
But that's not quite right.
Suppose we breed feral Teslas that need to find their own electricity every night, they can't just sit at home waiting for the owner to make decisions for them. We can run our genetic algorithms to reprogram them to calculate the distance to the nearest charging station, to make money from carrying passengers, to calculate how to spend money on tires vs electricity and balance passenger carrying time vs charging time, etc.
But none of this seems like it has to involved emotions and qualia. It's basically a question of either Lagrange multipliers and math or, if you don't know the exact numbers, run a bunch of simulations and see the range of outcomes.
So let's consider
Answer B: Think of a situation where you are like the feral Tesla, so you are putting together your budget. You may want a big screen TV, but you know rationally that rent and food come first, then utilities, then transport, etc. You may be disappointed that you can't buy your big screen TV, but you don't feel PAIN or any of the qualia of pain.
But you do feel (a version of) pain when you are hungry. So what's the difference?
Hypothesis is that emotions and pain are how you get a chemical system to ACT as a computer. When you're an animal hiding in your cave, of course you want to stay there and sleep. Outside it's cold and wet and dark and the other animals are very scary. If you had a computer brain you could run the simulations and decide whether the optimal strategy tonight is to go out and find food, or stay home and be safe. No qualia required.
But you don't have a computer brain, just a vat of chemicals, and the best you can do in terms of simulation is something like release one set of chemicals that vote for staying home (it's warm, outside is wet, cold, dark, dangerous), a different set of chemicals that vote for going out (hungry!!!) and see which set overpowers the other set. Qualia are the epiphenomena of simulating the world (roughly) via chemicals that have side-effects on the rest of the body so they're also triggering sensors and muscles god knows what else as they're doing their sloppy pseudo-calculation.
This doesn't exactly get at "why does it feel like it feels" but I think it does get at "why do animals feel and computers [or for that matter Chinese Rooms] don't".
Which implies that when you have a computer computer, whether it's running a standard program or a neural net, it doesn't need emotions-as-qualia, it can achieve whatever goals it is "programmed" to achieve via bloodless calculation.
This doesn't answer other issues like "consciousness" or even "the AI's will kill us all", but it does answer the question I found most interesting over the past month or so of hubbub.
What do you think? Makes sense? Or you can easily find some situations where it doesn't work and we have to keep searching?
Congrats on the growing family, Scott!
Around the Winter Solstice (for those in the Northern Hemisphere), I blogged last year about Vincent van Gogh and a famous quote of his: "in spite of us, and without our permission, there comes at last an end to the bitter frosts."
https://fragmentsintime.substack.com/p/on-the-winter-solstice
Near the end of this post, there are links to several extraordinary, van Gogh-themed works. I'd LOVE to see all of these find a wider audience – please view and share any or all, if you're so inclined!
* A trailer for, and a short "making of" video for, "Loving Vincent," "the world’s first fully painted feature film. … Every one of the 65,000 [or more] frames of the film is an oil-painting ..."
* A stunning performance of "Vincent" by Don McLean (of "American Pie" fame), on a German TV show in 1972.
* A segment from the "Vincent and the Doctor" episode of the long-running British science fiction TV series, "Doctor Who." An alternate, time-travel reality for a day in Vincent van Gogh's life.
What's the end game for the migrant crisis in NYC?
Are the migrants just going to stop ending up in NYC? How?
If they don't stop, where are they going to live?
If your answer is YIMBYism, who's going to pay for the new buildings? Because the government will have to either fund them themselves or force developers to lease the apartments to migrants with no money instead of many thousands of middle class workers who would live in the city if it were a bit cheaper or easier to find an apartment.
People believe, or pretend to believe, that the main problem with AI is this: someone, after many attempts and using sophisticated techniques, manages to get AI to print an offensive or sexually explicit message, or draw a picture that contains nudity, and it hurts their feelings. You act like this is a huge problem.
In fact, the problem with AI is that some dictator will get strong AI before democracies get it and use it to destroy democratic countries.
Democratic countries should develop AI as fast as they can.
so i was reading Bryan Caplan's The Magic of Education ( https://www.econlib.org/archives/2011/11/the_magic_of_ed.html?utm_source=substack&utm_medium=email ) and I had an obvious idea that could save trillions of dollars worth of waste. Instead of going to college, young people people could work relevant jobs that give them grades.
A big economic problem right now is that legible qualifications on a resume -- you worked at X for Y years -- don't provide enough information. You could have been an A+ worker at X, or a D worker at X. Any set of legible qualifications on a resume could correspond to an extremely wide range of actual productivity. So employers in tech give an all day on-site interview to test skills de novo and that's pretty expensive and time consuming and demoralizing to fly people out there and reject most of them. So they compensate by rejecting 99% at the top of the funnel and getting tons of false negatives. There are probably a lot of people languishing in jobs that are below their abilities just because there wasn't any way to make their abilities legible enough on LinkedIn. I would not be surprised if improving the legibility of labor productivity boosted GDP by trillions. We're already paying 700B/year on postsecondary education for a half-assed signal of labor productivity.
So Iranian/Yemeni militias have decided to just shut down all trade through the red sea and AFAICT the rest of the world is just... letting them? The US doesn't care enough to really get involved since it doesn't affect US trade, And Europe and China (whose trade it does affect) don't want to get involved because the Yemenis blame Israel and they don't want to seem like they're taking Israel's side. This seems... bad, in multiple ways.
I'm a pescatarian, because I draw the line of "enough probable consciousness to have rights" somewhere between fish and tetrapods - or at least the kind of tetrapods that people eat. Obviously this is completely arbitrary, it's not like I have an in depth knowledge of any of this. But given that I have chosen to draw the line here, can I eat cephalopods? I know they are considered honorary vertebrates, but do they have enough mentality to be considered honorary tetrapods?
I wrote an essay for the solstice and the new year to share with you about the ways we’re connected as data points jn history, and how we can celebrate the tiny glowing embers of our own stories before we return them to the dark. I hope it speaks to you, thanks for sharing the light this year : https://open.substack.com/pub/bessstillman/p/how-the-light-gets-in-a-solstice?r=16l8ek&utm_campaign=post&utm_medium=web
Year end prediction
There will be human activists asserting AI rights by December 2025.
Consequential prediction: the AI riskers will need to beef up their security arrangements. If you start believing in AI sentience their endeavours look like research into how to keep your field hands submissive and productive in the antebellum Deep South.
This is probably the last public Open Thread, or the one before the last if I'm unlucky and Scott decided to make another one before 11:59:59 PM 31st of Dec in any of the regions of the world where ACX commenters vibe.
Whatever it may be, I wish to all who read this a Merry Christmas, a happy new year, an excellent whatever-you-want-to-call-it holiday. Pray, wish well, or give blessing to people going through horrors in Gaza, Sudan, Ukraine and all the countless named and unnamed places where Man's beastly nature has trampled over its angelic counterpart and compelled him against his better knowledge to wreak havoc on his human brethrens.
--------------------------------------------------------------------
Veritasium, in what's very likely his last 2023 video as well, has discussed the Prisoner's Dilemma[1]. The prisoner's Dilemma is one of those things where you can say its name followed by the "All Else Is Commentary" declaration. In one English paragraph and a 2x2 table, you can summarize War, Peace, Nuclear Profileration, Sexual Competition, Social Media Comment Sections, Suicide Bombers and Altrusim.
Prior to watching the video, I knew that Tit-For-Tat is the "best" strategy, but I didn't know just how strong the basis for that is. It turns out the basis is surprisingly robust and impressive[2]. Robert Axelrod, an American political scientist who wrote a 1984 book on The Evolution of Cooperation, asked people from all around Academia to submit programs that will play the iterated Prisoner's Dilemma. He got around 15 entries, to which he added the entry "Random'', which simply and obliviously picks cooperation or defection 50% of the time independently of past behaviour by both the player and the adversary.
Tit-For-Tat won.
In a later rematch, Axelrod got more than 60 entries. Moreover, Axelrod published the results of the first contest, he published all programs and the leaderboard complete with statistics about how often each program won. People developed all sorts of "Nasty" (more on that shortly) programs that will try to exploit the goodwill and naivety of "Nice" (more on that shortly) programs, such as a program called "Tester'', which tries to detect if it's playing against the "Generous Tit-For-Tat" strategy, a strategy that allows its opponents two defections before it defects. Tester will test the water, defecting once and seeing if the adversary still cooperates after that and if so, defecting every other turn to take maximum advantage of Generous Tit-For-Tat's forgiveness. Despite all this, Tit-For-Tat still won the contest, again, against 60+ other contestants.
In the first contest, all top-scoring strategies were "Nice". In the second contest, all but one were. What's a "Nice" strategy ? Wikipedia describes it thus :
> [A Nice strategy] will not defect before its opponent does (this is sometimes referred to as an "optimistic" algorithm) [...] A purely selfish strategy will not "cheat" on its opponent for purely self-interested reasons first.
Along with that trait, top scoring strategies shared others :
> [2] Retaliating: The strategy must sometimes retaliate. An example of a non-retaliating strategy is Always Cooperate, a very bad choice that will frequently be exploited by "nasty" strategies.
> [3] Forgiving: Successful strategies must be forgiving. Though players will retaliate, they will cooperate again if the opponent does not continue to defect. This can stop long runs of revenge and counter-revenge, maximizing points.
> [4] Non-envious: The strategy must not strive to score more than the opponent.
The video doesn't mention [4] as it is, instead replacing it with "Clear", meaning the strategy has to be as unambiguous and clear in its pattern of behaviour as possible, or otherwise its adversary will simply find it easier to assume the worst than to understand it and behave accordingly.
So, Niceness wins right ? It seems so. Goodness pays. In an ecological simulation - playing strategies against multiple copies of each other and other strategies as if they're organisms - every single Nasty strategy went extinct or lost most of their numbers while the population of Nice strategies skyrocketed. The one Nasty player that thrived the longest was Harrington, the sole nasty top player in the second contest, and it was apparently only thriving because its method of play was suitable against other Nasty players, once all the other Nasties went extinct or nearly-extinct, it too diminished greatly in numbers against the thriving Nice players.
=======================================
There is one massive assumption you need to make, though, so that this analysis makes sense : Non-Noisy actions. Non-Noisy actions means that when a player makes an action, it's (A) Conveyed to the environment exactly as intended by the player (B) Perceived by the adversary exactly as it happened. Basically, a flavor of Perfect Information (just after the fact).
To appreciate how disastrous noise could be, the video asks you to consider 2 copies of Tit-For-Tat playing each other. In a non-noisy environment, that's boring, the first player cooperates, its adversary replies with cooperation, and the two live happily ever after from here in a skyrocketing, ever-higher mountain of trust.
But consider what happens if a player, meaning cooperation, "mistakenly" commits defection. You can imagine this as possibly the highly complex and unpredictable environment twisting the player's good-intentioned action, like someone who wants to carry groceries for you unintentionally stepping on your foot, or maybe it's just the player's propagandized and twitchy adversary perceiving a cooperation as a defection, like the soldiers of a certain army firing on unarmed people despite them waving white flags and screaming "help" in the soldiers' native language.
Whatever you imagine it to be, the impossible happens : A Tit-For-Tat strategy commits defection first. The other, naturally, commits defection as well. But then the first, unaware of the other's reasoning because in its eyes its last decision was cooperation, also commits defection. Ergo, a cycle of violence. The video discusses a solution for this : Tit-For-Tat should probabilistically retaliate, forgiving a defection 10% or so of the time. It's implied that this doesn't reduce Tit-For-Tat's fitness as a player. But this is just one strategy and one scenario involving it, the general problem is how devastating Noise could be.
I was planning to connect the Prisoner's Dilemma with the one conflict that has occupied an unrivaled percentage of my brain power and mental health for the past ~3 months, but this wall of text has already gotten too long for anybody to enjoy reading. So maybe next time.
[1] What The Prisoner's Dilemma Reveals About Life, The Universe, and Everything | Veritasium : https://www.youtube.com/watch?v=mScpHTIi-kM
[2] https://en.wikipedia.org/wiki/Prisoner's_dilemma#Axelrod's_contest_and_successful_strategy_conditions
Here's (https://aeon.co/essays/the-moral-imperative-to-learn-from-diverse-phenomenal-experiences) an Aeon article about variation in reported subjective experience, and how "real" these variations are. E.g., on aphantasiacs, who claim to have no ability to picture things in their mind’s eye:
“More light entering the pupil causes it to constrict. But [for normal people] simply imagining something bright like the Sun also causes (a smaller, but still measurable) constriction. Aphantasics show perfectly typical pupillary responses to actual changes in light. However, their pupils do not change to imagined light.”
https://www.youtube.com/playlist?list=PLgeXOVaJo_gnexNopBzUKdl3QKoADJlS8
Here are a few videos made by a guy who does "vigilante justice" to thieves -- nothing too serious, just a fart bomb, lots of color, bomb-like countdown, and recording their reactions.
On one level, it is fun. On another level, what the fuck is the police doing, so that this guy is left to fight crime during his free time and he seems to do a much better job.
(Especially, towards the end of the last video, he mentions a place where thieves sell stolen stuff. Everyone knows where the place is, because it was on a TV a year ago. And still, it is safe for the thieves to go there and sell stolen stuff, because... apparently, no one cares. What the fuck?)
For the time of year - History for Atheists and The Date of Christmas:
https://www.youtube.com/watch?v=l3R7tgRgkR0
Since I'm doing the preparations including the dinner for a diminished family, it really does make a difference when it's only yourself and/or one other, and the entire family. Whether or not Scott and family celebrate the Solstice, Hanukah or Christmas, this will be the first one as a family and subsequent years it really will be different now he has children.
Congratulations once again, and season's greetings to all on here!
TLDR: looking for rationalist medicine blog!
I’m looking for medical blog recommendations with a similar style to Scott’s but focused almost entirely on doctoring and medicine. Ideally one with thoughtful takes on both the day-to-day of medicine and interesting topics in medicine as a whole. As medical student, I’ve searched around for one but never found one anywhere near the quality of Scott’s
What's the research on the most effective way to get students to want to learn? I don't mean inspiring interest in any particular subject. I mean inducing a generalized desire in a student to seek knowledge.
Plato thought that goal of education should be exactly that. Even if you don't believe in the importance of training philosopher-kings it's easy to see the appeal, as it would make the job of any teacher much easier and presumably produce a much brighter populace for any society. But in a world of generative AI I find myself thinking that it's becoming ever more important that we figure out how to actually achieve it. My own experience with ChatGPT has been immensely positive. I use it as an expert tutor and it's help me make a lot of progress as a programmer and even been useful for brainstorming ideas for work. But the stories and anecdotes I hear from teachers about how it's being used by students are far more depressing (i.e. mass cheating).
Given that the generative genie is out of the box, there's no hope in simply telling students not to use it and likely no hope in devising ways to detect it's use that students won't be able to work around. Outright banning it would be simply impossible and incredibly foolish, given the very real benefits it can provide as a teaching tool. The ideal course of action then would seem to be encouraging in students a desire to learn for it's own sake, thus ensuring (or at least predisposing) them to utilize the tech in such a way as to enhance learning rather than avoid it. So, to return to my question at the beginning, what if any research is there on the best way to accomplish that?
Merry Christmas!
I'm trying to find a story (possibly a Reddit post?) that Gwern referenced somewhere (now I can't even find the reference), where a character is stuck in a time-loop until they can beat Stockfish (the chess engine).
I don't think I'm crazy but my Googling, ChatGPT queries and god help me Binging have all come up empty, appreciate anyone who can help out
While reading Varieties of Religious Experience by William James I came across something called “mind-cure” which turned out to be an early 19th century movement around the idea that the mind has actual healing powers over the body. Digging deeper into the rabbit hole led me to a person called Phineas Quimby, who, after being treated for tuberculosis with a tooth rotting crystal, came up with his own system of healing, which feels somewhat like early psychology. On top of that, one of his patients, Marg Eddy founded the Church of Christ, Scientist, giving a start to Christian Science. Fascinating stories, both of them. I wrote a somewhat longish piece about them ( https://open.substack.com/pub/valentinsocial/p/200-years-of-positive-thinking ). Does anyone know more about these people and their lives or any fun facts about the movements that they started?
I'm trying to figure out how US infrastructure spending (in dollar terms, total or per Capita) compares with other countries (specifically France, but any other countries would be good).
For the US, I found this table which implies that it's about 8% of 7 trillion total government spending, or about 500 billion a year:
https://www.reddit.com/r/dataisbeautiful/comments/6ccrqy/oc_americas_7_trillion_budget_all_levels_of/?share_id=cqnEoILmUCSLiRji7nzXI
I haven't managed to find much other data - google and gpt just give me things like "France plans to spend 40 billion euros by 2040", but I think that's under a specific program, not total infrastructure spending.
(Overall I suspect the US spends more total or per Capita on infrastructure than other countries with much better infra, but I'm having trouble finding good data to prove it).
Let's re-read Scott's ancient posts! "The place we lived when we were young" (https://pastebin.com/RLRWsach) is not at all about rationality, and all about Scott's childhood and melancholy.
Find a collection of the old posts here: https://archive.ph/fCFQx.
When doing probability calibration puzzles\predictionbook.com stuff, how do you handle updating your prior predictions with regards to new information?
Sticking to your prior commitment and not changing anything throws away the new data
Simply changing the probability estimate throws away the old data from your original estimate
Simply making a new estimate alongside the old one double-counts the data that went into the original estimate
The mathematically rigorous way would be to introduce a complexity penalty to all your predictions, but that it a whole other can of worms.
What is satire again?
For better or worse, I've decided to start writing about psychiatry as a way to look more critically at the things I think I've learned over the course of my training. I just finished writing and publishing my first essay on the commonly used sleep drug trazodone, and (spoiler alert) how bad the evidence is for its use, as well as some interesting things I learned along the way about how we study sleep, the mental pain model of depression, and some long dead psychiatric drugs.
If that sounds interesting to you, you can read it here: https://polypharmacy.substack.com/p/wots-uh-the-deal-with-trazodone
I have been hugely inspired by Scott, so I think it's only appropriate that I post this here.
Future topics in the works include a plea for people to please stop panicking about QTc intervals all the time, why it's important to have consult rejections be part of a consult service in a teaching hospital, and an analysis of the side-effects of ECT and whether or not we should think about using it much sooner than we currently do.
P.S. If this stuff is supposed to be relegated to the classifieds threads, just lmk and I'll delete it!
Belated congratulations on the twins!
Are people actually surprise-minimizers? It seems to me that people are surprise-optimizers, perhaps with a setting of keeping surprise fairly low. However, to various degrees, people get bored, take up dangerous sports, marry dangerous people, follow sports, and like actual surprises in fiction.
There are definitely surprises people don't want, like a body part failing to do what one intends. Well, usually. A lot of people like getting drunk, and a smaller number like doing extreme sports to the point of exhaustion.
Estimates for AI-Extinction of >90%¹ are unreasonable because it would mean that every dependency's complement is <10%. If one were to take aggressive, conservative priors for the path to PCM (paperclip maximizer), at least a few of them would be greater than 10% on their own², or at the very least, some non-intersecting combo of them would be as well.
I. Precondition failure
1. Moore's Law plateaus
2. the scaling hypothesis fails
3. appropriate warning shots arrive
4. controls arrive without warning shots
5. AI-safety theorists succeed
II. Known doubts about PCMs come true
6. instrumental convergence doesn't work as planned
7. the orthogonality hypothesis doesn't pan out
8. (I'm not an expert on PCMs, but those who are can fill in these blanks)
9. ...
10. ...
III. And of course, there are several unknown unknowns about PCMs, given how theoretical they are
11. ...
12. ...
13. ...
14. ...
15. ...
I forget who, but some professor who is a hero of this community, went back and calculated that estimates for atmospheric ignition pre-Trinity Test didn't appropriately factor in the probability of getting the maths/physics wrong (i.e., unknown unknowns) and that estimates should have been much higher. Meanwhile, the PCM story has less empirical support than atmospheric ignition, so the same "unknown unknowns" concern should also apply.
How is Eliezer's forecast for PCMs not an example of bad faith or a failure of imagination³?
---
UPDATE: My probability theory language wasn't great. Original version: https://pastebin.com/MeDpVK4E
[1]: https://www.astralcodexten.com/p/why-i-am-not-as-much-of-a-doomer
[2]: As an example, even a reasonable, well-informed Moore's Law hawk would concede a ~10% chance that it stops short somewhere. And by well-informed, I mean someone knowledgeable about Rock's Law, basic physics, etc.
[3]: It doesn't take much creativity to come up with scenarios where the PCM story comes off the rails: https://www.astralcodexten.com/p/open-thread-305/comment/44770931
On last weeks Open Thread I asked for advice regarding a wart treatment approach (see https://www.astralcodexten.com/p/open-thread-307/comment/45638770). Somehow this formed into a small study to determine if the treatment works. I don't have any website or so prepared yet, so here are the basics:
We compare three different methods (one was previously studied by professionals, one is a placebo, one is my new method). If a participant has more than one wart, it's ok to try several methods. Participants have to buy the material themselves, expect about 20$. Trial duration is 2 weeks, primary outcome that is measured is "is the wart gone after 2 weeks", but we try to also measure size to see if it shrinks.
Once I have results of about 10 participants (or if we don't find 10 participants then as soon as the ones we did find finished their 2 weeks) I will post the results (including the info what was the method under study and what the placebo) as a reply to this comment.
Send a mail to Sunnyafternoon123@proton.me to participate, tell me how many warts you have so I can assign the right number of methods.
I wrote about how rituals help us to find meaning at the end of life.
> But there is a conflict between following rituals and making the choice to live one’s life freely. I expect Sartre would say that rituals are good when they are freely chosen.
https://raggedclown.substack.com/p/what-is-the-meaning-of-it-all
I think I recall there being a reason for not allowing “top” sorting for comments here, but I can’t remember/find it. What was the reason?
I’m assuming it’s to avoid mainstream/hivemind topics or to allow for unbiased coverage, but I think there has to be a better method/algo. Or maybe allow readers to make the decision themselves?
Good morning, or maybe I should say Sad morning. My Youtube channel was cancelled the other day for "inappropriate content," even though it was mainly about woodworking. This was a real blow because I invested some precious cash in camera equipment. You can read more about this nonsense here: https://falsechoices.substack.com/p/modern-woodworking. Merry Christmas and Happy New Year's to you all.
I made a post about a day ago reflecting on what it really means when people share outragous-looking posts online.
Are certain groups really as insane as they seem to be, or is there another reason why it it merely seems to be that way.
https://open.substack.com/pub/samschoenberg/p/ideologues-in-the-zoo?r=tjkkm&utm_campaign=post&utm_medium=web
Would you be able to clarify the resolution criteria of this market? The correct probability is heavily dependent on what sort of errors you're going to count as "bizarre hacking-like tricks" vs. legitimate mistakes, and without further guidance we're not really able to place meaningful bets.
https://manifold.markets/ScottAlexander/in-2028-will-gary-marcus-still-be-a