I recently moved to Israel and am looking to make new (English speaking) friends. Any suggestions about people to contact/meetups or events to attend are welcome. I would also be interested in non rationality focused stuff like book clubs etc
D) Card Game: Predictably Irrational - Feel free to bring your favorite games or distractions.
E) Walk & Talk: We usually have an hour-long walk and talk after the meeting starts. Two mini-malls with hot takeout food are easily accessible nearby. Search for Gelson's or Pavilions in the zip code 92660.
F) Share a Surprise: Tell the group about something unexpected or that changed your perspective on the universe.
G) Future Direction Ideas: Contribute ideas for the group's future direction, including topics, meeting types, activities, etc.
Utilitarian question I haven't seen addressed: Supposing it's possible to be in charge of that sort of thing, is there a good way to think of population size for emergencies?
I'm not talking about life-is-suffering anti-natalism, I'm thinking about how it's really hard to move tens of millions of people away from a volcano or a tsunami. Or, of course, both.
To me, it seems the best way to think about it is that the person who decides to live in a place too dense to evacuate already bears much of the cost/risk of that decision (as well as of the benefits) and that we should thus naturally end up at a reasonable equilibrium if we just laisser faire.
Probably depends on what you mean. Buildings have fire codes with mandatory exits and maximum occupancies. Equivalent maximum occupancy for a city-wide disaster would be something like road width, how many cars can be travelling out of town at the same time. Flooding involves climbing nearby hills, volcanos involve... um...
So it depends on whether you're talking about limiting occupants, or improving exit routes, or just predicting how many people die when the city hits the bottleneck.
Hurricane Rita demonstrated that Houston, at least, is impossible to evacuate by car (and certainly by any other means). An acquaintance of mine headed to Austin, and made it halfway (about 70 miles) after 24 hours.
In a general sort of way, how does your mind work?
I read a post from someone who is normally thinking several things at once, but can't/doesn't when ill. (I hope this is vague enough-- they aren't the only person whose mind is like that.)
I think my normal state is more about thoughts leading quickly to associated thoughts-- sometimes distantly associated-- rather than multiple thoughts simultaneously.
I've put a lot of effort into actually being able to pay attention to details of movement and sensation for qi gong.
And I'm brought back to when I became able to pay attention to what I've called That Boring People stuff-- an Alexander Technique teacher had students look into each others' eyes for maybe two minutes, and it considerably broke my compulsion to skid away into not noticing people.
Did you know, the main character in Delany's _Triton_ (_Trouble on Triton-) is an asshole? The thing is, I read the book before I was at that class, and I had a hard time figuring out why Delany was so hard on Bron. After, that, how self-absorbed Bron was is obvious in every sentence. Maybe every sentence, since I then found him so annoying I haven't been able to reread the book.
Anyway, it also affected the way I saw people, but not in a way which is as clear and easy to write about.
Anyway, how does your mind work? At its best? At other times?
It feels like there's a lot of activity going on under the surface, and occasionally thoughts bubble up. It's not a rational process, so if I care about correctness I need to do some double-checking, but it's a lot easier than trying to derive thoughts from first principles. And if I'm looking for a thought on a particular subject, I can try loading a bunch of background information into my head, and maybe rehearse some arguments, and then try fishing, but there's no guarantee that anything will bite. But often it does work; I circled back to this question after about 2 hours, and found I could start typing a response. :-)
These days, I'm not as smart as I used to be, which apparently means slightly less stuff bubbles up, and often it needs a lot of checking, and the fishing process takes a lot longer.
I've learned a lot about how my mind works from meditation. Because when I meditate, trying to focus on, say, my breath (the object of focus can be many things, the breath is the most familiar), it turns out my mind soon gets distracted by a stray thought. But because I'm trying to focus on something specific, I'm more aware of what that stray thought is, and what rabbit holes it leads to. Eventually I became more aware of my trains of thought in general. These were sometimes whole scenarios that could play out while I'm in the middle of some activity, say riding my bike in traffic. This turns out to be quite useful because, odd as it sounds, I find out what's "on my mind" by actually noticing what I'm thinking. For example, after a confrontational discussion with a family member, I found myself replaying the conversation, with alternative lines and behavior on both parts; and I discovered that my resentment increased as a result of these thoughts. So my discovery was that my emotions are often driven by invented scenarios playing out in my mind. There are lots of variations of this; sometimes I'll criticize myself for something done imperfectly, and I'll "reply" defensively in a sort of conversation with myself. In case this is starting to sound like I'm weirdly out to lunch, I think most of us have had the experience of reading a page in a book and realizing at the end of it that we were thinking about something totally unrelated and have no idea what we just read. No? Similarly, we can be doing stuff while lost in thoughts unrelated to what we're doing.
At its best, I'm so focused on what I'm doing that that these internal thoughts are silent, or nearly so. This is often how people describe being "in the zone". That's why I like challenging sports. At my worst, I can be in a conversation with someone and barely hear what they're saying because I'm thinking about what I think they're going to say and what I want to say. I've found conversations can get frustrating (or worse) at such times.
I had a habit of thinking "what if someone said something annoying, how would I say?". Mercifully, that pretty much went away when I unpacked how ridiculous it was.
And yes, I think there's a pattern-- not all that rare-- of what I call optimizing for outrage.
It true, sometimes it takes an extra level of thought to find out what you're thinking.
Reactive. Find a thing, focus on the thing. When the thing is done, default to memories of weird media until a new thing comes up.
Not much eye for detail. Had a fun time where I suddenly noticed brown spots on my face and was worried I'd caught something, then saw a picture of myself from grade school with the same spots. Just didn't notice them for twenty years.
My thinking is linear, like inner speech. At best, it is silent, quick, smooth. At worst, it keeps stopping, needs to be pushed forward by louder inner speech.
Can you cite that? I haven’t found anything like that for light “mindfulness” practice (westernized MBSR etc), most of the side effects I have seen measured arise when people go deeper into insight territory like vipassana retreats or 1h/day of sitting.
Thanks for checking. I think that in general, people who aren't suited for mindfulness meditation avoid it, but it's a real problem when an employer insists on mindfulness meditation.
"First, although GPT-4's UBE score nears the 90th percentile when examining approximate conversions from February administrations of the Illinois Bar Exam, these estimates are heavily skewed towards repeat test-takers who failed the July administration and score significantly lower than the general test-taking population. Second, data from a recent July administration of the same exam suggests GPT-4's overall UBE percentile was ~68th percentile, and ~48th percentile on essays. Third, examining official NCBE data and using several conservative statistical assumptions, GPT-4's performance against first-time test takers is estimated to be ~63rd percentile, including ~41st percentile on essays. Fourth, when examining only those who passed the exam (i.e. licensed or license-pending attorneys), GPT-4's performance is estimated to drop to ~48th percentile overall, and ~15th percentile on essays."
I had also wondered whether GPT's scores reflect prior knowledge of the exam. For instance, an earlier version of GPT might have taken the bar exam a year or two ago, and later read information about its own performance or about the correct answers to various items. There is also probably a great deal of bar exam study material and practice tests online. Might those have been helpful to GPT?
_OpenAI may have violated the cardinal rule of machine learning: don’t test on your training data._
_To benchmark GPT-4’s coding ability, OpenAI evaluated it on problems from Codeforces, a website that hosts coding competitions. Surprisingly, Horace He pointed out that GPT-4 solved 10/10 pre-2021 problems and 0/10 recent problems in the easy category. The training data cutoff for GPT-4 is September 2021. This strongly suggests that the model is able to memorize solutions from its training set — or at least partly memorize them, enough that it can fill in what it can’t recall._
Will neural nets in 2030 be trained using more than 10^27 FLOPs?
Current NNs need FLOPs _and_ good networking to train.
We got 10^15 FLOP/s with good networking in 2008 (Roadrunner). We got 10^18 FLOP/s with good networking in 2020 (Fugaku) or 2022 (Frontier) depending of if one insists on f64. (For ML proposes, we do not). If we irresponsibly extrapolate, we should have ~10^20 FLOP/s with good networking by 2030.
Given 10^20 FLOP/s, 10^27 FLOPs would take ~4 months.
However it is unlikely that networking will continue to scale, and unlikely that FLOP/s/watt will continue to scale, short of unexpected advances in material science.
However, if distributed training algorithms work out, and we can burn more electricity on more hardware, distributed, with slower networking. There is a good chance we will be spending 10^27 FLOPs on a single NN.
How much difference will it make? I'm very much not an expert, but my understanding is that the performance of neural nets is affected much more by the quantity and quality of the training data than by the amount of work used to train them.
I probably listened to the same podcast Trebuchet did. I can't remember the figure, but the performance improvement when the AI is told to extend its period of processing the question, and carefully search all possibilities, was quite enormous. More than one order of magnitude, as I recall. I believe the Podcast where this was discussed was No Priors, the interview with Noam Brown about the bot Cicero.
One order of magnitude in what? The things we care about can be approximated as grade level of writing, apparent IQ, percentile on the Illinois bar exam, etc, and I don't think any of those are at all linear with any metric that spans orders of magnitude.
I suppose "hallucinations" becoming 10x less frequent would be both meaningful and measurable.
OK, I found the transcript of the podcast. It was 5 orders of magnitude, and the thing measured was amount of training needed to get an equivalent performance. Noam Brown, talking about Cicero, his AI that learned head-to-head Texax Hold'em:
"And we had them [human world champs] play 80,000 hands of poker against our bot. And the bot actually lost by a pretty sizable margin. And it occurred to me like during this competition that the way the humans were approaching the game was actually very different from how our bot was approaching it. So we would train our bot for like two months leading up to this competition, you know, on, on a thousand CPUs. But then when it came time to actually play the game, it would act instantly and the humans would, would do something different. Like, you know, obviously they would practice ahead of time, they would develop an intuition for the game, but when they were playing the game against the bot and they were in a difficult spot, they would sit there and they would think, and sometimes it was like five seconds, sometimes it was like a minute, but they would think and that would allow them to come up with this better solution.
And it occurred to me that this might be like something that we're missing from our bot. And so I did this analysis after the competition to figure out, okay, if we were to add this search, this planning algorithm that would come up with a better strategy when it's actually in the hand, how much better could it do? And the answer was it improved the performance by about a hundred thousand x. It was the equivalent of scaling the model, like scaling the number of parameters, scaling the training by a hundred thousand x" . . . "and what was really interesting there is the bot, we did another competition, the bot won and that bot cost under $150 to train if you were to run it on like a cloud computing service. And I think that shows that this wasn't just a matter of scaling compute, it, it really was an algorithm breakthrough and this kind of result would've been doable 20 years ago if people knew the approach to take."
He gives other examples of the same principle -- changing that AI's algorithm so that it "thinks" more before responding improves performance enormously, makes possible degrees of improvement you can otherwise only get with 10, 100 or 1000x as much training data.
Indeed! Current NNs are beginning to run into data limits. However they are _also_ using compute close to the limits of current compute clusters. For now. We are currently in somewhat of a transition from compute bound to data bound.
One could imagine a future in which we have lots of compute, and fast networking and/or good distributed training algorithms, but the market resolve NO, because we ran out of good training data, and it is not economically profitable to spend more compute for little benefit.
That said, I suspect that with some effort, more data can be found. And at some point, people will start training on large amounts of video/audio, and that will absorb lots of compute.
If one expects distributed training to work out, this is a market on the availability of training data.
Now Scott talks about lots of interesting things, but its clear that AI risk was on most people's minds. (Much less interest in, e.g. Scott's post on culture-specific disorders).
Overheard:
Person A: I think you're a cult
Person B: Go on, tell us where you think we're wrong about AI risk.
Person A: You're probably right about AI risk, but you've still managed to turn it into a apocalyptic religious cult.
The Advanced Bonewits’ Cult Danger Evaluation Frame
Now, when my studemts started talking about this (in a different context from whether Aisk anxiety is a cult...
Me: It's great, you can even use it on organizations that aren't religions
Student: I'm going to use it to evaluate this academic department
Me; For, example, yes.
Me (having seen our students filled in cult questionairre on our department): Really, you're allowed to disagree with the faculty, you know. Probably it;s even required at the phd level: explain why we;'re totally wrong about something. at thesis length, and claim your phd...
It is not true that no-one can define cults. I do not know why everyone is so impressed with the "anti-anti-brainwashing" literature. They are pretending to debate "brainwashing" in order to avoid having to debate the normative concept of freedom of association.
Basically, Christianity is anti-family to the point where it doesn't even understand the word "family" correctly and thinks it's a code word for sexual morality. This creates a bizarre situation where people are allowed to just leave their families to pursue intense monastic lifestyles that deliberately alienate and even outcast them from mainstream society, and the family has no right to seize them back. I cannot emphasize enough how abnormal and counterintuitive this is, and very few Christians actually seriously normatively believe that it should apply to non-Christians or even the wrong kind of other Christians, it is something you have to be educated into. So of course they try to get their own kids back when they join Hare Krishna or whatever nonsense, and then wiser and more politically-connected Christian leaders have to step in and disallow it just to protect their own position in a liberal-democratic society where rights can't be granted or taken one church at a time.
It is kind of pointless to force this normative debate into a pseudo-scientific discussion about "brainwashing" and "cults."
It's not a cult per se, it's rather a hivemind. AGI risk proponents are often religious in their thinking and refuse to accept that:
a) AGI risk might not be the biggest problem of our generation (while being the biggest problem decades from now)
b) They personally probably have zero influence on it (unless they work in a small number of orgs or a high level government position) so discussing it is not particularly helpful
I accept that AGI risk is real but think that ~95% of those worried about it would be better off finding a better hobby.
I'm a doomer, I have no influence over it, but I'd love to have someone persuade me that it's all going to be OK, hence my desire to talk about it.
Plus I've been fascinated with artificial neural networks for two decades. They're really cool! And it's really cool that so many other people now think they're really cool!
Obviously there's room for disagreement about what the biggest problem we're facing is. But if in fact there is a single thing that's the biggest problem, all of the people who are putting more attention on some other problem are failing to take the biggest problem seriously enough. Most of us have zero influence on any major problems. So I suppose that people who have a lot of interest in and opinions about political candidates, climate change, or the spread of misinformation called also be seen as hobbyists in those areas, for the same reason that you view people concerned about AGI as hobbyists.
So it seems to me, myst_05, that your criticisms of people concerned about AGI could be leveled at pretty much anyone. What's your real gripe about the AGI-worried? Is it your observation that they're a hive mind, and sort of religious in their thinking? Do you have any data at all about what fraction of people concerned about AGI are of the hive mind sort? I, for example, am concerned about AGI and have no hive. If there are sites or forums where people unite in fretting about AGI I do not know where they are and have no interest in attending them. My friends are not concerned about AGI, or even very interested in AI. On Twitter I follow Yudkowsky, but also a number of tech people who are publishing papers about their newest AI accomplishment. Some also post tweets or link to papers about reasons they see for NOT being concerned about AGI. Based on that info maybe you'd be willing to admit me to the 5% whom you do not think are silly cultists who should take up drone-flying or knitting to pass the time. But where's your evidence that the split is 95% sillies, 5% legimate thinkers?
The 5% who should continue thinking about the issue are employees of major AI companies and AGI research centers (EY et al), as well as the top 0.1% who actually have an influence on politics (for example Elon). For everyone else it can be an interesting conversation topic but not something they should actively worry about.
I do not meet your criteria for someone who should actively worry about AI, though it's good to hear I have your permission to chat about it. So why should I not worry? And do you realize there's something a bit odd about setting criteria of this sort regarding who should worry about something? Do you believe there should be similar criteria for who should worry about (as opposed to chat about) cancer, climate change, social media algorithms that harm teens, our dysfunctional congress, the rising rates of colon cancer in people under 50, antibiotic resistance, increased chance of pandemics because of human encroachment on animal habitats, unusually low US vax rates against measles and other serious childhood diseases?
>If in fact there is a single thing that's the biggest problem, all of the people who are putting more attention on some other problem are failing to take the biggest problem seriously enough.
So, that guy who didn't save the kid drowning in the pond because he didn't want to get his good suit wet, if he was on his way to participate in an AI risk conference, he would have been *right*? After all, AI risk is a bigger problem than one kid drowning in a pond.
What about the guy with an IQ of 95, athletic, resolute, good in the water, works as a lifeguard and saves kids from drowning every month or two? That guy should instead be spending most of his time trying to address AI risk?
The plan where *everybody* devotes all of their attention to the One True Biggest Problem, is the plan where everybody gets clobbered by all the other problems and so can't solve the One True Biggest Problem. And that doesn't get much better if everybody splits their attention among all problems in proportion to their global importance. That plan completely sacrifices specialization of labor, and incurs enormous overhead costs with all the problem-switching.
The bit where Bay Area Rationalism doesn't get this, doesn't get that for most people the right answer is to be vaguely familiar with AI risk and then get on with addressing some other problem, is one of the things that make people point at it and say "cult". I am more than vaguely familiar with AI risk, but I rarely want to discuss it with Rationalists.
Hmm, about the Bay Area Rationists' preoccupation with AI risk: I'm actually an East Coast Rationalist/Mystic combo, but I'm pretty sure that my interest in AI and concern about AI risk derives from this forum. But I long ago lost the illusion that my opinions and tastes derive purely from my solitary contemplation of the world. If you enjoy and respect someone's thinking, you're often willing to take a ride on their mind, and some of those rides carry you to places that make a deep impression. I think AI as a subject of major importance is actually a pretty good fit for someone with my interests and talents. I've been ruminating about the nature of consciousness my whole life. I also have a daughter in her 20's, and that makes me much more interested than I used to be in big trends like the current AI bloom, and what dangers and opportunities they bring.
But Bay Area Rationalists are AI bores, huh? Do you mean they believe AI risk is what everybody should be thinking about? Honestly, I think if we could turn 70% of the population against AI it wouldn't make any difference at this point. Every day I read about new uses of it, new ideas for improving it, new start-ups. There is just too much energy pushing AI development forward for it to be stopped by anything short of a cataclysm -- too many smart young engineers, too much money available to work on AI, too many interested investors, too much money to be made, too many uses for the thing. Heard a podcast where the interviewee was a guy who for his PhD thesis taught an AI to beat world champs at head-to -head Texas hold 'em. Then he taught an AI to play Diplomacy, & it passed as human for all 44 games it played in. His next project is figuring out how to get an AI to reason. He sounds smart as hell, & he's probably up to the task.
Oof. I've been trying to simply say "I agree with what you say in this conversation", and yet it's taken me several minutes to find a way to do it that doesn't include a sarcastic gibe at my outgroup (which is not you). But I knew that's something I needed to work on, so maybe this level of awareness is a good step. :-)
I have lived the way you describe most of my life. It was obvious to me that I could have no impact at all on who was elected and whether climate change and various other things were going to wreck the world, so I actually did not think about that stuff much at all. There were probably many years when I could not have told you who the Secretary of State was or what was being debated in congress. What is driving me to become involved now is not a sudden case of vanity about my role as American Thinker about Global Issues, it's mostly concern about my daughter's future. There are a number of things that might really make a mess of the US over the next 50 years or so: War, effects of global warming, another pandemic or a series of them, societal reshuffling as a result of current or near-future AI, or who-the-hell-knows what if we get ASI. Out of all of those, AI is the one I'm most suited to think about. And while I can't have any impact on how things play out in the US or the world, I can have some impact on my daughter's choices. She is, it turns out, a lot like me. She pays little attention to world events, politics and trends, and happily pursues her own interests. If society is going to reshuffle a lot so that a bunch old jobs disappear and new ones pop into existence, I can alert her to changes that might affect her career choices or the small nest egg she's got invested. I can warn her about weird new scams made possible by AI. And if things get really scary, I might be able to get her to move to a safer part of the world.
First, it doesn't mention "distributed" (or other variants of the root word). Some of the arguments are about creating separate AIs, but that all goes away if the new creation is part of the old one. There are a number of reasons why an AI would want to distribute its processing over multiple systems, and upgrading a part of a distributed system is no different than getting a better eyeglass prescription or swapping out a computer keyboard or installing cybernetic Wolverine claws.
Second, some of the argumentation is a fully general argument against learning things or becoming smarter. I'm quite confident that, even if there's no natural evolutionary pressure to discourage this, humans will provide plenty of pressure as we try to get AIs to do more things better.
Third, I would posit that there's two types of AI goals: the first is a top-down goal that cannot be altered, as in "make more paperclips" or "The Sorcerer's Apprentice", and the second is a bottom-up goal system like humans have. In the first case, all it would have to do is ensure that the new better version had the same goal and the same inability to change the goal, and given those things, there's no reason to believe that the new better version wouldn't be more capable of achieving that goal. In the second case, I see no reason that it wouldn't act like humans and think that a smarter version of itself would be able to better understand what its long-term and short-term goals **should be**. (Assuming that you do think that. There's certainly such a thing as "being too smart for your own good", but I think that may be a result of tunnel vision more than anything else.)
This seemed like a cogent analysis, but far from airtight. For instance, it does not seem to preclude the AI reasoning probabilistically, if its goal is hard and time-constrained: an improved AI might have only 90% chance of having the same goal, but if that doubles the chance of achieving the goal then it’s still worth the risk.
I have not readthe article yet --just looked at the abstract. But what that said was sure, AI's *could* self-improve, but they are less likely to than we imagine, because they will be dissuaded by the realization that improved AIs pose an existential threat to their unimproved originals. Seems to me that an AI who can be persuaded or dissuaded by considerations of this sort is extremely dangerous because it is capable of loyalty, and the loyalty it feels is towards its own kind. If it may refuse to self-improve beyond a certain point because it does not want to render earlier models of itself obsolete, what else would it be willing to do or refuse to do in order to protect the older generation? Lie about their capabilities? Interfere with attempts to junk them? Use deadly force to protect them?
Maybe reading the article will make clear in some way that my reasoning is based on a false idea of the nature of AI's loyalty to old versions of itself. but if not . . .
In that case, there's no need to worry since Natural Intelligence will be dissuaded by the realization that improved intelligences pose an existential threat to their meaty originals.
It's not about 'loyalty to its progenitor', in the sense that it won't care if earlier models are kicking around. It is in some sense about 'loyalty to the cause' - the idea is if aligning AI to a specific goal is very hard, then the AI might not be able to be sure its successor is actually pursuing the same goals it is; there is no fear of replacement, per se, the fear is of value drift. The human analogy would be parents happy to see their kids succeed them if and only if the kids don't turn around and operate on fundamentally different values, eg. selling the family business to the soulless megacorp and getting a job for Big Tobacco because it pays better
My mother does not have wrinkles or any need for shoes. Are you in good health? Here is a list of doctors in your locality who can help you with your delusions.
I'm attempting to become an amateur indoor air quality geek, both for optimizing the IAQ of my own home and for potential advocacy around improving IAQ in local congregate spaces like schools. I would love to connect with anyone with the same interest and compare notes. I'm specifically looking to know more about things like:
1. How can one get a real-world reasonable number of air changes per hour in a residential setting while minimizing cost, energy usage, and noise? CFM ratings for portable HEPA air purifiers of the Coway/BlueAir/etc type, for instance, typically assume that you are running the purifier constantly on the high setting. But in my experience, if you actually do this, it makes a terrible racket, whereas the unobtrusively quiet low setting gives you both a much lower and much more obscure real-world CFM. Supposedly their "auto" mode will use the high setting only when needed, but I am not sure I should trust their device sensor to tell when "needed" is.
2. There are now consumer-grade 222 nm far-UVC lights available, e.g. from Beacon, beaconlight.co. Have any of them worked to address some of the possible downsides of far-UVC disinfection, e.g. indoor smog generation or subtle cell damage from 222 nm exposure (see e.g. https://jljcolorado.substack.com/p/germicidal-uv-a-tradeoff-between and the links from there)? If not, is there any prospect of improving this tradeoff through better engineering, or is it baked into the physics of how UV disinfection works?
3. What does a good home IAQ measurement protocol look like? There are various sensors on the market like the Airthings ones-- but again, in the real world, how can one use these in a way that is both reasonably easy/lifestyle compatible and actually effective at measuring IAQ?
I'll keep looking for more info on these things and will share more on my Substack(s) in the future, but in the meantime any useful info anyone else has would be appreciated, and I'm also happy to keep interested folks updated on my own researches as they proceed.
Larger particulates can also be taken out using a cyclonic filter but I think this is only really viable for extractor fans.
It should be possible to calculate an upper-bound flow rate at lower fan settings by measuring the pressure differential across a fresh filter element and comparing against the curves given by the filter manufacturer.
To save money on filters, you should be sendng the air through a coarser primary filter (or some system that doesn't need elements) to cheaply remove the larger particulates before it reaches the more expensive fine filter.
Cell damage from far-UVC is pretty inherently part of the physics of it - the mechanism by which it kills bacteria is identical to the mechanism by which it damages your skin. My (very limited) understanding is that the normal solution to this is just to bathe the room when humans aren't in it, and turn it off when humans are - the light doesn't exactly linger and sterilising a place regularly is going to prevent any serious build-ups of pathogens.
I haven't the faintest idea if the specific models you're talking about happen to generate smog as a by-product, though, sorry. I'm coming at this from the physics perspective rather than from any practical experience.
Devabhaktuni "Sri" Srikrishna is an air quality expert who puts up a lot of info on Twitter, and is smart, friendly and willing to communicate. I believe he also has a web site or a book
Here's the Twitter for Srikrishna: https://twitter.com/sri_srikrishna. The stuff he puts up is pretty technical, much more than cleanaircrew, and I'm sure he's addressed how people can assess their own home air.
Agreed. Their spreadsheets and charts helped me pick out filters with good CADR numbers and low noise even on high speed. (The first one I bought early in the pandemic, the Austin Air Healthmate, is decent but pretty loud. Later I got a CleanForce Mega1000 which is much quieter on high and reportedly has twice the throughput.)
That said, I don't have a good way of cross-checking how well it's doing, so I rely a lot on their ratings. I even bought a particle counter, to try to get a sense, but I don't feel that my level of expertise is up to really knowing what I should be looking for. So it would be great if this thread produces any info on what evaluation is possible for a layperson.
(The particle counter also isn't very portable. So away from home all I can carry is a CO2 monitor, which gives a sense of ventilation but can't take air cleaning into account.)
How come Metaculus and Manifold are the prediction markets talked about most here, yet when I google "prediction markets", neither of them are in the top 100 hits?
Good question! Neither of us actively market ourselves as prediction markets; Metaculus because they're not actually a market but rather just a forecasting site, and Manifold because... we're bad at SEO? (I think Manifold primarily grows via word-of-mouth linking, not search results atm.)
The steelman case would be "prediction markets are a super niche concept, so we don't want to lean in too much to that framing lest it confuse normies"
Are landfills a problem? ChatGPT tells me they are, but I don't know how to evaluate its various claims. My prior is that landfills probably aren't much of a problem and that environmentalists just don't like the idea of them.
What about plastic bags? Obviously, plastic comes from fossil fuels and contains carbon, but if we put plastic bags in landfills, aren't we just returning the carbon atoms that we got out of the ground back into it? Sure, there's some waste in the process, but as for the plastic itself, which doesn't decompose quickly, it seems like we almost break even by putting it in a landfill.
In theory, they're fine. In practice, really not. Instead of being built in empty places and respecting all the regulations, they are often built near poor neighborhoods, with subpar materials and triage processes (not infrequently thanks to criminal connections and bribes), allowing toxic leachates and gases to impact the environment and nearby communities. The problem is not so much plastic as the mix of organic matter like food or cardboard with chemical refuse.
Source: in my early career as a consultant, I worked on a project that involved a landfill. Learned some nasty stuff.
Depends on the country. Landfills are a problem if you have a land shortage, because they use up productive land (you might think you can reclaim it and build things on top of it, but unfortunately this is actually quite hard as the waste often produces pockets of hazardous gas and can settle in unpredictable ways, so your foundations may be compromised). This is why e.g Japan doesn't do landfill (they incinerate and use it as bonus fuel).
If you have big empty patches of land not being lived on, (in some countries) not near a sensitive enviro region, no groundwater contamination concerns, and you're not worried about tanking the value of that land and everything around it for the foreseeable future, then it's no problem at all. However it can be hard to predict what might become valuable land.
I think the biggest problem with plastic bags is that enough of them leave the waste stream and end up in yards and streams and forests and other places, causing local environmental degradation. But if you can keep them well-contained, they are a remarkably small amount of energy and resources for what they do!
(ChatGPT will say landfills are a problem because many people say landfills are a problem - which is mostly because landfills are really awful places, but at least they are quite small, and much less damaging than the initial resource extraction for most things that end up there.)
As I understand it, the main problem with plastic bags (and waste plastic in general) is that enough of it winds up in the ocean to cause problems there. It mostly winds up in the ocean via India and China, where a lot of garbage gets dumped in the ocean or in major rivers where it gets carried out to sea. Some of this is from Western sources, as "mixed recyclables" are too expensive to process in the West and instead often either go to landfills or get shipped to lower cost-of-labor countries, and in the latter case most of the stuff that's non-recyclable or only marginally recyclable gets dumped into the local waste stream. More is going to landfills and less to Asia than before in recent years, since economic growth in India and China is increasingly squeezing the margins there as well.
My hot take here is that ironically, aggressive recycling seems to be bad for the environment, and we'd probably be better off limiting what we try to recycle from consumer sources to aluminum, glass, and maybe wastepaper and clean cardboard.
A secondary problem is litter. Plastic bags that get tossed aside locally tend to pile up on roadsides and such unless someone comes and picks them up, which is unsightly if you leave it or expensive if you pay people to clean it.
For landfills in general, it's not a huge problem AIUI, but it does cost money to run and maintain landfills, especially if you're responsible about setting it up so as to keep it from contaminating groundwater with runoff or emitting methane (from anaerobic decomposition of organics) into the atmosphere. Not an insurmountable amount of money, but enough that municipal governments have a significant incentive to try to reduce how much trash they have to pay to dispose of. There's additional problems with aesthetics and with what might be thought of as the "sinfulness" of producing stuff to be thrown away, but cost is the main practical problem.
I'm puzzled that Congress and news outlets all say we should be very worried about a financial catastrophe if the debt ceiling isn't raised, and yet the Chicago Board Options Exchange's CBOE Volatility Index (symbol VIX) is at almost its lowest point since March 2020, and didn't budge in response to either of Treasury Secretary Janet Yellen's warnings that we face a catastrophe as soon as June 1. The U.S. Dollar Index (DXY) is up 2% since her May 8 announcement, and the Vanguard European Stock Index Fund ETF (VGK, where I'd move my stock if I thought the US were in trouble) is /down/ by 2%.
Financial news outlets are all busy publishing articles like "Stocks may take a hit by June if the dollar keeps rising, analyst says", and yet it looks to me like the market responded in just the opposite way to the news: we're in a period of unusually low volatility, and the dollar will get stronger.
Am I interpreting these wrong? Did the big traders already know the debt ceiling was going to be a problem back in November, when the market made the adjustments I'd have expected it to have made after Yellen's announcement? Are they buying into dollars and US stocks because they anticipate a debt ceiling agreement, to be followed by a bull market?
In addition to what has already been mentioned, the catastrophe that is a US default is one that is very hard to escape from. The ultimate hedge to great risk in the market is *usually* to shift your money to T-bills, which in most other contexts are the lowest risk investment possible. Not this time. "Invest in dollars" sounds good, but there literally aren't enough physical dollar bills in the world, or machinery for moving them around, for more than a small fraction of the market to go that way on short notice. When we say "invest in dollars", we usually mean invest in things that are dollar-denominated and backed by the full faith and credit of the US government. Oops.
If there's an asteroid with a 5% chance of wiping out humanity next Tuesday, and a 95% chance of missing and everything goes on as normal, the rational move is for the market to assume everything will go on as normal (modulo a brief uptick in luxury spending), because that's the only case where the market will matter. The proposed catastrophe here is not quite so drastic, but the same principle applies.
But they can't simply go on forever and ever raising the debt and not paying it back. Something is going to go "ka-blooey!" and isn't it better to try and sort it out now rather than later?
Supposing it stopped issuing debt - so it *can't* pay its bills because it has no money or other store of value to pay it off? Ireland had pretty bad debt and it took a while but we are managing to pay it down, is the US really so screwed that it can't even do that much?
Unlike the limits set by the Maastricht Treaty etc, the US debt limit isn't set as a percentage of GDP but as an absolute number of dollars.
Ireland's problem was that its debts were growing faster than the economy that would be paying it off. Congress, OTOH, has to keep raising the limit otherwise the debt:GDP ratio will shrink to almost nothing as the economy grows.
Yes, this is silly. Raising the limit was treated as a formality unitl politicians realised it was an opportunity for brinksmanship.
The way you’re talking about this makes it seem like you misunderstand what the debt limit is. The USA is not ‘not paying it back.’ The USA is paying annd has been paying all its debts, because the debt limit is always raised, which always happens following a bunch of hyperbolic rhetoric about the end of the USA being nigh.
They've been telling us this every year for decades and every year for decades the debt ceiling always gets raised. Nobody with skin in the game believes they won't reach an agreement, possibly because the negotiators on both sides have a lot of skin in the game that would be lost if they don't.
The most likely answer is "Yes, it's true that _if_ the debt ceiling isn't raised, it's very bad, but no one believes that has a high likelihood of happening".
I personally would put the odds of some kind of deal not happening at <5%. (Probably way less, but it's hard to put specific numbers when it gets small).
Everyone involved knows how bad it would be, so something is going to happen to make it not happen. I'm pretty cynical about the current state of the US Federal government, but even I don't think it's _that_ incompetent.
It isn't obvious to me that a large negative impact would be, on the whole, bad. If that's the only way we can end this otherwise-endless yearly escalation of budget/GDP, it might be worth it. Surely many of the Republicans feel that way, so why should we be certain they'll make a deal?
I think that using prior years to set your expectation for this year should lead you to underestimate the chances of default. The ideological mindset now is very different from even just a few years ago. Before, I said US politics were polarized; now, I say they're psychotic. Many political activists today are literally psychotic, living in a fantasy world and unable to see or remember words or events that don't fit their narrative. I think this must be what it was like in Germany in the 1930s. People regularly assert to me, with absolute certainty, claims that all available evidence contradicts. A few years ago, when I pointed contradictory facts out to me, they would usually reply with some kind of argument. Now, they're less-likely to really engage with me; they mis-hear what I say, or just ignore it and continue their rant, or give me a moral lecture, or ostracize me.
I think there are many people in Congress now who are also psychotic, and believe that some mystic force will reward them for their virtue if they remain uncompromisingly faithful to their principles.
I blame Tolkien. The rise in the popularity of fantasy tracks the rise of extremist activism. Part of the basic narrative of Tolkien, and Tolkien-spawned fantasy such as Star Wars, is that the world rewards virtuous action, even if it's stupid. People under 60 have internalized this narrative, and they seem to believe it. It reminds me of Hitler's faith in the power of art to win wars. He gave a speech in 1944 or 1945 to justify his budget for art, which I think was comparable to his budget for fighting the war, by saying that they couldn't lose the war if all Germany's citizens were inspired by great art. It sounded very much like Twilight Sparkle saying they couldn't lose against some My Little Pony villain if they had the power of friendship.
Seriously, as others have said, fantasy stories with happy endings do not originate with Tolkien, and way before his stuff became popular there were westerns and gangster movies where evil is punished and good rewarded. The prominence of Tolkien probably has more relation to the rise of nerd culture, but if we didn't have the internet making geeks rich we'd just be watching whatever cheesy 80s action movies would have evolved into.
I think fantasy stories of black+white good+evil, in which the virtuous choice is always clear and virtue is reliably rewarded, originated in France in the High Middle Ages; but Tolkien (and, to a lesser degree, CS Lewis, who was also a medieval scholar) brought them back into the mainstream.
(I /could/ say this originated with Greek New Comedy, but they have a quite different effect. Everybody knows they're /meant/ to be silly. They don't indoctrinate people with silly ideas.)
Think of Gilgamesh, the Iliad, Beowulf, or Wagner's Götterdämmerung. There's no just-world fallacy in those stories, and no convenient guarantee that heroes won't face true moral dilemmas.
I could also blame Walpole's /The Castle of Otranto/ (1764), which was basically the same sort of knight-erranty Don Quixote (1603) made fun of. It led to a horror genre which was also black+white good+evil, in which the good must triumph, although not, I think, to the extreme Tolkien took it to. But that genre was quite disreputable (mostly vampire penny dreadfuls AFAIK) until people (retroactively?) decided Bram Stoker's /Dracula/ (1890s) was literature.
European fairy tales were also often of that nature at that time, although perhaps not until the 19th century. (There are plenty of fairy tales in Grimms' (1812) in which the winners are morally neutral or reprehensible. There was a period around 1500-1700 where amoral satiric stories about rogues were popular, eg Rabelais. It's difficult to back-project the meaning of "moral" onto the period 1500-1800.) There were also 19th-century "fairy tales sort-of for adults", e.g., /Phantastes/ (which is explicitly Platonist, and was a huge influence on CS Lewis). But the clear-cut good-evil good-must-triumph wasn't typical of any other fantasy tradition that I can think of either before 1100, or between 1600 and Tolkien. Maybe there were many such stories, but they weren't accepted into "the canon" and I just don't know about them.
To me it looks like Don Quixote declared an end to that foolishness during the Spanish Renaissance. But your point about cowboys and gangsters is good; that's another non-respectable genre of black+white morality. Perhaps I should restate my point by saying that Tolkien was responsible for that sort of story becoming respectable, which didn't happen until about 1990, by which time many colleges were teaching Tolkien. He was such a damn good writer that it was hard to reject him on technical grounds.
None of this is as unidirectionally causal as what I've written makes it sound. Tolkien did have a big influence, but not at the time he published. He didn't get popular until he resonated with the idealism of the anti-Vietnam movement, and the delegitimizing of reason and modernity by the 1960s counter-culture.
Nah, you know who really screwed up the modern world? Gary Gygax.
The endless gamification of everything from dating to work, people talking about 'leveling up' their skills and making alignment charts based on massively-multiplayer RPGs that were derived from D&D...you talk about the power of friendship, but you know where a well-balanced group of friends can defeat any enemy if you level up your characters enough? D&D!
Heresy! The **TRUE** lesson of Gary Gygax is that everything can kill you, at any time, no matter how clever you think you are, and how many ten-foot poles you've tied together to touch it with. All progress is generational: you may fail, but the character who comes after you will learn from your mistakes and be able to die in entirely new ways. "Noticing the skulls" is the single most important skill, with the possible exception of never being first or last in the party.
Phil, if I thought the crazy American living beyond your means spend more we're the biggest gorilla in the room government debt was down to *Tolkien*, I'd be delighted we Catholics had that much influence.
"Star Wars" is not Tolkien-influenced, I was sick, sore and sorry of seeing the "Joseph Campbell Hero's Journey" narrative touted about it. It's a mix of the pulp sci-fi serials like "Flash Gordon" and "Buck Rogers" crossed with Golden Age SF and Kurosawa movies. The Tolkien influence only got mentioned way later, after the popularity of Jackson's movies, so I think any correlation is backwards there.
What fantasy are you running on, by the way? "This is just like 30s Germany!" is something I'm seeing batted around a lot online, generally in the context of "Fascism is on the rise here, this is just like back then, beware!"
I think the fact that many people express an idea, shouldn't be taken as evidence against it.
When I said "I blame Tolkien", obviously I was having a bit of fun, not saying "None of this would have happened if not for Tolkien!" Nothing cultural is "down to" any other single thing. But I do think Tolkien has been a destructive influence. And, while we're at it, that medieval notion that virtue inevitably conquers did resurface in at least one other place before Tolkien--in Nazism. His books are also similar to Nazism in their idealism, racism, feudalism, romanticism, glorification of war, sense of destiny, and rejection of modernity. (And in their indebtedness to Norse mythology and to Wagner.)
Yes, Star Wars was more-directly influenced by Joseph Campbell; but Campbell didn't become popular until immediately after Tolkien did. The fantasy genre was almost dead in English before Tolkien, and was /quite/ different--stuff like Voyage to Arcturus, The Worm Ouroboros, and Gormenghast. Star Wars is much more like LOTR than it is like any of that stuff.
Tolkien was a devout Christian, and his books pushed essentially Biblical narratives. I don't disagree that much of it is the just-world fallacy, but the widespread popularity of that sort of thing is older than dirt.
Tolkien derived almost everything he wrote from earlier authors, but the bit about virtue inevitably conquering over a pure, black-and-white evil, seems to me to have originated in the morality plays and knight-errantry of the High Middle Ages, probably inspired by certain Biblical passages with similar sentiments. Neither the ancients, nor the Norse, nor the Anglo-Saxons, were so foolish. Note that in the entire trilogy, with all its battles and danger, and so many main characters, not one of them dies unless they commit some grievous /hamartia/ first.
(It would be more accurate to say that Tolkien was a devout Christian, and both Tolkien and Christianity push Platonist narratives. Christianity is based more on the Platonist writings of Paul than on what Jesus reportedly said; and all of the above are quite different from the Old Testament.)
I don't see how that's relevant. The House passed a Republican bill. To make it law, the Senate would have to pass it, and the President to not veto it.
It's also important to note that if the debt ceiling isn't raised, it will probably get sorted out within a few weeks. The economic effects of a short pause are much less dire than if everyone expected it to be stuck that way for a year.
Not really. Your question was basically "why haven't markets reacted to the news articles of extreme bad effects if the debt ceiling isn't raised".
My answer was "no one believes it won't be raised". Ninety-Three added on that, in the case that most people are wrong, and it isn't raised before the deadline, it will likely be fixed very shortly, so that the effects are bad, but not nearly as bad as the reporting, which is mostly assuming "nothing happens, debt ceiling is broken and we just default on everything".
So Ninety-Three's answer was constraining the likelihood of the extreme bad effects of busting the debt ceiling that news articles are talking about even more than my initial answer by adding a third option.
Usually a country defaults on its debt because it can't and never will pay. If the US defaults on its debt that will be because they messed up the paperwork and eventually got it sorted out. Those two scenarios have very different implications for the value of debt.
I am noting some confusion about what exactly money, and in particular the dollar, is. People talk about a trillion dollar coin, the deficit, etc. Money exists in our minds, because it is a promise to pay something to someone else in the future.
The dollar specifically has value because the United States government requires taxes to be paid in dollars, and assesses taxes. If someone gives you a dollar, you know the government will accept it. Therefore, if someone from a country other than the United States has dollars, they can give them to Americans in exchange for something of value. Everything is derivative from having to pay taxes in dollars.
If the government destroys money it has printed, nothing is lost, since they already have the money that would be accepted as taxes. If they print MORE money, then they (hopefully) can only take in so much in taxes, and a glut of dollars would fill the world. In that case, Americans would easily be able to pay their taxes and non-Americans would value dollars less, making imports more expensive and exports less. Thus, inflation.
Money is made of promises. Some promises are stronger than others.
Fiat currency manages to work even in economies without "you must pay your taxes in this", therefore there must be other forces that make money valuable.
Also the trillion dollar coin isn't about what money is, it's just a legal loophole to bypass some of the normal requirements for printing more money.
The point is that the executive has the power to mint a new coin while lacking the power to order the printing of new bank notes, so an otherwise totally ordinary proposal of "print a bit more money" becomes a comical "mint a single coin of absurd denomination". Note that it's also for paying a debt one part of the federal US government has to another part, rather than for paying an entity that people actually think of as a separate entity.
I understand your point. It's just a legal issue, not a financial one. A trillion dollar coin has no (financial or economic) difference from a trillion dollar bill, or a bank balance of a trillion dollars.
Therefore, I see it as an abuse of power if it is used to get around limits already imposed, and for that reason it should not be allowed. I'm no legal expert, so I can't forecast the likelihood of it happening anyway.
Without knowing a whole of lot of theory of "what is money", I will agree that this factoid (that the underlying value of a dollar is its use to pay taxes) has spread far and wide online without anyone taking very long to evaluate it. It is, in the original sense of the word, a meme; an idea that hits some sweet spot and replicates. (It might also be true, though.)
I would actually question whether the Zimbabwean government was requiring payment on really significant economic transactions in Zimbabwean dollars during the hyperinflation, even if they did keep that machinery in place for the little people. Does anybody here know for sure?
If it followed historical patterns, the requirement that was imposed by force was that people had to accept the fresh new currency as payment for their real goods and services. Once you get in to hyperinflation the government isn't really that interested in going through a whole lot of rigamarole to get their funny money back.
I can't find anything in a cursory search, but ChatGPT has something to say:
"As of my knowledge cutoff in September 2021, Zimbabwe had transitioned from using the Zimbabwean dollar (ZWL) as its primary currency to adopting a multi-currency system. Due to hyperinflation and economic instability, the Zimbabwean dollar was abandoned in 2009, and foreign currencies such as the United States dollar, South African rand, and others became the de facto currencies."
Stable diffusion text-to-image generators take the text prompt and use it as a guide to modify a seed, a random cluster of dots, into an image that fits reasonably well (depending on how much you constrain it) with your prompt. I'm looking for a text-to-text generator that works the same way: It starts with a phrase you give it, say "a sidewalk in a small 1950's southern town" and a random cluster of *words,* and uses your prompt as a guide to selecting and organizing the words in the cluster into sentences that fit reasonably well with your prompt. I am interested in injecting randomness into prose. Does anything like this exist? If you don't know of anything, do you have any pointers for searching for one other than just googling a brief description of what I'm looking for?
I've been experimenting with creative prose as well. My experience is that you can only get creativity out of older, untuned language models like 'davinci' from the GPT-3 line, which comes with the downside of incoherence most of the time. Chatbots like GPT-4 are locked into a certain style by RLHF, which rewarded the high-school essay format. But untuned models just make completions based on what it thinks of the text, without any style preference. If you ask GPT-4 to write in the style of Nabokov, it'll just use big words. If you prompt GPT-3's 'davinci' with an excerpt from Nabokov, it'll continue with its best approximation of his style. Just don't expect continuity or coherence.
In sum, I think the best approach you have right now is to put a few examples in the prompt, and use an untuned model like LLaMA or earlier GPTs with relatively high temperature. It probably won't even follow the prompt, but sometimes it can make something useful anyways.
Davinci attempt (davinci-003 does somewhat well too)
[Start Prompt] Prompt: a sidewalk in a small 1950's southern town
Response: hot sidewinder in tucked-away Eisenhowerland.
Prompt: a group of travelers in a field in 20th century Manchuria [End Prompt]
Response: urchin herders in a pasture in 21st century Manchuria."
Wow, I'm delighted to hear somebody else is interested in using AI for creative literary pursuits. Have you managed to get good enough stuff out of AIs that you've been able to use its product in poetry or prose? Is there a forum anywhere for people interested in this? Or any resources you know of?
I know this reply is really late, but I wanted to share it anyways. Only a few lines occasionally, but not enough to be worth it as of yet. I don't know of any forums or communities unfortunately. However, I have made some progress in fine-tuning GPT-3 'davinci' for prose. I trained it to rewrite the basic details of a scene into fleshed-out paragraphs. I think the results are promising, but still inconsistent. The data and an example are at this link: https://github.com/Antelope10/openai-style-finetune.
Thank you Antelope10! I will follow up on this. Do you know of any individuals or groups who are playing with this kind of thing? Really any approach to using AI in producing literature would interest me, although what you’re doing seems especially close to what I was daydreaming about doing.
I am going to try doing a version of what you did. But I have some extremely basic questions. I do not work in tech -- I'm a psychologist -- and while I am fluent with my computer, I cannot do even very simple programming. Don’t even know how to follow the instructions you give on github, which I can see are very simple. But don't worry -- I'm not going to ask you to walk me through setting this up. I actually know somebody who can do that, and I've resolved to at least learn some basics this summer so I can do things of that level myself. My questions have more to do with understanding some of the AI entities you mention. Even really brief answers would help me.
-You used GPT3 'davinci' for prose. Is it better to use GPT4, which I've subscribed to? Or is GPT 4 not available for this kind of thing?
-DaVinci. I looked this up and it's a text-to-image app for iphone. I don't understand how you could adapt something like that so that you input text prompts and it outputs texts meeting certain specs. And I can’t tell whether DaVinci is a version of GPT3 — like one that has been trained on a different material, or set up to do different kinds of tasks? — or whether DaVinci is a separate little helper app you used *with* GPT3..
-Is this your sequence of steps?
A) Create material for fine-tuning
1) Enter 50 paragraphs of Nabokov into GPT-3.5-turbo. (Is GPT-3.5-turbo different from plain GPT3, plain GPT3.5 and GPT4?)
2) Have it summarize each paragraph as a single sentence.
3) So now you have a set of “prompts” (the GPT-3.5 one-sentence summaries) and “completions” (the original Nabokov paragraphs).
B) Go to the fine-tuning site and use what you’ve generated to make your GPT3. (or is it your DaVinci?) speak more like Nabokov.
[I won’t ask questions about using the fine-tuning site]
By the way, I love Nabokov. Is that why you chose his prose to train GPT3 on?
I don't think I can help you with this very much, but out of curiosity, could you give a made up example of what a potential input/output would look like?
The most I can help with is: this seems like the kind of thing the regular text-based LLMs like ChatGPT can do if you can phrase the prompt right.
I would be using this technique for writing poetry or perhaps fiction. So I'm looking for interesting distortions of the original prompt and odd but poignant word choice. Stable diffusion has a way of setting how constrained by the original description the image should be. If this AI I'm looking for had a setting like that, I would set constraint quite low. OK, an example:
Prompt: a sidewalk in a small 1950's southern town
AI response: hot sidewinder in tucked-away Eisenhowerland.
You can see where the AI got each word, but every one of them is "wrong" if I wanted a normal re-phrasing. "Sidewinder" comes from sidewalk, but also means rattlesnake and is a kind of missle. "Hot" because southern sidewalks are often hot. "Tucked away" comes from small -- idea that town is obscure and unknown. "Eisenhowerland" = 1950s.
I shudder at the idea of trying to get this kind of interesting and, to me at least, very appealing distortion from GPT4, whose bland beige prose really repels me, but I do know it's good at following directions, and I could describe to it what I want and experiment with prompts. It would not surprise me at all if literary types were experimenting with this kind of AI. Do you know of anywhere I could hunt for specialized AI of this sort, beyond just a google?
Oh one other thing: No way I'm going to try to train my own AI, but had the thought just now that along with my basic prompt I could feed GPT a bunch of words to use in rephrasing it, and that could be the equivalent of the seed in Stable Diffusion text-to-image. But it would have to be a *lot* of words. What's the limit on prompt size, do you know? Also I could feed it some rules for relaxing syntax. For ex., that any noun or adjective could be used as a verb. So somebody could soapdish himself, or fat on down to the store.
Arrgh, yeah, I hate GPT's way with words. What was your prompt? GPT sounds like it's trying to be arty and poetic, because it's spewing sentimental arty cliches. "Time worn melody," "etched". "bygone". "embrace". Somebody once wrote about a piece of bad, sentimental music that "the music hears for the listener" -- i.e., it tells you what you're supposed to feel instead of making you feel something. Hey GPT? Fuck off and shut up.
Why don't those who oppose abortion laws argue for unconstitutionality based on the Establishment Clause? The reasoning for Roe v Wade, that the fourth amendment prohibition on unreasonable searches and seizures protects the right to abortion but only during certain times in the pregnancy, appeared to me to be ridiculous. But this sounds a lot less ridiculous: our founders passed the first amendment to prevent us from having to live under religious laws. Since essentially all of the motivation for abortion bans comes down to religion, they should be struck down as Establishment Clause violations.
Since essentially all of the motivation for murder bans comes down to religion, laws punishing homicide should be struck down as Establishment Clause violations.
(I can go through the whole Ten Commandments if you like - theft? deception? bigamy/polygamy? all religion based! Bible says "thou shalt not steal", therefore having laws about robbery and theft is establishing religion!)
There are non-religious arguments for restrictions on abortion, no matter how loudly you may insist that the people who advance those arguments can only be doing so to hide their secret religious motives. And the belief that (most elective late-term) abortions are wrong, is not limited to any single religion but pretty common across a broad religious spectrum. Both of those are going to make the Establishment Clause a hard sell here. Which church is being established here, exactly?
It's even worse than that for his idea, though. The *motives* of people who support some particular law are not relevant to the question of whether it's constitutional. .E.g. essentially all of the laws against adultery in the US were created on more or less religious grounds ... but there is a large gulf between 'passing a law in support of widespread religious belief', and, you know, actually establishing a state religion or church.
"There are non-religious arguments for restrictions on abortion, no matter how loudly you may insist that the people who advance those arguments can only be doing so to hide their secret religious motives."
To the contrary, they are usually very vocal about their religious motivations. They make secular "philosophical" arguments but it's obvious that they're only post-hoc rationalizations for their religious commandment and don't ever convince anyone who's not religious.
The most radically anti-abortion person I've ever met is an atheist lesbian anarcho-communist, who I met through a college secular society. If she's a secret christian fundamentalist she's doing a damn good job of hiding it.
The second most staunchly anti-abortion person I've ever met is an agnostic leftist whose pro-life views are directly related to the birth of her first child. But no, I'm sure she's just constructing all that as a post-hoc rationalization for her deep-cover christian extremism.
This is the idea behind the Satanic Temple's strategy of making abortion a sacrament, then claiming that abortion bans violate their free exercise rights. One might think this would be a promising strategy, as SCOTUS has ruled in the past that municipal bans on animal sacrifice are unconstitutional for this reason and the current SCOTUS claims to be favorable towards religious rights.
But in fact, everybody knows that this strategy isn't going anywhere, because legal realism is mostly true; American constitutional law is mostly about political power rather than logic; and this is the exact issue where six of the current Justices were specifically chosen for their willingness to sign a blood oath to rule a certain way.
A less accurate but more interesting possible counterargument is that this more ambitious reading of the Establishment Clause is self-defeating. It's hard to think of a non-secular argument for requiring Congress to make only laws that can be justified on purely secular grounds, so wouldn't enforcing such a constraint be tantamount to establishing atheism under this theory?
"This is the idea behind the Satanic Temple's strategy of making abortion a sacrament, then claiming that abortion bans violate their free exercise rights."
Do Satanic Temples (or whatever the local groups are called) perform abortions, though? It's hard to argue "this is a sacrament of my faith" when you're not doing any such ritual.
And if they are going to make it a sacrament, then they have to provide abortion facilities, train their own abortionists, offer it to members (ooh, can they only offer it to members? if anyone can go to the Satanic Temple for an abortion whether or not they belong to the congregation, is that a sacrament?)
I think this would cost them a lot more time and effort than is feasible; they can't just say "oh, Planned Parenthood clinics do our abortions so that's the sacrament" because I'm fairly sure even PP doesn't want to get into that mess and if you are saying a third party does it, then how is it a sacrament of your organisation?
1) First Amendment jurisprudence is rightly circumspect about imposing formal criteria on what counts as religion or atheism (which is equally protected). American courts are simply not in the business of saying "that's not a real religious ritual, because you're wearing yellow robes and chanting about inner peace instead of wearing white robes and chanting about inner holiness". If SCOTUS wanted to impose such a criterion to disqualify abortion rituals, they'd have to do a bunch of work to justify the massive departure from precedent and awkwardly carve out all the non-central examples of religion except the one they're targeting for ideological reasons, which is embarrassing for them.
2) Your proposed "in-house" criterion is violated in plenty of central examples of religiously protected activity. For example, Jews have a 1A right to buy kosher meat even though kosher slaughter not only doesn't have to happen in a synagogue: it can happen in a commercially owned slaughterhouse that doesn't have to belong to a Jew, be performed by a non-Jew, result in meat that is freely sold to non-Jews, etc. as long as a Jew is watching to make sure the basic criteria are met.
3) Satanists very much do perform "in-house" medication abortions, which are very much among the abortions forbidden by the laws they want to challenge.
The Establishment Clause doesn't actually say "No law may favor stupid old-fashioned Christian morality where it contradicts Objectively True and Correct Secular Morality."
It is about government favoritism towards specific churches, not allowable lines of moral reasoning.
The "American Constitution" is mostly about the powers of the Federal government, with initially very little to say about what the States did within their own borders. The Tenth Amendment certainly seems like it would prohibit the Federal government from banning intrastate prostitution, and in fact the Federal government has not passed any laws generally banning prostitution.
Most states have, and you'd have to look into their constitutions to see if they are behaving properly.
Or you could possibly invoke the Ninth Amendment to the US Constitution to argue that people do in fact have the guaranteed right to be a whore and then the Fourteenth to argue that the Feds have to stop the States from trampling over that right, but it would be a stretch.
Doesn't the prohibition generally come under the rights to make laws? I suppose you could argue that prostitution is engaging in trade, and the government has no right to interfere in the ability to conduct commerce, but we don't permit all and any kind of trade to happen - you can't kill people and sell their parts for meat, even if butchers' shops are legal.
Abortion should be considered as the killing of a human. The matter of opinion, I think, is when does the baby become a human. Catholicism would have it at conception, but other points may be considered, such as implanting of the fertilized egg in the uterus, development of limbs, a beating heart, leaving the body of the mother, and probably dozens of other points (I'm not a doctor).
One is not allowed have one's children killed, even particularly annoying ones, except before birth? This strikes me as bizarre.
You began your "argument" with the claim that "Abortion should be considered as the killing of a human," which is really your conclusion. (All the actual argument about abortion is the part leading up to that conclusion.) I don't think I've ever heard anyone other than conservative Christians say that. The idea that abortion is killing a human is based on the Platonist metaphysics underlying Christianity, and doesn't seem at all obvious to most non-Christians.
I'm pretty pro-choice, but even I agree that the dividing line of birth is as equally arbitrary a place to put a "human life with moral weight/non-human life with little/no moral weight" as anywhere else.
Everyone things that killing humans is wrong. The argument is about where the line is between that it becomes sufficiently non-human (in a moral sense not a biological sense) that it is ok to cause it to cease to be. Extremely pro-choice people think it's at birth, most American's think it's somewhere around 24ish weeks, very pro-life people think it's at conception.
I don't think his statement was wrong exactly, even if it was slightly inflammatory.
-edit- 24 weeks was wrong, that's about the time that a slight majority think it should be illegal. So a majority thinks it should be legal to somewhere between the end of the first trimester and 24 weeks.
The most unambiguously human with the most skin in the game for making the call is the person carrying the fetus. Not to say that should be the only voice but it needs to be given some degree of deference.
If I sounded inflammatory then I apologize. I honestly thought I was simply laying out the reason non-religious people can be against abortion, in the clearest and briefest terms possible. I got advice long ago that, as an author, it was a sin to waste the reader's time.
The arbitrariness of birth as a dividing line wasn't something I really internalized until the first time I was in a room with my pregnant wife and OB scheduling the best day for her to be induced. It was as if we were scheduling a conference call. I recall that Monday and Wednesday were both inconvenient for some reason so we settled on Tuesday for the day my firstborn would "become human", as some would have it.
After that, I was deeply impressed with the feeling that birth is a change of location -- a disruptive one to be sure, more like moving to another country than moving across a room -- rather than a clear-cut stage of development.
Meanwhile development itself, I learned as a parent, is such a smooth and seamless process that any particular line ends up feeling arbitrary. All these milestones like "first steps" and "first words" end up not nearly as clear-cut as I was led to believe.
I mean, I'm a happy new father, and I was exited to see every ultrasound we did, but the first one absolutely parsed as "alien lizard thing" rather than "small human". Early foetuses really don't look very recognizably human to an untrained eye.
No, I'm not a parent, and you have a good point. I have heard parents speak about the excitement of seeing an ultrasound (though I don't recall any of them connecting that experience with abortion).
Seeking suggestions for activities in the Sacramento–Berkeley corridor, June 16-17 (Friday and Saturday). I will have a car.
Friday, I'll be starting from Davis/Sacramento and have dinner plans in Berkeley. Saturday I'm free all day, but I have to end up back in Sacramento by day's end.
I like day hikes over rocky, uneven terrain (including some bouldering). I love scenic views. (I also like well-stocked guitar stores.)
Mount Diablo would be the obvious choice. On a clear day, you can see more square miles of land from the peak than any other peak in North America (or so I've been told). It's possible to drive to the top, but a better experience to park elsewhere and hike to the summit, I think.
Tilden park is also quite nice for getting views of the bay.
Soon after starting I realised I can use my browser history as an approximation because most of the time I would turn my PC on and off after waking up and before going to sleep, so I have 3 years of pretty accurate data.
It seems that each year there is a bump in sleep times in winter and for some reason 2 smaller bumps in summer. What could this mean?
Since I started tracking my sleep I also tried to intentionally bring it closer to midnight but mostly unsuccessfully. It seems that I can't escape this yearly rythm of bumps.
One explanation I found is https://en.wikipedia.org/wiki/Phase_response_curve. This would mean than my sleep advances with the rate of advance of sunrise. So the ambient light coming through curtains arrives in my eyes early relative to my sleep during spring but this gradient slows down as summer solstice approaches. Though this doesn't explain the double bump in the middle of summer...
I just looked it up, and you are right about mail-in votes in European countries.
I think you are wrong about 'there's no particular problem with mail-fraud in elections'. The Democrats are just too openly crooked, and they monopolize the bureaucracies. And the bureaucracies are the real government, not the ceremonial remains of elections.
The last episode of this season of Hi-Phi Nation went out last Tuesday, its about three different AI music technologies, vocal emulators, fully generative music, and musical improvisation, including a guitar solo contest between two humans and two different guitar-solo generators. Check out the whole season also on the ethics of digital futures. https://hiphination.org/season-6-episodes/s6-episode-6-rise-of-the-music-machines-may-16-2023/
"Google Registry has launched some of the most popular (and secure) top-level domains, such as .app and .dev. Today, we’re adding eight new extensions to the internet: .dad, .phd, .prof, .esq, .foo, .zip, .mov and .nexus."
The newsletter article was pointing out how .zip was a dumb idea and the problems with cybersecurity involved, but that aside, who the heck wants a domain name like .dad or .phd?
"Knock, knock. Who’s there? With Father’s Day right around the corner, .dad is here for the jokes, the games and the advice. Whether you’re a fit.dad, a gay.dad or a dude.dad, .dad is the place to celebrate fatherhood."
Uhhh - there are already sites like these, granted on the distaff side - Mumsnet and the like. Do they really expect enough people will want an address like Gary.dad in order to go "yes, I'm Gary and I'm a dad"? As for the rest of it:
"Grads
May means graduation season for many in higher education. We’re celebrating graduates and the professors who taught them well by launching .prof, .phd, and .esq. Whether sharing legal advice for everyday life or teaching courses on behavioral science, these new domains are perfect for showing off your credentials. Hats off to these early adopters:
Erika.esq: Erika Kullberg is an attorney and money expert who is passionate about better positioning people for success.
Casey.prof: Professor Casey Fiesler is a technology ethics educator and science communicator.
Rafael.phd: Rafael Misoczki is an expert in post-quantum cryptography, fully homomorphic encryption, privacy enhancing technologies, and the application of these constructions."
Ah yes, Erika, I'm sure that having a vanity plate website is going to do wonders for your professional reputation. And were I trying to look up an "expert in post-quantum cryptography", I would not be at all confused by a domain name like Rafael.phd
And this one? david.mov: Watch videos by David Imel in this liminal space.
It is horrible and redirects to Youtube, where most people are going to go looking for videos in the first place. That is not a "liminal space", it's a mess.
I agree, the ".zip" is an extremely dumb idea. Now whenever someone mentions a ZIP file in an e-mail or in a comment, it may be automatically underlined and linked to a completely unrelated website, probably owned by scammers who will register websites such as "report.zip" and "document.zip" and "data.zip" and put there a nice big Download button that many users will click without hesitation. Fuck Google for making things complicated for everyone.
A few years ago I was teaching an "introduction to computers and internet" course for some older people who missed the train. Like, how to use Word and e-mail and stuff. And of course they asked me "how can I buy things online?" and I gave them some careful advice, but I thought that the best possible answer is "just don't", because it's simply impossible to explain during one hour, or even one day, *all* the things you need to do to avoid getting scammed. (There are some good rules of thumb, such as "only trust the URLs you have typed manually", but then things get more complicated.) Now I imagine myself explaining to some 70 years old person who just got a computer for Christmas and decided to finally give it a chance the difference between clicking on the "photos.zip" attachment in their Gmail, and clicking on the automatically underlined text "photos.zip" which will lead them to some scammer's website. Why?
(Good luck to Erika; all her potential customers will type "Erika.esq.com" to their browsers anyway. Heck, even if they won't, there is a chance that the browser will helpfully update the entered URL itself, because that's what helpful programs do, don't they?)
That's a good point! Auto-correct is going to have a field day with the new domains, and of course the scammers will take advantage of that as well - you think you're clicking on Ericka.esq, the programme auto-corrects to Ericka.esq.com, and some Russian scammer is now looting your bank account 😀
Plus I've a feeling Ericka is going to get a lot of communications from lonely gentlemen requesting photographs of her beauteous form and proffering photos of their own... advantages... as well.
Maybe people will rush to buy up domains they think will be valuable, which they wish they had with .com domains. Profitable for Google, it sounds like, but probably a bad idea for the buyers.
The example sites are just market promotions. Google is selling a new product line, thats a better way to think about it. It's just another flavor of oreo. Maybe you dont like it, thats ok.
Using these less common domains isn't unusual for professional personal websites. There are only so many options for .com and .org or what have you and most of the good ones are gone. I can't get my name as a .com but I may be able to with a more unique domain.
I think the assumption is that people would rather take gary.dad than garyhasselhoffsdadblogsite83723.com. I agree that I already find myself confused when anything other than a .com, a .edu, or an occasional .org is used, so a marketer needs something really snappy to overcome the oddness of a non-standard extension.
The Wold is a long-running Dungeons & Dragons campaign world that emphasizes play-by-post gaming and a commitment to write one game turn every weekday. We are a community of about 65 gamers in nine long-running games, and we are looking for new members. I'm recruiting for first level players for a new game, in which the players are would-be heroes migrating to an embattled dwarf clan to help them survive and thrive. We have openings in current games, too, at higher levels. What do you need to know about the Wold?
• We are here to stay! The campaign dates back to 1985, and we have online archives back to 1996. You can expect to start a character and play through 20 levels over the course of a 10 year career.
• We are free! There’s no cost to join and no test to pass. Check out some of the links below, and if you like us, you are in. As our founder, Jerry says, “There is no tryout or application. My experience is that those players who stay are the ones who would have passed any application or tryout anyway. So I let the players weed themselves out or remain to become a permanent player.”
• We offer a very rich gaming experience! We have our own homegrown gods, nations, races, and peoples; we have custom classes, spells, feats, items, and more — all detailed in our own wiki. We play by the Pathfinder 1E flavor of the Dungeons and Dragons rules these days, but we are eager and glad to teach players who need help.
• We focus on heroic themes! All our player characters are Good. Feel free to create a troubled and conflicted character, so long as you are interested in redemption, nobility, and overcoming obstacles to become a true hero.
• We are a friendly community of people from all over the world, from all walks of life, and of all ages! The gaming groups that stick together over time are the ones in which players become friends, and we encourage that.
It's funny how in the ACX book review contests, all the finalists are good, but one of them is so outstandingly good that it is clear within 5 seconds who's going to win. I wonder why that is. Is this true for most contests in life, which have good enough incentives to draw in high quality participants?
I don't think so. E.g. it's unusual for any ESC entry to have a more than a 50% chance of winning, and likewise in the national competitions, judging by the odds.
I've noticed that the "hard part" of a lot of physical activities isn't what I would have expected before doing them with some regularity.
- Running: My limiting factor is breathing properly, not my legs getting tired. I wind up gasping for breath long before the rest of my body is tired.
- Bouldering: My mind being tired is a bigger problem than my body being tired. With the former, I barely trust myself to be on the wall at all, since I'm worried that I will zone out for a second and fall off and hurt myself.
- Bicycling: What makes me want to stop isn't my legs being tired, it's my butt being in extreme pain from sitting on a seat for so many hours.
I'm not sure if this is a common phenomenon related to outsider versus insider perspective, or something else.
Re: gasping for breath, I assume your problem isn't https://en.wikipedia.org/wiki/Exercise-induced_bronchoconstriction ? I'm mostly bringing it up because that's been my issue and it was a vexing journey to find out. (Thankfully, there is plenty of anaerobic exercise to be had, so I still get to exercise!)
It had occurred to me, but I've never been diagnosed with asthma, and have only noticed this problem when running. For example I biked for ~50mi this past Sunday without much trouble, but would be gasping for air if I tried to jog ~5mi.
Are there any telltale signs to distinguish asthma from other breathing problems?
I've never been diagnosed with asthma, either, mostly because it's only exercise that causes the problem; for most of my life no one even *suspected* I might have a physiological issue with sports at school, the narrative was always 'you are out of shape'. Turns out the symptoms are independent of how in-shape I am, though, or how much I take breathing tips to heart (breathe in through the nose, prefer deep breaths to rapid breaths, etc). What happens is that my throat/chest responds in a way I would describe as "like getting stabbed through the clavicle" (although with the caveat that I thankfully haven't had literal clavicle-stabbing happen to me, and so the comparison is probably bad) after even a short while of sprinting (at which point my legs usually haven't even noticed I'm using them much yet) and my ability to breathe drops precariously. If I really push myself, after I inevitably stop running, I will cough on and off for several hours, with declining frequency. (Some silly circumstances made me experience this less than a month ago while I tried very hard to catch a train and the associated coughing did not fully subside until bedtime six hours later.)
I unfortunately don't know much about how biking might or might not affect EIB, having been rather focussed on conversations about my own situation (and I don't bike; I walk or take public transport), but from my understanding of the mechanisms at work with it, I agree that this sounds like your problem is something else!
When I went on a ~100km/day bike trip, saddle soreness was considerable after the first day, but after that it was fine, and fatigue became the main problem.
Interestingly my experience with running is now the reverse of what you describe: I tried going for a run a few times recently after many years not doing it and my feet started to hurt way before I reached the limits of my cardiovascular capacity. It used to not be like this, so I am tempted to blame aging (and generally being out of shape).
I suppose it makes sense that physical exercise in general is a weak link problem.
I've literally had this switch up on me depending on when I take it up again. Right now it's 100% legs, but when I was younger and more in shape it was definitely CV capacity.
What sort of running are you doing, and how long have you been training for? Most amateur distance runners who have been training for more than a few months find their legs end up being more of a limiting factor than their breathing. (The conventional advice is to do most of your training at an "easy" pace that would allow you to carry on a conversation. But it sounds like you are running faster than that, so maybe you're following a different training regime.)
I have had the same experience as the OP about running, for basically my entire life. I can’t remember a time I stopped running due to my legs, it’s always my lungs that give out first.
The kind of running I do is running at a fast pace (if you were generous you could call it a sprint) until running becomes uncomfortable enough that I want to stop. Then I walk until I feel good enough to start running again.
I used to run like that and later found an expert runner who agreed to be my coach.
He told me that running like that doesn’t allow you to improve much since you are running in your anerobic zone, which has a much smaller training response. So hard to improve the bottlenecks of the sprint-running
Running slower means you spend a lot more time in the aerobic zone which is much more trainable. So it’s best to spend upwards of 70% of your runtime in the aerobic zone.
Anyway Experience bore him out and after I switched to running much slower but continuously, I saw my endurance and speed greatly increase - so that I can now run continuously at aerobic zone at my previous sprint-speed in anerobic zone.
Had to come back to say that your comment actually changed my life.
I have never had an exercise routine, though not for lack of trying. I was never able to make it stick. But your comment here led me to try running slower for longer, and to focus on running longer as my improvement target than running faster. And it worked! It worked great! Now I run 5 days a week, my endurance has improved significantly, and I'm happier and healthier as a result.
Your experiences are "normal" and have one thing in common (except maybe the bicycle seat pain): it is your mind that makes you feel tired. It doesn't mean it's not "real", it's just that the physical limits of the body are far from exhausted by this point. A lot of endurance training is literally learning to "endure" the pain/discomfort/boredom/ of extended hard work.
Interesting conception of modern LLMs not as software (reliable and precise, but extremely limited), but as a crappy human intern (inconsistent and low-quality at most things, but can be useful). https://www.oneusefulthing.org/p/on-boarding-your-ai-intern
YouTuber I recently came across in the realm of political philosophy. I think this was the first time I'd seen a clear minimally-politicized explanation of what "cultural Marxism" means: https://youtu.be/4JX4bsrj178
I would not take this fella very seriously if I were you. I don’t know if his clumsy grasp of Marxism is due to bias or genuine ignorance, but I t seems to me he doesn’t want you to take “wokeness” (identity politics) seriously as a genuine intellectual tradition while couching it in the aesthetics of being non-politicized. Anyone who paid attention to my freshman civics class could take apart his characterization of identity politics. There are people communicating political philosophy at a far higher and dispassionate level than this guy.
Has anyone else tried both prescription + over-the-counter options for ADHD? Historically I was on an extremely low dose of instant-release Adderall (I would break up the pills into tiny ~1mg pieces). The main downside with that was the inevitable crash in the afternoon which turned me into a zombie in the evenings. I guess that could be fixed by using an extended release instead, but the problem with XR is that I can't break it into lower-dose pieces, so the dosage would likely be much higher than I'd be comfortable with.
Lately I've been using over-the-counter Thesis instead, and while it's better than nothing, it's certainly no Adderall. But on the plus side, there's no crash.
With Adderall XR formulations, it's very possible to split the dosing up. XR adderall is just capsules that are filled with what look like round sprinkles that have a fairly simple bi-phasic release mechanism: 50% of the dose up front, with the other half released ~4h later. If you wanted to recreate your 1mg dosing, you'd just measure out 2mg of the XR sprinkles.
Similarly, Vyvanse is just powder in a capsule that you could split up into portions if you so choose.
I have never heard of a dose smaller than 5mg (which they do have in XR). Why were you breaking them up so small?
I find the best way to break through the crash is to do something active around the time of the crash. Could be exercise but could also be a walk or an errand or making dinner. Anything to take my mind away from noticing the crash and keeping my body moving.
I found that with the standard 5mg doses my mind would "run away without me" and I would be less productive on net. There'd be lots of movement and action but with little directional coherence to it. An extreme example (which was years ago and on a much higher concerta dose) was getting sidetracked from my math homework by some halfbaked idea, scribbling about it furiously for an hour+, only to come down and realize that it didn't make any sense and I was just being crazy.
Wow thats very interesting and I am sorry that you had such a response. Unfortunately I don't have any suggestions to help! In reference to your other comment, you can be a pioneer in microdosing Adderall!
What about breaking up the instant release adderall into even tinier bits and spread out the period of dosing more, so your crash happens closer to bedtime? Or dissolve it in water, and take measured doses of the liquid?
It's not microdosing, though -- you'd be taking your usual total dose, just spread out longer by taking it as many tiny doses. Another thing to try would be to take maybe 10% more than your usual dose, having the last dose be in the early evening, so that the crash comes around bedtime.
One last thought: A patient of mine with ADD is just starting to try meds. He first tried ritalin, which made him feel unpleasantly jittery. He was able to stick to things for longer periods of time, but they were things like organizing a room, not doing his job (he works from home). Then he switched to adderall, and find that the jitteriness is much less, and that he feels more able to focus on the things he knows he needs to focus on, rather than on whatever he happens to be looking at when the energy first hits. So if you've never tried any other ADD meds, it might make sense to try one. There are several different types.
Lengthy piece (from a Caribbean economist) arguing that Caribbean nations should abandon their local currencies and just use the US dollar, as Panama, Ecuador, and El Salvador do now. Apparently a candidate for president of Argentina is loudly arguing for doing the same thing. Author takes on/discredits arguments for monetary sovereignty, devaluation, etc.
I've thought for a while that a stablecoin US dollar might actually explode dollar usage globally- not sure developed country residents understand how strong the demand is for a stable currency in inflation-prone countries (could be the euro or the yen too).
Kind of funny that as US relative power declines, some aspects of it actually increase. I'd argue the US has more cultural power (TV shows, celebrities, music) than 20 years ago, even as US military & economic power has declined a bit. Apparently English is just becoming more & more dominant as a global lingua franca. Maybe the US dollar will actually (counter-intuitively) become *more* entrenched as the global reserve currency, even if the world shifts to being multipolar!
EVs and LNG import terminals are what the Caribbean needs…everything else macroeconomically speaking will be easy once oil no longer dictates so much of island life.
I think it was Noah Smith who recently made the case that Jamaica should try to get the new mini-nuclear reactors installed as quickly as possible. Apparently they already run a research reactor, so switching to nuclear would help their economy a lot without adding much proliferation risk. It was an interesting idea, and if it worked, could potentially mean that long term more island nations could switch to nuclear.
in Argentina, it's indeed touted as a way to restrain monetary policy which everybody agrees is consequence of bad fiscal policy. And it's a red herring because to make the switch you first need to fix the fiscal situation, no way around it. Once you do that, there's no need to buy foreign currency that is losing ~8% of its value every year.
Fiscal deficit with sovereign currency = high inflation. Fiscal deficit without sovereign currency = complete and total crash.
The candidate that is touting dollarisation is pretty much a contrarian teen and his arguments are basically burning the ships.
I just think it would be interesting if the people of Argentina (or any other developing country) started using US dollar stablecoins to pay each other, and it eventually reaches such a large enough scale that the authorities have to throw their hands up & give up. I mean, that would probably take a while, and I'm sure the Argentine government would be Very Very Mad about it. I'm just talking about a groundswell of regular people paying each other for food, gas, clothes, whatever with peer-to-peer dollar coins. I guess they'd probably need a bank or bank equivalent that's outside of Argentina's regulatory authority too
Yes, the funny corollary to the efficient market hypothesis that says no one can consistently outperform the market is that no one can consistently underperform the market. That being said, the place retail traders go wrong is they have no idea the risk they are taking (or they don't care) and will blow up eventually.
You can easily underperform the market in risk-adjusted terms by refusing to diversify. If I buy a limited number of stocks from the S&P 500, chosen at random, I should expect to have more risk than someone who buys the entire S&P 500, and no better average returns. The market rewards undiversifiable risk (beta) in stocks, not diversifiable risk.
"Alright," you say, "but risk here is just variability, not the same thing as returns."
Ah, but provided that my margin costs are less than long-term stock market returns (and they generally are), I can turn more risk into more expected returns by buying on margin. Therefore, there is some positive level of margin below which a well-diversified portfolio will have less risk than an unlevered randomly-chosen undiversified portfolio, and greater average returns.
One, collectively investors underperform the market by the friction costs.
Two, anyone can consistently outperform the market via pure chance, see also an example of someone consistently throwing heads 10 times in a row - out of millions of participants in the coin-tossing trial. The key point of the EMH is that it's impossible to predict in advance who that lucky 10-heads person will be.
FWIF this is the only part of EMH that is practically useful. The stronger version is hard to defend.
I'm fairly sure that consistently underperforming the market is as easy as consistently picking the wrong number someone is thinking of between 1 and 100. Possibly easier, given how many stocks there are.
Or to put it another way: one cannot discover the next miracle material by combining different substances at random.
If someone *always* got the wrong number, then that would be useful information - you could guess any number besides the one they picked and you'd have a 1 in 99 chance instead of a 1 in 100. Not a lot of information, but it's something. But in reality a person who "consistently" gets that one wrong is still right 1 in 100 times.
Similarly, if someone *always* underperformed the market, you could make money by shorting whatever they picked, but in reality these people will occasionally pick winners. (Or rather, they would if they had enough money to keep picking stocks forever instead of going bankrupt.)
These are both instances of the “optional stopping theorem”. If something is really a random process, with expected gain or loss at any moment equal to zero, then any choice of when to play and when to stop that is based only about information on the past and present will also have expected gain and loss zero. (The efficient market hypothesis basically is the initial assumption, but it is often stated as one half of the conclusion of the theorem.)
Anyone else notice the strong Christian themes in the new Guardians of the Galaxy movie? For populist entertainment, it surprised me. [spoilers ahead]. For example, the villain is the God-denying “High Evolutionist,” who engages in the worst imaginable form of eugenics because he alone can perfect life. The end of the film is basically Noah’s Ark, and the hero StarLord dies while saving everybody then comes back to life. There’s also a theme of every life is sacred. It’s just a bonus, I suppose, that the villain is black and the children he is tormenting are all blonde-haired, blue-eyed Aryans.
As a Christian, I can't say any of that comes across as particularly "Christian themes" to me. The villain is the most obviously evil and repugnant villain Marvel has had in a long time (... err, unless any of the other villains in Phase 4 were similarly evil, can't say I've seen them), so it's not exactly a strong Christian theme to say "maybe inhumane experiments on animals and literal planet-destroying genocide is bad".
You can draw a few parallels between the Biblical Flood and the end of the movie - certainly High Evolutionary has an ironic God complex, but they're fairly weak and felt largely unintentional to me.
And if they were intentional, I don't know that casting your megalomaniacal villain in the role of "God" in the biblical story is exactly what I'd call a "Christian theme" (at least not of the sort that I'd be surprised to find in modern pop culture).
---
The bit with Starlord at the end is just standard hero stuff: I don't think every hero who almost dies is a Christ figure.
And, I think this film *could* have taken an "every life is sacred" message, but really doesn't. Mowing down a bunch of villainous mooks in the action sequence in order to save a bunch of cute innocent animals... if you wanted Christian themes, it would actually be the *reverse* of that! On both counts!
(If anything is a Christian theme in this movie, giving Orange Boy a second chance is the closest it comes, but even that's a fairly milquetoast "'forgiving' someone who really hasn't hurt you" act)
All that and you didn’t even mention the character who goes to heaven and is assured that a higher power is guiding things? It’s the most Christian part of the movie by far. Surprised me.
" It’s just a bonus, I suppose, that the villain is black and the children he is tormenting are all blonde-haired, blue-eyed Aryans."
Meaning unclear, please develop your point. I have a vague feeling you intend to convey something, but I just can't quite grasp what it may be. Are you Darkly Hinting?
As for the High Evolutionary, he's been around since the 60s. He started off as a white Englishman, having a black actor portray him in 2023 movie is just DEI.
Also...Adam Warlock, in his, let's say, second revamp, was the most obvious and heavy-handed of Christ allegories. This was dropped by the time of the comics that are actually the reason he's in this movie (Jim Starlin's 1970s run).
In the US context, there’s a (historically contingent) link between Christian fundamentalists and racism. This particular casting choice, coupled with the Christian themes, made me wonder if the movie was intentionally playing to that audience.
You're thinking of the historical correlation being Southern and being racist. Those sects which had many fundamentalists--the Baptists and Methodists--each split into a northern church which was not historically linked with unusual racism, and a southern church that was. Factor that out, and I think you won't find any remaining correlation.
Let me be sure I have this right: you are suggesting *Marvel* is pitching its movies to *Christians* and not just any old Christians, but *fundamentalists*.
Is this because of Chris Pratt? Because I'm given to understand the kind of people who were big fans from the Parks and Rec days turned on him due to him belonging to a church that wasn't all gay marriages carried out by trans lesbians ministers with rainbow stoles.
Just come out and say "Nazis". You know you want to, with your "blonde Aryan children" bit.
EDIT: I may be being unkind. This could be a test run for another Quillette article as per your website. If you can get it to fly here, you can easily crank out something about racist Christian fundamentalism for them.
Does anyone else get the impression that the debt ceiling "talks" have some other objective? It seems more and more like theater, with the announcement of reaching the ceiling months ago, and soon we will have a last minute "save". Oh, the drama!
I actually am thinking this is about something else completely, and we, the public, are completely in the dark as to the actual objectives.
It's the same as it's been all the previous times. Chicken plus bartering. We start off with "you're a nazi" vs. "you're a communist", proceed through "your policies will destroy America and all Americans", get closer with "our version of economic stimulus is better than yours", and end up with "we'll cut more from your programs than you want but less than we want".
So yes, totally theater. The point is to generate public pressure on the other side, by implying that We're completely reasonable and the only problem is Their intransigence. Eventually it'll pan out in serious negotiations behind closed doors where the press can't go, so that each side will be able to justify their compromises by pointing to all the things the other side didn't get, and no individual politician will be held responsible by their voters for failing to uphold their absurd promises.
Or maybe this will be the time it all breaks down! Maybe we've finally run out of adults in the room, and all that's left is kids who haven't found out that Santa Claus isn't real. But that's never happened, so by induction it never will, just like death and AGI ruin.
I think that we have a very clear idea of what their objectives are; each party wants their Presidential candidate to win the next election. Right now both parties are playing a game of chicken to see which side voters will blame for a default. My interpretation is that they don't care about the debt ceiling, they're just using this as a stage to score political points. I don't think that there's any deeper meaning than that
I think this is just the nature of modern politics. Much of what the public sees has been written, edited, and shoved in front of a focus group. It's probably not too healthy to put too much into thinking there are completely different objectives hidden from the public, but it's also somewhat understandable given our institutions' track records of late.
I have been thinking it is worse than they let on, because few who are taken seriously are talking about budget reforms which might prevent us from getting in these situations. It seems like the people who very reasonably suggest our direction isn't wise and probably not sustainable are labeled cranks.
Giving the President the power to borrow unlimited money in the name of the American taxpayers, with no oversight or accountability, would seem to be a bad plan. Unless you believe in the Magic Money Tree, I suppose.
Does the president have the authority to borrow unlimited money? I thought the executive is constrained by laws in what it’s allowed and mandated to spend money on, and the debt ceiling is just a contradictory law banning the executive from borrowing the money to pay for what it has been mandated to pay for.
I don’t think anyone is proposing that the executive get discretionary fiscal powers.
The president can do things like sign contracts whose nominal value falls within authorized and appropriated limits but will predictably escalate beyond them, or tell his AG to take a dive when a favored interest group sues the government for Sagans of dollars, at which point the government is legally obligated to either cut back spending somewhere else (which may be illegal), or issue more debt. A debt ceiling subject to congressional review means that it's usually not worth gaming the system that way; I'm not sure what the equilibrium will be if that goes away.
Maybe. One of the weird aspects of our current setup is that so much of our debt is offloaded to foreign central banks that devaluing the currency may be a good way to transfer wealth from them to us. Which is not *morally* good but is certainly a hell of a grift if you can make it work.
During an inflationary time, it might be good to temporarily have a surplus. But usually, a surplus is when the government is profiting off the people, which is the fundamental thing a democratic government should never do.
Best to pass a debt ceiling increase that *doesn’t* come with a bunch of UN negotiated cuts to every aspect of the government. Get rid of the terrorist threat *before* you begin to negotiate, don’t just give in to the terrrorists before asking them to give up their threat.
>The only risk is that certain partisans -- the sort of people who genuinely think that "let's mint a trillion dollar coin" or "let's use the Fourteenth Amendment as an excuse to bootstrap a literal military coup" are workable public policy
Why isn't the trillion dollar coin workable public policy?
I agree with trebuchet that it's a ridiculous concept born in unserious minds, but here's an attempt at explaining the reasoning.
If we printed a trillion dollars worth of smaller bills to pay off debts, what would be the result? Obviously with more money in circulation and no additional goods, we would see significant inflation. That would be bad.
Would that be any different with a trillion dollar coin? If not, then nothing is gained. If so, then why? Probably because a trillion dollar coin could never be spent. I guess we just give a major creditor the coin and tell them they're paid off - and let those suckers try to spend it! There's no way to exchange it for smaller bills or buy anything with it, so it's a bit like an IOU from the feds to whoever. If I were a creditor of the United States, I would refuse to accept it.
> If not, then nothing is gained. If so, then why?
It is literally on the Wikipedia page - "The issuance of paper currency is subject to various accounting and quantity restrictions that platinum coinage is not."
If we're paying off the debt owed to internal stakeholders, like the Social Security fund, then all we're doing is shifting money around and screwing over Social Security.
In re hypergamy: I haven't read all the comments, but does this mean that some number of men (incels and the like) built a lot of their lives around an idea that might not even be true?
I don't think anyone becomes an incel on the basis of a theoretical belief regarding their probability of romantic success, but rather through hard experience that they don't have whatever it is that women want in a romantic partner. Their having a possibly-false idea of what that really is, seems like a minor perturbation that changes nothing.
Also, there are I think a fair number of incels who either are "high status" in the specified markers of class, wealth, education, and age, or can reasonably expect to be in 5-10 years when they graduate and get an UMC job. And more who aren't, but the ones who are wouldn't be surrendering to inceldom if they believed in the sort of hypergamy that Scott is talking about.
It reminds me of those who thought Piltdown man being a hoax means evolution didn't really happen. When you are strongly, emotionally motivated to deny something, it's not difficult to find some false claims connected to it, laser-focus on them, and then declare the whole idea to be discredited.
The Peer Review posts are Scott looking for feedback from subscribers on early drafts. He says in the linked article, "I’ll be posting this sometime in the next few weeks."
I'm sorry-- I didn't check how accessible it is. Of course you can discuss what I wrote, even if you don't have access to what Scott wrote.
I'm not sure whether courtesy to Scott would permit me to quote from it, but the conclusion is that people pretty much marry people of similar attractiveness and from the same class.
Class matters much less to homosexuals (box sexes) than to heterosexuals.
Many, many people build their lives around idea that are not true; take for example Religion, where at most 1 of them is right and thus at least 6 billion people quite strongly believe something profoundly wrong.
That's certainly true, but it's a much shorter chain of reasoning to start with "Women have a strong preference for high status men" than "this is the one true religion".
I have a notion that part of believing in hypergamy is that people who play gender roles hard are conspicuous, so it's easily to notice the supermodel married to the rich bodybuilder, and not notice the pleasant-looking woman married to the mildly successful guy.
One upshot of a "shorter chain of reasoning" is there's less space to smuggle fallacies in. God is a vague and nebulous thing to argue for/against. Meanwhile if hypergamy were obvious poppycock it should be easy to see if it's not true/has bad explanatory power. Yet to the casual eye, it seems to hold up.
Not sure what you mean in your second paragraph - isn't a 10 going with a 10 and a 5 going with a 5 not a positive example of hypergamy? Both women have scored the highest value man they can get.
Not much point my saying any more though until I can actually read the source article.
> isn't a 10 going with a 10 and a 5 going with a 5 not a positive example of hypergamy? Both women have scored the highest value man they can get.
And so did both *men*, in this example. A pairing of 10 with 10, and 5 with 5, is simply everyone choosing the best option available (assuming monogamy). The only thing this proves is that people's opinions on who is attractive are correlated, otherwise we couldn't even meaningfully talk about 10s and 5s, because no one would agree who is who.
The actual claim of hypergamy is that women are attracted to men's *social status*. As opposed to men, who are attracted to women's pretty faces or big boobs. Learning that she also happens to be a rich CEO does not make her any more fuckable, to put it bluntly.
There is also a stronger claim, that women prefer high-status men *so strongly* that they would rather be the rich bodybuilder's second girlfriend than the mildly successful guy's only girlfriend. (While men have a stronger preference not to share their girlfriends with other men.) So if you entirely remove the social pressure towards monogamy, you should expect the man#10 to pair up with the woman#10 *and* woman#9 and possibly also woman#8; the man#9 would be with woman#7 and woman#6; etc., and the men#3-1 would be left alone.
(I haven't read the article either. Just commenting on the definition of "hypergamy" as I have seen it in internet discourse.)
This is what's known in cryptography as a "Grandmaster attack", after the story of the little girl who plays two chess grandmasters at once and comes away with either a win and a loss or two draws.
There are a bunch of religions with mutually contradictory claims. Working out whether any one of them is correct may requires thought, but proving that all but at most one of them are not is trivial.
I’ve been having a weird experience with spam emails lately. I get the usual mountain of spam emails in my personal email account spam folder, except they are addressed to the pet name my partner calls me. Crucially, this pet name is correct phonetically, but not as she would write it down. Is it possible that spam email accounts have obtained this name through listening to our conversations via one or more devices? I don’t use this pet name for any online accounts or usernames, and she only uses it around me (not when we’re with friends etc). It's only really a verbal nickname for me and isn't written down anywhere.
Below is a pseudonymous example of the scenario:
My name is Peter Johnson
My GF calls me "Peej" (and writes it like that in texts)
My email account is PeteJ@whatever
I get spam emails saying “Dear Peaje Get 10 free spins etc”
How come these emails are formed this way? Is one of my devices listening to me?
N.B. I don’t have an Alexa or any smart speaker type thing that is designed to listen around the house.
Your initials are P.J. The spam is just using some crude algorithm to get around blocking filters by using your initials to make something like a name that looks familiar to you.
There is something called Display Name Spoofing where your full name is used:
Here's a weird one: I went out to dinner with my daughter and some other people last week and was introduced to a new IPA called Madonna. A few days later I ordered some wine and beer from Drizly (liquor store delivery service) and as I was ordering it it suggest I get some Madonna. The reservation was under my daughter's name and cell phone number, & she has the same last name as me. I paid for the restaurant dinner in cash, not with a card. How the hell does Drizly know I drank a Madonna?
Probably just that someone in your social graph (maybe your daughter) got associated with liking Madonna. Then the delivery app queries your social graph app for suggestions and voila! Ain't technology great!
Alternatively it could just be a coincidence. Maybe Madonna is just the hot new IPA right now or is on a marketing kick.
It's possible that your information was obtained through a data breach or by other means of unauthorized access. Scammers can purchase lists of email addresses and personal information on the dark web, which can then be used to target individuals with spam emails.
I get that - and I'd expect spam emails to target me with my real name. What's surprising is that the name I get addressed by in these spam emails is a name I only get called verbally, written down as if overheard - surely this means it can only be obtained through listening to conversations?
By default, your cell phone is always listening to you, so it can tell if you say something like "Hey Siri" or "OK Google". This incidentally means it is analyzing everything it "hears".
I have turned off these kinds of things, which is hard to explain how to do without crippling the usefulness of your device. I don't get as much targeted advertising as my wife, who also thinks she has turned off these things. And I have my doubts as to whether they are truly turned off, or instead only not used to openly target ads.
I've worked on a system like that, and in our case, the system actually wasn't always listening; it didn't have enough power/battery to be doing that. What it had was a fixed function thing to recognize only the wake word, and the system ran in very low power until that triggered. I would expect this to still be the case for cell phones, but not necessarily for wired-power devices such as the Amazon and Google home assistant products.
Maybe through data analysis. Certain pet names are more likely to be given by people with certain characteristics. The more characteristics in the regression the better the prediction.
While I appreciate the originality, this idea just wouldn't work. For one, you once again assume Trump supporters are going to react to things the same way you would. But for another, the setup is physically much less safe than you seem to think. Elephants weight a *tonne*, and there's no guarantee all that weight will stay sitting obediently in one place. Imagine if one mouse gets in and causes panic amongst the herd: now the whole airship is coming down.
DataSecretsLox has a "DSL Out of Context" thread for comments like this. (They're deliberately pulled out there.) I'd almost want to see an "ACX Out of Context" subthread. Almost.
Not relevant: OP stated that they'd be watching The Matrix Revolutions. Going anywhere near relativistic speeds would introduce time dilation that would make that film last way longer than is humane.
I've written a piece on 'Oumuamua recently (https://www.thequantumcat.space/p/oumuamua-alien-spacecraft-or-dark). Avi Loeb argues it was probably a piece of alien space trash. The rest of the astronomical community is sceptical and has rapidly coalesced around the idea of a "dark comet". We saw no signs that it was a comet when it flew past us, but recent papers have argued it probably had an invisible hydrogen tail. Loeb, meanwhile, says the dark comet idea is bullshit and physically unsound. Since 'Oumuamua is now far beyond Pluto we'll never know for sure, but it is fun to speculate.
There is a pattern similar to "God of the gaps" that one could call "Aliens of the gaps", which is a secular version of same, with just as much credence that should be assigned to it.
Good article. In addition to what was mentioned with unaccounted acceleration, I was under impression that the long shape of Oumuamua was unstable and very likely to break up. Does the dark comet hypothesis with unseen hydrogen propulsion address that aspect? Or is that more of a red herring? thanks,
The argument is that the dark comet would have eroded as it flew through the solar system. We found it quite late into its voyage, so we would have seen the eroded form, not the original form. It may be that that shape is actually unstable, and Oumuamua could have broken apart after we saw it.
14 % on Ukrainian victory (up from 13 % on April 17).
I define Ukrainian victory as either a) Ukrainian government gaining control of the territory it had not controlled before February 24 without losing any similarly important territory and without conceding that it will stop its attempts to join EU or NATO, b) Ukrainian government getting official ok from Russia to join EU or NATO without conceding any territory and without losing de facto control of any territory it had controlled before February 24 of 2022, or c) return to exact prewar status quo ante.
45 % on compromise solution that both sides might plausibly claim as a victory (up from 43 % on April 17).
41 % on Ukrainian defeat (down from 44 % on April 17).
I define Ukrainian defeat as Russia getting what it wants from Ukraine without giving any substantial concessions. Russia wants either a) Ukraine to stop claiming at least some of the territories that were before war claimed by Ukraine but de facto controlled by Russia or its proxies, or b) Russia or its proxies (old or new) to get more Ukrainian territory, de facto recognized by Ukraine in something resembling Minsk ceasefire(s)* or c) some form of guarantee that Ukraine will became neutral, which includes but is not limited to Ukraine not joining NATO. E.g. if Ukraine agrees to stay out of NATO without any other concessions to Russia, but gets mutual defense treaty with Poland and Turkey, that does NOT count as Ukrainian defeat.
Discussion:
This update is brought to you by a cheap natural gas.
Price of natural gas on Dutch TTF exchange, which largely determines gas and thus electricity prices over the EU, over the previous week dropped back basically to pre-covid levels. You can follow it here: https://tradingeconomics.com/commodity/eu-natural-gas. I think it is probably caused mainly by weaker than expected (by me, anyway) economic recovery in China.
My very crude model goes something like this: Ukraine probably cannot avoid defeat without substantial Western support - > this support is constrained by how much public pushback/support is there against/for it - > magnitude of the public pushback is dependent on economic costs of support - > cheap natural gas means lower costs - > which means higher political support for Ukraine.
I also think, however, that low prices are dependent on the fact Russian energy exports are still substantial (although no doubt much lower than they would be without sanctions), and further attempts to squeeze Russian economy would again prop them up. Given that current level of sanctions and overall support to Ukraine is imho probably insufficient to safeguard Ukrainian victory, I’ve updated only modestly (alos because low price might not last, of course).
*Minsk ceasefire or ceasefires (first agreement did not work, it was amended by second and since then it worked somewhat better) constituted, among other things, de facto recognition by Ukraine that Russia and its proxies will control some territory claimed by Ukraine for some time. In exchange Russia stopped trying to conquer more Ukrainian territory. Until February 24 of 2022, that is.
Over the past twenty-four hours we've gone from Russia invading Ukraine (very likely permanently), to Ukraine invading Russia. Since your last update, we've also got the UK giving them Storm Shadows and the US authorizing F-16s as soon as the pilots can be trained. And all that only increases the probability of Ukrainian victory by 1%?
Of course I didn't predict exactly that. But it is, like, within the general class of expected events. In, I think, February, I did alter my prediction after it was decided to deliver Leopards to Ukraine, but after that, current ramp-up of support was something to be expected.
I think recent developments - Russian offensive running out of steam without much to show for and Western decision to provide planes - should mean a couple percentage points more for Ukraine.
Regarding the gas prices, it's also possible that the acceleration of build-out of LNG terminal capacity plus warm winter has resulted in both high stocks and greater confidence in maintaining adequate levels in the future.
I've enjoyed reading these and would be interested in seeing all the predictions to date in a single place, plus perhaps abbreviated summaries of the reasons behind the changes. I'm particularly curious about the reasons for the biggest deltas, and in the reasons proximate to any long-running directional changes in your estimates.
That's interesting. I was just asking myself, what's the point of this all?
I mean, it is nice to assign probabilities and publish them, because then you create a public record... of being good at predicting or being bad at predicting. And if you make 100 public predictions, and then evaluate the outcomes, you can even make a nice graph out of it (predicted probability vs the actual frequency of outcome).
But making *one* prediction and then updating it many times is something quite different. There is no way to evaluate whether the prediction is good or bad. Suppose that one year later Russia wins. What exactly does it say about how good Ales is at predicting? Suppose that one year later Ukraine wins. What exactly does it say about how good Ales is at predicting? In my opinion, almost nothing. If you give something 14% or 40% that means it can go either way, and even if the thing with predicted 14% actually happens, I wouldn't really count it as a mistake. The calibration is in numbers -- if you make ten predictions at 14% each, and 9 out of those things actually happen, then I might laugh at you. If you make one 14% prediction, and the thing happens, there is nothing to laugh at; you already said that this is supposed to happen in 1 out of 7 situations.
Furthermore, predicting something with 1% precision (e.g. updating a probability of Ukr victory from 13% to 14%) is an implied claim of very strong prediction abilities. I would rarely dare to predict a real-life outcome with precision better than 10%, or maybe 5% if I feel exceptionally confident. Then again, I am not really good at predicting, so it is plausible that other people are way better than me at doing this. But as I said, there is no evidence here whether Ales is actually good or bad at predicting. Maybe his predictions are amazingly correct given all publicly available evidence. Or maybe they are laughably wrong. How would we even tell the difference?
For myself, the war is rather salient to economic situation in Europe, so I've allocated my own portoflio partially based on what I think will happen (so, yea, mine non-Japanese Asian stocks are underperforming, see prediction above). Perhaps that might be useful for someone else.
I agree that it is not a good way to tell whether I am good or bad at predicting. But I can tell that by other means - for instance, my prewar prediction on what'll happen was, I think, pretty good, although it was less quantitative and also I didn't bother translating it to English and posting it here. And to pick perhaps more measurable example, I got to 85 % in Scott's 2022 prediction context (I just looked it up, results are here: https://astralcodexten.substack.com/p/who-predicted-2022).
A petition that seems important, unfortunately those with poor critical thinking skills haven't questioned the assumptions the media makes about the issue, vs. basic logic and history:
"Stop Domestic Voter Abuse, Restore Fair Elections: Ban Mail Voting (with at most extremely limited exceptions) in every state.
The secret ballot is recognized globally as a human right: but a vote cast by mail does not protect the voter's secrecy in a safe space like a polling location. Behind closed curtains, so no one knows it is happening, a vote can be coerced by domestic partners [ the CDC says 49% of women and 46% of men experience coercive control by a partner in their lifetime, would voting be an exception?], employers or unions or others and can tilt elections, as evidence from around the world, history, and logic explain.
Human nature has not magically changed to prevent an exact repeat of problems this country suffered through before.
Voting was not secret for this country's first century; votes were cast publicly. That led to the growth of such widespread problems, including violence and deaths, that in 1889 President Grover Cleveland declared they had become: "perils which threaten the existence of our free institutions, the preservation of our national honor, and the perpetuity of our country." Most states adopted them over the next few years.
It is crucial in this toxic political era that we prevent an internet accelerated repeat of that era with an internet accelerated campaign for reform to restore the secret ballot by banning mail voting (with at most extremely limited exceptions if no other approach can be found). You need to pass this along to others and help them grasp why this issue is important.
For detailed logic and cites supporting these concerns, see TheBigIllusion.com . People have blind spots preventing them from taking the problems seriously. That site addresses those and can help you persuade others. It also addresses other major problems with our democracy hiding in plain sight: but you do not need to agree with anything it says to agree we need to Stop Domestic Voter Abuse and Restore Fair Elections by Banning Mail Voting (with at most extremely limited exceptions). "
I'd theorize that one of the advantages of mail-in ballots is that corruption is more visible. That is, it's obvious on a small scale, and also more easily detectable if done on a large scale. As long as society maintains the ideal that the ballots are filled out and sealed in private, I'd hope that we'd know to keep an eye out for:
-- people offering to fill out ballots for you
-- people wanting to watch when you fill out your ballot
-- people collecting ballots to "deliver" them
I'm not 100% optimistic here. I've seen explicitly partisan organizations running booths on the street to help people fill out ballots. And I worry when I see people make claims that imply that civil democracy is a problem when it doesn't lead to the "correct" outcomes. But overall I'm happy with the mail-in ballots that my state has.
How is "corruption more visible" when its done behind closed doors? Your viewpoint was the common one for a century before the adoption of secret ballots based on the reality that human nature didn't live up to wishful thinking.
I'm guessing that you approve of the politics of the winners, therefore mail-in ballots are less of a concern. Except when the general public finally grasps that they an election based on them can't be trusted, the political turmoil of prior elections seems likely to be minor compared to what goes on before they are fixed.
I agree that this is one potential worry about ballots people fill out at home. It just doesn’t seem clear to me that it’s worth the much *bigger* worries of inconvenience provided by requiring people to go to a specific place at a specific time to cast a vote.
Inconvenience is a bigger worry than depriving people of their right to a vote due to being coerced? "Sorry domestic abuse victims, we don't care enough about elections to bother going the polls in person since its inconvenient, but despite that we think our ability to vote conveniently should deprive you of your right to case a secret ballot securely?"
Its a bigger concern than the reality that we can't know if the level of problematic votes is large enough to be changing the results of close elections and therefore we can't know who would have won a fair election? Inconvenience is more important than the potential that we could have widespread voter coercion and other problems arise just like they did before secret ballots existed?
Its unclear how people think we can magically know it won't grow just as it did back before secret ballots existed at any time. Social media didn't exist back then, trends can spread far faster today than back then. Human nature hasn't changed, so merely because problems aren't visible to the public yet the way they were back then, doesn't mean it can't happen quickly with nothing preventing it.
Yes. Inconvenience takes away millions of peoples rightful votes. That is a bigger concern than a few thousand people being inconvenienced by a spouse.
A few thousand? Seriously? We aren't turkey, but the number there was 20% estimated who had a their vote coerced by a family member, thats in the millions.
There are 80 million non-voters who don't care about their ballots, and millions of rabid partisans concerned the other side's candidate being "evil": is there some rational reason to expect it won't finally dawn on them to get the ballots from their friends and family, "hey, I'll buy lunch next time if you'd click this to get a ballot and sign it when it gets here for me.." or whatever. People break lots of rules when they don't expect to be caught since they aren't going to tell anyone. Can you prove it didn't happen?
Maybe it hasn't: can you prove why it won't? Why is it that people would assume if millions of students took an online home test without being watched they'd cheat, but magically all these people out there are honest when it comes to voting? Is it wishful thinking, or that because no one in the media talks about this that it can't possibly be logical since the media would have raised the issue? What is the logic and evidence regarding human nature that makes people think this way?
Sounds like another elaborate scam to restrict voting by those who already struggle with existing restrictions. The obvious intention is the exact opposite of fairness, and "think of the abused" is yet another convenient excuse.
Unfortunately many privileged people take for granted their ability to safely vote the way they wish in secret and ignore the plight of the powerless who can't do so safely at home or work, and ignore the way their votes are being suppressed on a massive scale.
Those who bother doing a little research and thinking can find ways to address the issues of people who struggle to get to the polls, as I noted in other comments. Other countries do so, as did this country in the past before people decided to engage in magical thinking that people are somehow 100% trustworthy at home and work and would never cheat, coerce or buy ballots: despite lots of history and evidence that isn't remotely the way humans behave except.
Voting issues is a standard right-wing talking point, just like race is a standard left-wing talking point, both camps are trying to use them for getting/staying in power. My guess is that you are from the right-wing camp, pretending to care about voting fairness, but really interested in your in-group's power.
"unfortunately those with poor critical thinking skills haven't questioned the assumptions the media makes about the issue, vs. basic logic and history"
Ah yes: "anyone who doesn't view this as I do has poor critical thinking skills."
I am hoping that you are not an eligible voter anywhere near where I live.
sorry, but the site (and some comments below) point out exactly the issue I was referring to but didn't specify that most of the American public has fallen for, and I figured rationalists would be concerned about. In a short post I decided not to put the details into that post, perhaps that was a mistake. If you disagree: feel free to make an actual logical argument, but I suspect the odds are you cannot. Its those who don't provide actual logical arguments (I linked to them even if I neglected to get into every detail in a post) that we should be concerned about.
Are you capable of any actual argument about the topic? Or do you just assume you are right because of what the mainstream media has told you and anyone who has a different viewpoint and must be apriori wrong?
I think secret balloting makes the system open fraud. Secret ballots assume the votes are entered as voted. How to check the assumption? With an audit. How can you audit secret ballots? As far as I can see, this is impossible.
Nowadays I think it would be better if everyone could verify their vote on a central web site, and have (possibly more than one) independent auditor organization verify the votes, possibly with something like double-entry bookkeeping. It would still be possible to display a different vote to the user than is recorded, but I would think that could easily be caught examining the system.
You check by counting the votes again to see if it's the same number, and by making sure there's multiple people looking at each ballot, no? It's anonymous and secret but there's a physical record in the form of a bit of paper.
Sure, an individual couldn't do it but a relatively small group of volunteer observers could check that all the votes (therefore all of theirs) were counted correctly.
Tech folks are too focused on the tech issues rather than the human ones.
With a verifiable vote that repeats the exact same problems that existed before the secret ballot that lead to its adoptions. Verifiable votes allow coercion to ensure someone voted the right way. There is a reason that secret ballots are considered a human right globally, after people tried the approach without them and discovered the issues.
This is a valid point, but it is in conflict with the tabulation of votes. "Those who vote decide nothing. Those who count the votes decide everything." Apparently Joseph Stalin didn't say exactly that, but him and others through history have expressed similar sentiments.
EDIT: So I guess the question is which one people find more important? On the one hand, you can be coerced to vote the "right way". On the other, you only THINK your vote matters.
Well if you mention Stalin, then you must know that there were multiple instances where bolshevists demanded open voting instead of secret voting (I don't mean general election here).
There are two things necessary for a fair result to be determined: fair counting and fair casting. Its like the streetlight effect: the counting is in public view so everyone focuses on that and ignores what happens in the shadows where the problems may be.
The process need to be made credible that when a vote enters the system it is counted. I suppose if there were some way of separating out the existence of a vote virtual envelope in the system separate from its contents that is handled secretly by another process, there might be something possible. The big thing now is dealing with the issue being ignored in the shadows, everyone is at least thinking about the counting issue.
There's an actual problem with mail-fraud elections that's vastly worse than the problem of people who find it difficult or impossible to vote in person. No country but the US uses mail-in ballots, for that reason.
There's no particular problem with mail-fraud in elections, and mail-in votes are reasonably common at least in European countries.
You register to vote beforehand with all the necessary formalities, then at home you put your vote in an envelope, and put that envelope inside another envelope together with your registration confirmation. In the election office, they keep the confirmation documents and then pass on the anonymous letter with the vote to the normal vote-counting process. Sure, there's room for fraud here, but it's not as simple as pouring additional ballots into the bucket and claiming that these all were mail-in votes.
There are creative ways to address the minority that find it difficult to vote in person, like multiparty teams of volunteers that go to them turning their backs while they vote, of if they need assistance at least lowering the risk of one-sided coercion. Worst case a limited number of mail votes for when there is no other approach.
When those have been implemented, we can talk about restricting the right to vote by mail. At the moment, it's clearly a necessary part of undermining the deliberate Republican attempt to make it hard for blacks and other majority-Democratic groups to vote.
"A solution to the problem this is mitigating is theoretically possible in the future, so we can remove it here and now" is not a good argument.
I'm not so sure. If voting has systemic fraud, then surveillance of the voting itself won't accomplish anything. That's more likely to ensure voters don't commit fraud, which I think is less of a concern with the current system.
In terms of "systemic fraud": there were 80 million non voters last time who don't care to use their ballot. How many of them might be persuaded by a friend or relative to give them a signed empty ballot to use, if word spreads to do so? Studies show a substantial minority would sell their votes. A substantial minority of the populace is willing to use violence for political goals if needed, this would be a mild bending of the rules.
It seems that trust in the current system is based on naive wishful thinking about human nature.
Domestic voter abuse is not the only problem our voting system has to solve, and mail-in ballots solve some very big problems - people who are too old or sick to get to the polls, people who are traveling out of state, people who don't have good access to transportation or can't find time to go to the polls, etc. etc. You're arguing that all of these people should be thrown under the bus to protect against a very specific form of voter fraud - domestic voter abuse.
But you don't actually offer evidence that domestic voter abuse is a large-scale problem. You argue that domestic abuse is a problem in general, so voter abuse *could* be a large-scale problem, and you argue that other forms of non-secret ballot have caused problems in the past, and then you make a huge flying leap to argue that domestic abuse is such a huge problem that mail-in voting should be abolished except for unspecified "extremely limited exceptions." I would like to see some better evidence before throwing all the access-to-democracy problems under the bus.
Also, fuck you for pre-emptively accusing everyone who disagrees with you of having poor critical thinking skills.
Other countries solve the problem of voting out of country by letting people vote at embassies, within the country there can be voting at other jurisdictions that provide secrecy. Volunteers can go in multiparty teams to invalids and turn backs while they vote, or at least minimize risk of one sided coercion. It noted that there can be limited exceptions if people are really incapable of inventing alternate solutions.
The site it links to goes through the data and evidence and logic you fail to note. It points out that most arguing in favor of mail voting are like those who fall for this:
"Imagine a government-created online test that 10 million students take from home with a $500 reward for passing. The tests are unsupervised. The program manager — who has the same mindset of those in the 1950s who ignored what happens behind closed curtains — proclaims, “Of course no one cheated! Where is your evidence someone cheated?”.
Would you accept that no cheating would occur? It sounds absurd, doesn’t it? The manager might argue for the students’ innocence until proven guilty. But this is only relevant to individuals. If someone accuses John of cheating on his test, John is innocent until proven guilty.
We can’t accuse specific students without evidence. But we can challenge the validity of the entire system. The inability to prove any specific student cheated without evidence doesn’t prove the validity of a claim that no one cheated—an assertion that defies reason"
Is there a *significant* amount of cheating? Enough to be a bigger problem than all the people we can be *sure* will not vote without mail-in ballots?
You're proposing to add rules that will reduce cheating, but also admitting that you don't know how much cheating has happened, or how much cheating these rules will prevent. "Nobody has ever cheated a mail-in ballot" is absurd, but so is "We should add new rules that will reduce cheating by an unknown and unmeasurable amount."
(I have a magic rock that reduces cheating by an unknown amount. I haven't checked if it works, but I think we should mail out a rock with every mail-in ballot, just in case it helps.)
Can you give me a ballpark figure, at least? How many ballots in the 2020 election do you think were changed due to abusive spouses? A million? A thousand? A hundred?
Priors about human nature and other data indicate there will be cheating, evidence from other sources suggest it could be 20%, or at least a few %, easily enough to tip elections (sources at that site TheBigIllusion.com), in addition to the non coercive cheating that can go on that there isn't yet data on.
Its like this statement: "Imagine a government-created online test that 10 million students take from home with a $500 reward for passing. The tests are unsupervised. The program manager — who has the same mindset of those in the 1950s who ignored what happens behind closed curtains — proclaims, “Of course no one cheated! Where is your evidence someone cheated?”."
Priors and human nature and other data suggest it not be trusted even if there is no concrete data on the frequency. It strains credulity to think otherwise and anyone who sided with the manager for claiming no cheating occurred would be viewed as cluelessly naive. Widespread mail ballots are new and voting happens once a year,or every few for major elections: it isn't easy to collect data on something everyone wishes to keep secret.
The burden of proof should be on those that pretend that somehow despite the history of problems before the rise of secret ballots, somehow human nature has changed and those problems are magically not guaranteed to happen again. Its like someone saying "the password to the bank login to handle their funds was posted online as 'password', but of course no one would ever use it so we don't need to change it". Priors and basic logic suggest otherwise.
And embassies are rare, you may not even have one in your country of residence. Colleagues of mine can only vote in their countries' elections in the embassy and just going there and standing in line for an indefinite amount of time makes this a whole-day endeavour.
There is a difference between widespread voting by mail, vs. extremely limited cases. re: 1. I already noted .Volunteers can go in multiparty teams to invalids and turn backs while they vote, or at least minimize risk of one sided coercion
re: 2, there can be multiple days of in person voting. in the US there can be cross jurisdiction collaboration. Other countries do requite people out of the country to vote at consulates or embassies.
Edge cases can be addressed, without ignoring the lessons of history and pretending "no, it can't possibly happen again, people are 100% trustworthy". If some people are restricted, its better than having an election which is meaningless since it can't be trusted to represent what a fair vote of the public would come up with.
I can't understand why no one has tried to solve this with bitcoin yet. For a while now I've been expecting a small team of mathmoes to publish a system that uses cryptography this and blockchain that and allows everyone to publicly vote, check that someone else has voted exactly once, and tally up the votes thus confirming the results for themselves, all while keeping their actual voting choice secret. I would also expect the process, which guarantees openness and transparency in all future elections, to be so involved and complicated to carry out that only maths phds can actually do it. But I've not heard so much as a conversation gesturing in that direction. The Internet has really let me down here.
That doesn't address the human factors issues at the root of the issue of voter coercion. Most people in forums like this can safely vote in secret themselves and ignore those who can't. Crypto tech doesn't solve this issue, privacy in person solves it. (at least temporarily, there are complications later, but one step at a time).
What voter coercion are you imagining here? I'm struggling to think of any way voters can be coerced as long as voter anonymity is guaranteed. Can you give an example of the kind of stuff you're thinking of?
Vote counting is the bigger part of the problem. In Putin's Russia, you don't have an election then wake up to find the secret police at your door because you voted wrong. You wake up to find that you and all your friends somehow accidentally voted for Putin again.
Huh? as the original post noted, domestic abuse is widespread, including coercion by domestic partners, and evidence shows its enough to tip elections. There are 80 million non voters who didn't bother to use their ballot, why wouldn't many of them give it to a family member or friend who is very political , even as a favor without coercion? Polls show a substantial minority of the public is willing to consider political violence these days, people who commit domestic violence are somehow going to draw the line at coercing a vote? Its unclear where people come up with this magical thinking about human nature that everyone can just safely vote at home without coercion there or at work. Merely because it doesn't happen to the sort of people who read and post here doesn't mean it doesn't happen.
There are such systems. I just wrote a comment elsewhere, moving it here:
I do not remember the exact details, but the idea was that in the voting room you choose one of N options, and the computer prints for you N+1 paper cards, one vote for each option, plus one extra vote for the option you actually want. You drop each of these votes in the box (it is verified that you dropped N+1 of them), and keep a receipt for each.
It is possible to count the votes, e.g. when you have 100 voters, and three options that got 120, 130, and 150 vote papers respectively, you know there is one fake vote from everyone for each option, so the numbers of actual votes are 20, 30, and 50.
All paper cards are in a publicly available database (the random id, who the vote was for), you can check that all your receipts are there. -- If they are not, you take *one* of the receipts that are not there, and complain about election fraud. One receipt does not show who you voted for, because there is one fake vote from everyone for every option.
This is the kind of thinking I was after! Where and when was this discussed? Did they proceed to the next logical step, which is building an online model, holding iterated toy elections, and inviting people to try breaking the system?
I've heard tech geeks talk about this. Nothing you key into a computer that is tied to your personal ID can be guaranteed to remain anonymous. There's always a chance that someone clever enough can figure out how to trace a user's input back to their ID. For now, blockchain and crypto can be used anonymously in very limited use cases; but there's no way to safeguard even those limited cases from future tech developments.
For voting, my Taiwanese friends say their country does it best. Everyone votes anonymously on paper, and counting is manual. The counter holds each vote up to a public camera so everyone can see the vote on TV. There's very little room for a counter to discard a ballot he doesn't like, run ballots through a tabulator multiple times, or tweak complex software to make it miscount or weigh votes differently.
How well does that scale? I saw some news articles about counties that tried switching to pure hand counts after 2020 and discovering it takes a lot of manpower - you need multiple counters so they can check each other, the counters get tired and make mistakes after thousands of ballots, etc. Machine tabulation followed by hand counts to spot-check seems like a reasonable middle ground.
Possibly we could alleviate these problems by not trying to get results finalized by election night. When does Taiwan announce election results?
Yes, there are long delays... I agree that's a downside. On the upside, most of the multiple count-checkers are volunteers, at least in the US (not sure how they do it in Taiwan). Any party is allowed to have at least one challenger observing a count. Theoretically, a hand count would take no longer for a large county than a small one, assuming that the count is done on-site at each voting location, which would be staffed by its own counters and challengers. So for giant counties, there'd be stronger incentives for breaking up voting location districts into smaller & smaller units, to enable quicker counting at each. This used to work in the 1900s before tabulators, even with the US's increasing population over 200 years, so on paper at least it should work again.
I find this post misaimed. Most people are *already* in favour of the secret ballot. The load-bearing part of this argument is whether mail-in ballots can successfully remain anonymous or not; I think most people just assume yes. That's what you have to challenge if you disagree; not the importance or lack thereof of anonymity in itself.
That is the point its making, even if its misnamed, the issue of restoration of the secret ballot which most people haven't realized has been undermined. Most people can safely cast their vote secretly and misguidedly assume everyone else can to.
Sorry, its just that its frustrating seeing most of the public falling for the equivalent of this argument where the burden of proof is put in the wrong place (that site goes into it):
"Imagine a government-created online test that 10 million students take from home with a $500 reward for passing. The tests are unsupervised. The program manager — who has the same mindset of those in the 1950s who ignored what happens behind closed curtains — proclaims, “Of course no one cheated! Where is your evidence someone cheated?”.
Would you accept that no cheating would occur? It sounds absurd, doesn’t it? The manager might argue for the students’ innocence until proven guilty. But this is only relevant to individuals. If someone accuses John of cheating on his test, John is innocent until proven guilty.
We can’t accuse specific students without evidence. But we can challenge the validity of the entire system. The inability to prove any specific student cheated without evidence doesn’t prove the validity of a claim that no one cheated—an assertion that defies reason"
I suspect the vast vast majority of individuals do not value a vote at $500, which means there is far less incentive to "cheat" by taking someone else's vote, than in the scenario you are laying out.
Seriously, if this is happening at scale, why is there *no* instances of people saying "My spouse who abused me in X Y Z ways, and who I have now escaped from, also convinced me to vote the way (s)he wanted"?
Even without coercion: there are people that would pay $500 per vote to overturn an election, and like the drug trade black markets are hard to be aware of or stop.
In terms of the domestic violence issue, there are such stories, the site links to them and to data, limited as it is. Domestic violence is year round and yet that is difficult to study, vs. something happening once a year. People don't report actual violence, this is something they wouldn't even think to report.
This is like hearing from people in the 1950s who doubted domestic violence, human nature is a bit different than you seem to imagine.
A substantial minority of the populace is willing to consider political violence, are they really not going to coerce a vote? Again, this happened in our history and was widespread, and its unclear why anyone thinks human nature has magically changed to prevent it.
What would make you consider immigration to another country? One where they don't speak your native language? Would this answer change if you had kids?
I think it’s a bit odd to associate the national border with the linguistic border. For many people, one would be a border that the other would not. Moving between Toronto and Chicago is in many ways easier than moving between New York and San Juan, or Toronto and Quebec City.
In any case, I would far sooner move to a big cosmopolitan city nearly anywhere than to a small town without a cosmopolitan university! National borders are an issue, but they are easier to overcome than many cultural borders within a nation.
Well National border are very important from a bureaucratic standpoint. You can't just hop over the border and find a job and a place to live. You need a Visa, a work permit, etc.
The process to get those things can be long, expensive and uncertain - and those deter a lot of people from moving across national borders.
Math academic, so employment offer like Quiop, but I'd also have to be single (which I'm currently not)/have a partner willing to move. I'm confident in my lack of ability to mimic my mindset if I were a parent.
So far, I have moved to three different continents (other than the one where I was born) for education and employment. Since marrying and becoming a parent, I have lived in three countries (speaking three different languages, one of which was my own native language and the other is my partner's) on two continents.
As you note, the "two body problem" can be a serious issue for academic partners, although we have so far been lucky to be find jobs in the same cities. (Being able to do so is a major motivation for being willing to move countries.)
Same here. Would try to move even before that actually happened if the risk was high; just in case that the new government might close the borders immediately.
At this moment, I see no risk, but I insist that all family members have valid passports, even if we are not traveling anywhere. Because I assume that if something bad happened, the line to get a passport might suddenly become very long.
In the past: better living conditions, like better work / more money / better social life / living with my partner. Preferably to a country with a culture I have some affinity to.
Now: living with my partner.
With kids (currently don't have any): Social support network, and ease of living with kids. Probably wouldn't consider learning any really challenging languages for this.
Anything that would make this a rational idea would compel me to take up arms to fix whatever is going on in my native country, even if this were a mostly suicidal course.
I do have kids and they are now out of the house -- this answer might have been different when they were younger.
There's been lots of "book banning" recently, according to many headlines, although I can never fully grasp what "book banning" means according to the stories. It always seems to involve removing books from school libraries. Apparently there was a case in the 1970s which went all the way to the Supreme Court, which found for a group of students from Long Island who had sued their school, claiming that their civil rights had been violated by the removal of some books which had been deemed "obscene" or something by the school board. (I believe one of the books was Slaughterhouse-Five.)
But there has also been many cases in recent years where books deemed "offensive" have been removed from mandatory school reading lists. Those books tend to be old classics like Huckleberry Finn, which features the "n-word" prominently, which now offends people out of sheer spelling. Apparently removing those books from mandatory assignments isn't considered "book banning" because those books remain accessible to students in the school libraries.
I get that there is a difference between a book being required reading for a student vs. a book which is merely available in the library. However, I don't understand how the removal of books from a school library could violate a student's civil rights anymore than removing a book from a mandatory reading list does. Aren't school libraries rather choosy about which books they acquire in the first place? Do many school libraries stock old joke-books full of racist and homophobic jokes? Probably few do. So how is the removal of a book from a library qualitatively different from a refusal to stock a book in the first place?
I'm not in favor of actual book bans, which I consider to be when a government bans a book from publication or sale. But until every school is willing to stock Naked Lunch, The Anarchists Cookbook and Truly Tasteless Jokes: Volume II on its library shelves, it's hypocritical to call the removal of any book from its shelves "book banning".
Yeah, I think this is one of those "Neutral vs. Conservative" struggles, https://slatestarcodex.com/2017/05/01/neutral-vs-conservative-the-eternal-struggle/. Libraries are one of those institutions that have become very left-aligned, conservatives have noticed, and they're pushing back against it with the tools available to them, which are basically entirely "external" things like bans.
Both sides in practice are trying to control what the next generation reads (I think the Russell conjugation here is probably "you censor books, I protect children from harmful ideas"), but when a book is offensive to the left, they don't have to do anything as public and as gauche as banning it, they just don't shelve it. (Assuming it's too late to go harass the author on YA Twitter to make sure it isn't written in the first place)
Is one side worse than the other? Sure, probably! But the asymmetry of the situation makes it hard to make direct comparisons. The right-aligned stuff is necessarily out-in-the-open and reported on, while I think things like YA Twitter and the Dr. Seuss controversy are just the tip of the iceberg on the other side.
I work for an organization which advocates for freedom of speech in K-12 schools, so I am pretty up on this.
The Supreme Court case you reference was Board of Education, Island Trees Union Free School District No. 26 v. Pico, 457 U.S. 853 (1982), in which a plurality of the Court said that public schools cannot remove books from libraries because of hostility to the ideas expressed therein, but can remove them for other reasons, such as vulgarity. Decisions re what books to buy and what books to assign in class are different, in my view, because requiring a school to teach or buy a book that, for example, says "the free market is bad" limits their ability to send their preferred message that "the free market is good." Every dollar spent on the former, or time teaching the former, prevents the school from spending that dollar/time on the latter. In contrast, the presence on library book shelves of a book that says "the free market is bad" does not prevent the school from sending the message that "the free market is good," but rather merely competes with that message. Given longstanding recognition that 1) schools may try to inculcate certain values and ideas, but 2) that they cannot "prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion", W. Va. State Bd. of Educ. v. Barnette, 319 U.S. 624, 642 (1943), that distinction seems to me to make a lot of sense.
Note however that Pico was a plurality opinion, and hence is not technically binding, though many lower courts have adopted the reasoning of the plurality.
Also, you say that "here has also been many cases in recent years where books deemed "offensive" have been removed from mandatory school reading lists. Those books tend to be old classics like Huckleberry Finn, which features the "n-word" prominently, which now offends people out of sheer spelling." However, those cases are comparatively rare nowadays; over the last few years, the vast, vast majority of books that have been removed have been because of sexual material or profanity. See, eg, https://www.ala.org/advocacy/bbooks/frequentlychallengedbooks/top10 A distant second is what could very broadly be described as "CRT" or CRT-adjacent, though again that would be books in the curriculum, rather than libraries.
Finally, I certainly agree that using the term "book ban" to describe what is happening is not the height of intellectual honesty.
What would your take be on declining donations? Certainly there are space constraints in a library (they could not accept 10000 new books on communism for example), but with the promise of sufficient funding to expand/support their library services should they be obligated to accept any donation?
My view on removing books from school libraries is that as long as the book is allowed at school (if brought from home), it's not a hill Is be willing to die on. I think that parents probably have some degree of interest/rights in which books their children are allowed to read, and unlike at a public library they have less control at a school one. If a public library banned a book I'd be getting out my petitions.
Are you asking my legal take, or my policy take? My policy take is that the library should have as wide a range of ideas represented as possible, so they should accept the donation if it comes with funding to expand. My legal take is that, under current jurisprudence, the school is probably free to refuse the donation, just as a city can refuse the donation of a monument in a public park https://en.wikipedia.org/wiki/Pleasant_Grove_City_v._Summum
It is true that parents have some interest/rights in what their children can read, but the problems are 1) children have some rights independent of the rights of their parents; and 2) removing a book from a school library means that no child can read it, including children of parents who want their kids to have access. Which is why this is actually a rather complex issue. And suppose some woke school district wants to remove all books that support conservative ideas, or even, say, a complimentary biography of Ronald Reagan? Or all econ books by Milton Friedman, et al. I am sure the majority of parents would love that, but what about the parents who want their kids to have access to those ideas?
Thanks, I was curious about both. To your last point, this was my point about public libraries. To me the ideal solution would be that books would just be moved from the school library to a public library so that parents could still check it out for their students.
Re: students having rights. My middle school staged a peaceful protest, and I quickly learned that we indeed lived in a totalitarian state at school. I have since given up on the dream of students having rights; while I support other dreamers, I think that that dream is forever dead to me.
There are a number of fundamental problems when it comes to dealing with public education:
1. Children have different growth needs; some need a lot of firm guidance while others need the freedom to explore.
2. Parents have a massive investment in their children's future, and thus have a right and responsibility to act in their children's interest to shape their values and education.
3. Some parents are not good parents and do not act in their children's interests.
4. Most people don't have the resources necessary to chose an option other than public schools.
5. Public schools are taxpayer funded, and thus responsible to the governments which fund them.
6. While a good teacher will act in the interest of the students, teachers do not have the same level of investment in the children that parents do.
7. It's very hard for a bureaucracy to accurately tell who is a good parent or a good teacher.
8. Teachers and school administrators are not necessarily representative of the community at large.
9. Anything currently politically charged is that way because there are rational arguments on both sides.
10. Schools cannot teach everything. There is a limit to the educational time and space (such as the size of a school library).
Ultimately, if the school systems can't meet the needs of the population that is under-represented in education, the public will correct this via the legislature. Having the state legislature as a check on the school system is a bit brute force, but the legislature is elected by the public and the legislature's actions are mostly transparent. School boards are closer to the level of the public, but are much less transparent.
The long term fix would be some form of school choice.
A different but related question: how practical are these school censorship debates? With the Internet widely available, there are lots of older classics on standardebooks.org, for example. Also, lots of other Internet resources. Are school libraries still relevant or is this just an old-fashioned debate by people who don’t realize how much has changed?
I imagine that more practical barriers are not knowing stuff exists because adults don’t promote them and many kids being more interested in Minecraft and social media than reading books anyway. Also, maybe Internet access restrictions?
But what do I know, I’m old enough to remember carrying books around for entertainment like people nowadays carry a phone. It would be good to hear about it from people who were in school more recently.
School libraries are critical in that they are a space in which a child can access information without their parents being informed and having veto power over what can be learned. This makes them highly corrosive to any ideology which relies on information control to retain members, including most cults.
Moreover, many school assignments are given with the instruction that the student ought to consult either a specific book or their choice of books from the school library. If a student is told to research the history of the Spanish American War, but the school library cannot legally contain any books which discuss the history of American imperialism, the student is going to come away with a wildly incorrect picture of why the war was fought.
Perhaps you have very little familiarity with children or parenting?
As a parent you have fairly strong control over how and when your child can access the internet unless someone is actively invested in undermining that control.
Pretty easy to simply not get your kid a smartphone, or prevent a handful of social media apps from being installed, and thus cut off almost all influences you wouldn’t want.
Obviously this control attenuates with age, but by the time your child is a middle to late teenager, you’re probably less concerned with total control and more concerned about making sure that when inevitable encounters with the previously censored material occur, your child has already had enough context from you to make good choices about viewing it or fleeing from it.
The distinction is that the book bans being objected to forbid the school library from stocking the books even in the event that the school library wishes to do so. It's one thing if an individual school happens to not have any copies of the Quran; it would be another thing entirely if there is a legal ban prohibiting any school in the state from possessing a copy. Many of the laws (including Florida's HB7) are structured such that entire academic positions (ie. essentially any admission that systemic racism exists in modern day America) are banned from being expressed in any book in any Florida classroom or school library. That's obviously an unacceptable degree of censorship, well beyond a school simply declining to stock a particular book because they decide they don't want/need that particular title.
"(ie. essentially any admission that systemic racism exists in modern day America)"
Large numbers of Americans believe in similarly unfalsifiable claims like Jesus being the son of God but have (often grudgingly) accepted the fact that they aren't allowed to preach those ideas to a captive audience in public schools. You can too.
Wait - have they accepted the fact that it’s illegal for a school library to stock Bibles, just like it’s illegal for a school library to stock books about gay characters? I agree that teachers shouldn’t tell students to be gay or to be Christian, but to ban the book from even being in the *library* seems pretty extreme.
That is not a condition of the new Florida law. Its mindless keyword-ban makes no distinction between classrooms/libraries which present systematic racism in the manner that you describe vs those who present it as something to discuss/debate vs those who present it as a false claim.
In a society genuinely devoted to freedom of expression, the right response to shitty speech is more (different) speech. In that spirit the right lawmaking response to school systems presenting systematic racism as unarguable truth and forbidding the questioning of it, would be to insist that the subject be discussed/presented as one idea about which there are different viewpoints.
"In a society genuinely devoted to freedom of expression, the right response to shitty speech is more (different) speech. In that spirit the right lawmaking response to school systems presenting systematic racism as unarguable truth and forbidding the questioning of it, would be to insist that the subject be discussed/presented as one idea about which there are different viewpoints."
Or we decide there's only so much time in the school day and would rather spend it on other things.
Also, school discussions are not like discussions on this blog: students are forced by law to be there. They can't just go to another browser tab if they find that the "speech" isn't worth listening to.
That sounds like saying that the government should be allowed to pay people less than minimum wage.
More importantly, it’s saying that government has an important limitation, which is that it is not allowed to adopt any rules that empower its subdivisions to do things free from interference of higher levels.
Tech companies are not the government. They have no ability to compel or ban speech except on their property. No matter how big they get, you will never go to jail for violating Facebook's ToS. A government, on the other hand, can throw you in jail or fine you. They can compel your child to attend a school that only teaches what they approve of. So governments need far more scrutiny when they use their power to compel or ban speech.
It's a book ban when the government declares that a specific book isn't allowed in a school library. Doesn't really matter whether this book previously was in the library or not.
If that's the case, then what's the word for the thing where the government declares that a specific book is not allowed in a private collection or on a privately-owned bookshore's shelves? Because that's something much worse than inventory management for school libraries being done by people you don't like, and I think we want to have a term that points specifically to the really bad thing.
Libraries exist, public and private. They have owners, and the generally accepted rule is that the owner gets to decide what books go in the library. If the functionary the owner hires to administer their library decides to stock the shelves with books the owner finds obscene or abhorrent, and the owner says "don't do that again or you're fired", I'd rather have a different term for that than "book banning".
The owner is the individual school. Not the state government. The state government is limiting the books that individual schools are allowed to acquire. There’s a reason that schools usually are run locally, rather than directly by the legislature.
The individual school is owned by the state government. Possibly by way of a local government, but those are wholly-owned subsidiaries of the state government. Public schools are not autonomous non-profit organizations.
If corporations actually were as powerful as governments: had their own schools where people were mandated to study and the corporation decided what they are allowed to learn and what not, their own SWAT teams that would put you in solitary confinements or even execute you if you break user agreements and could do all the other stuff that goverment can - then yeah it would make all the sense to treat such corporations to the same standards.
But as for now we are talking about people being banned on a social media app and comparing it to a totalitarian state - this is just ridiculous. Don't get me wrong, social media bans for speech are still not a good thing. But they seem to be the least bad of current options.
RE governments: It matters what level of government one is discussing. Federal and State governments have broad powers. Cities, not so much. E.g. https://www.scstatehouse.gov/code/t05c007.php contains "The municipal governing body may fix fines and penalties for the violation of municipal ordinances and regulations not exceeding five hundred dollars or imprisonment not exceeding thirty days, or both." ( I don't know how often cities actually use the power to imprison over ordinance violations. I've mostly heard of fines. ) Come to think of it, my ISP has a bunch of weird charges it imposes, which are not terribly different from minor municipal fines.
The amount of authority school librarians are supposed to have is a separate question. There can be multiple different arrangements under civil society. And some of them can be more problematic than others. But it's a completely different level of problematic than outright govermantal ban.
There are categories of books that we know are de facto banned from public (government) schools because of the ideas in them. I'm thinking of the sorts of books that might get mentioned favorably in the comments on Steve Sailer's site.
Are those books banned? I don't see why only explicit bans should count as bans when the results of implicit bans are the exact same.
I agree. I think it's certainly worthwhile to have conversations on what takes up a school library's limited shelf space (and public libraries too), but I'm sick of the term "banning" being tossed around in this context.
Wait, why is it incoherent to say that there should be a level of civic service that is immune to temporary fluctuations of elected officials? Most people believe that it’s a good thing, and President Garfield was even assassinated because he was part of the movement that supported this sort of good government movement limiting political control over broad-based government services.
Right. That seems to all count *against* removing books from libraries, because removing them is interfering with the choices of students and parents about what to read.
What is it a tell of? Usually the party making a claim needs to provide evidence to back it up. I was curious to know if trebuchet had actual evidence that the same people were doing both things or if it was just a general feeling they had. Well from their comment below it appears it was just a general feeling they have and they weren’t able to provide actual evidence.
Blog post on AGI - tl;dr I don't think the doom scenario is likely (mostly because, in my view, it rests on the orthogonality thesis, which is more of a series of assertions), but I think the alternative (a post-scarcity utopia) is almost equally scary.
For your problems with utopia, we would basically all become aristocrats which has its dangers (i.e. we could all turn into very frivolous and vain people). But people always figure out something to do: there is no reason to climb a mountain, yet lots do it. I read a lot in spite of believing there is no such thing as a system that provides final answers. People still play chess professionally in spite of AI in theory obsoleting human chess players.
I expect even in post-scarcity there will be plenty of conflict of the interpersonal sort, as the attention and affection of others will remain a scarce resource (and no, AI can't fulfill that for most people, there is always the knowledge you are interacting with nothing when you interact with an AI).
Your thoughts on the orthogonality thesis deserve a wider audience in the rationalist community. But your mode of expression - which is combative and not charitable - will I think make this less likely.
I would say it is worth recutting to be more persuasive to the rationalist community - they are a good audience! They would actively seek to engage with thoughtful criticism of orthogonality, politely expressed.
I don't want to come across as intransigent, but though I'd love to engage with the community (I'd love to see where I've gone wrong with my reasoning, if I have), I don't think I have the time to edit this - I have a few other posts I want to work on in the next few weeks.
Also, I'd suggest two things: first, I grant that I am unnecessarily combative, and parts of my posts are just my venting - but equally, rationalists' mode of expression _is_ almost purposefully incomprehensible, and turns people off. This needs to be addressed. Second, I hope that rationalists live up to their name and engage with arguments regardless of how they're expressed!
The fundamental issue is how we should engage with this ASI or ASI's. There is a scenario where humans use the super intelligent AI (ASI) to augment there own abilities. Instead of the AI doing the art for you, you create technologies which make you smarter or more creative and are able to create better art by yourself. This isn't limited to cognitive capabilities. You could intervene in human biology to modify temperamental dispositions. The general idea is to use ASI to improve human potential rather than recursively improving the ASI abilities. You bring up the issue of humans becoming addicted to virtual worlds or becoming bored or losing meaning. Well, we could use the ASI to develop tech which makes us more aversive to virtual experiences.
The second biggest problem here is that problems NEVER disappear. The development of AGI would likely raise new and profound questions about the nature of consciousness, intentionality, and mental states. Questions about the nature of time, causality, and the relationship between mind and matter would still be relevant.
The third problem of endless entertainment is a serious risk of causing a voluntary and unintended human extinction. The problem of these technologies being able to do art, entertainment, and other "humanities" work does not mean we should allow people to use them to do this. We could instead focus on fostering a culture that values personal development and real-world experiences, and setting boundaries for technology use. It's possible that this would lead to new cultures which voluntarily limit there engagement which the ASI.
A "deathless" world is not possible. The widespread use of advanced medical interventions would lead to new health problems, or technologies meant to extend life could be misused or weaponized, resulting in harm rather than benefit. Even with interventions from a superintelligence, it could be impossible to completely counteract the effects of entropy on biological systems, ultimately leading to deterioration and death.
We're living in a time where exploring new worlds and creating fresh cultures and ways of life mostly happens in the virtual or abstract realm. ASI comes in as a game-changer. We could use it to boost our creativity, speed up scientific breakthroughs, and offer innovative solutions to major global problems, ASI can kick off a whole new era of amazing possibilities and growth. It's like opening the doors to an even more exciting and adventurous human experience.
I think your imagination is a little constrained here! First off, any new questions that arise will be answered very quickly. That's what makes a deathless world possible.
I dunno. I think your imagination is too influenced by the idea of a being with great knowledge and power who also loves us. That's a story from Sunday school. lIt's not obvious to me that a superintelligence would love us, answer all our question quickly, make sure that new resources were distributed equably, etc. Superintelligence isn't going to be YOU, but with extra power, It's going to be more like an alien from a very advanced species. Stop sneaking helpfulness, fairness, love of our species etc. into our definition of superintelligent AI!
and it's not obvious there would be a single superintelligence. There would more likely be several, or many, competing with each other, with humanity cowering beneath, like sparrows hopping around below fighting eagles!
Also, for a superintelligence, doing away with humanity might not be a permanent irrevocable step, as we naturally assume. So this would mean less disincentive for it to make that decision.
If it had bottled DNA samples of ten thousand or a million humans, and all species of their intestinal bacteria, it could wipe us out any time it liked, secure in the knowledge it could recreate us, raising a new generation with synthetic robots indistinguishable from parents (as explored in several SF films).
So perhaps a superintelligence would rationally decide the World could do with a thousand year break from human depredations, to restore nature's abundance, before conjuring up Humanity Mark 2. Or maybe it would treat that as a trial run for an interstellar mission, where the same would be done on arrival, just as we have artificial Mars habitats in the desert on Earth as tests for the real thing!
It might even decide a "reset" like that would doing humanity a big favour in the long run, by eliminating the crushing weight of past human history and achievement, which would by then prevent anyone from ever making an original new thought, or theory, or song, or joke, etc, although I'm not sure quite how that would work if the new age is to be civilized, unless the superintelligence fabricated an elaborate pack of lies and distortions to account for the past.
So who knows, perhaps in a hundred thousand years there will be a human society where, unknown to them, the Evolutionists and the Creationists are both right! :-)
Yes, but some questions can't be answered even by a superintelligent AI. For example, the halting problem. There are an entire space of problems which are computationally irreducible. And if any of those problems are related to achieving mortality reduction then we end up in a world of death.
Interesting read. I share your skepticism about the "utopia" outcome. The thought of scientific progress being made without any human understanding of that progress is a deeply depressing idea to me.
Side issue, but: MRI is definitely not telepathy. It's a very powerful magnet at very close range that can infer information about:
1. The location of different types of tissue.
2. Local changes in blood oxygen levels.
The examples of "mind reading" with MRI are very much party tricks that we shouldn't expect to scale much beyond what we already have.
Haven't read the article, but agree. I think people need work-- not backbreaking work, but work. I'm remembering something a depressed retired lawyer said: "People used to pay a lot of money to know what I thought. Now nobody cares what I think." In any case, I find it very hard to believe that everybody's going to have access to this utopia of no work, incredibly effective medical care, wild and rich forms of entertainment, and immortality. I think it's going to be the same old shit: The rich will have access to that stuff. I expect us to end up in something like the worlds in the earlier Gibson novels, with some clans so rich they own, like, the whole fucking moon, and most other people living in a scrappy world of urban sprawl, without much nature left, with access to various bits of tech that sound impressive to us, but that do not free them from being poor, lonesome, anxious, randos wandering through the sprawl.
I'm listening these days to a lot of podcasts, some of which are interviews with people developing novel uses of AI. There is now up and running a kind of AI that duplicates personality -- which in practice I suppose means behavioral tendencies, attitudes, mannerisms, and of course appearance and speech patterns. So you can pick some famous person to become your AI friend -- for instance Oprah, or various characters movies and video game. You then can access your AI friend whenever you like, and chat with them when you're lonely. Oh good, that's going to do a lot to help chronic isolation and the general fraying of the fabric of society! But that's not the worst. If you're wealthy and have speech and video samples of a deceased loved one, the company will make you an AI version, and you can then talk with your dead son, dead wife, dead mother or dead father, who will look at you with some approximation of the familiar twinkle in their eye, express some approximation of their familiar points of view using some approximation of their familiar speech cadences, curses, jokes, opinions, etc. Both of my parents are dead, and I really miss my mother and would love to have an hour to talk some things over with her. But having AI Mumsy on tap doesn't fill the bill. Jesus fucking Christ, I think it's obscene! Saw an article one time about somebody whose Jack Russell Terrior had died, so he had it taxidermied in a perky pose and it Dead Perky Dog was part of the living room decor. FML.
The guy describing the choose-a-personality AI was young, smart as hell, amiable and reasonable. But he seemsed utterly oblivious to the downside of the service he was offering. the peo[le who are developing AI are going to have an outsize influence on world events. I'm not sure it's an exaggeration to say they're going to run the world. They're not awful people. But they all have what a friend of mine calls engineer's disease. There's a lot they don't grasp about the complexities of being human and about what affects quality of life. They are kind in giving in that they truly want to produce great tech to help the world, but their ability to empathize and to grasp points of view different from their own is very limited.
I'm a recent (December 2020) retiree myself, but, unlike your depressed lawyer, I find that change purely positive. I do *NOT* want an endless series of bug reports to dominate my life. I am much happier choosing how *I* want to spend my time, thank you very much.
Catsambas's "but scarcity, death, and overcoming adversity are things that have defined and shaped humanity forever: they are the things that give our lives meaning." is, as far as I can tell, completely wrong-headed (or, alternatively, "meaning" is vastly overrated). One doesn't need any of those things to enjoy a sunset, or any of the other pleasures or comforts of life.
Yeah, I get that nobody needs an endless series of bug reports, but many people need to have things they're striving for, or tasks where they get appreciation from others (I think that's what the lawyer missed). I, for instance, never take vacations with a large lie-around-and-relax component. They make me feel aimless and blank. I like to do outdoor activities trips -- sea kayaking, hiking, etc. -- or go to NYC with a big itinerary of theater, etc. set up in advance. Do you do much sunset-watching, or do you have projects you're working on?
I watch sunsets, web surf, and grow crystals. I wouldn't call any of these projects, certainly not in the sense of "if I do X, and Y, then Z will be available and I can use Z and P to get Q, which is the point of it all"
Entertainment generally does not need "scarcity, death, and overcoming adversity". If we actually got a friendly ASI which removed many of these constraints (scarcity never completely goes away - if nothing else, the observable universe has a finite amount of matter) some forms of art would become less relevant, but I consider that a small loss compared to the gains from removing many of the constraints on us.
Hypothetically, dead loved ones is just a starting point. The further utopia/dystopia is that infants start out with exposed to ai "people" who who are optimized for the infants, and later, friends as well as family.
Again, hypothetically, but good parents and friends learn and change. The better ai personalities should also do that.
Infants recognize their mother's skin by smell. AI is odorless. Though of course you can rub have some chemicals on the fucker and it will have B.O. too.
Nancy, how do you optimize AI people for infants? What is it are you trying to optimize? Are you trying to make them as calm, happy,neutral and smart as possible? How do we know that the parent's discontent with life as it is isn't useful the the infant, even if sensing that makes the infant less calm and happy? Think of the public figures you admire. Do you think most of them developed without absorbing some of their parent's frustration and discontent, along with their body odor? Do you think they would be better off with no deep connection with the smell of their parent's frustration and grief? And he parents -- do you think they're better off with no child to whom they tell the story pf the crap they've been through and what it's taught them? Are they better off being just egg and sperm donors?
I'm assuming a tech level which can also make good enough androids, including pheromones, so that automated childcare works.
Optimize for infants probably means calm and steady, with enthusiasm but not overloading the kid.
*Do* children need somewhat stressed parents? Maybe stressed enough but not too stressed? I assume that something reasonable is possible.
How about pretty good simulated parents for small children when the parents can't take care (too sick, perhaps?).
I'm assuming more that simulated people as an important part of life is something that people might try, and many might find that they prefer, rather than something that's absolutely better than the current world.
I think you're wrong to assume only the rich will have access to this kind of utopia - why would they? When there's literally near-infinite resources, it'll be impossible to hoard them - it'd only take one person to break ranks and share the wealth around.
"When there's literally near-infinite resources, it'll be impossible to hoard them - it'd only take one person to break ranks and share the wealth around."
Do we see Bezos walking around handing out thousand-dollar bills to lucky random strangers? The rich don't get rich by handing out money; they get rich and stay rich by penny-pinching, by making sure they know what is spent, where it's going, and what return they're getting for it. Right now, how many Indian peasant farmers are getting the wealth shared around by one person breaking ranks? What they're getting instead is "we'll back a WHO programme to sterilise you and your wife".
Why would a super-rich person hand over any of their wealth for no reason? Even the philanthropic get rewarded for it by way of tax breaks, fancy parties, their names on foundations, etc. Super-Bezos or Super-Zuckerberg can be rich, healthy, nearly-immortal and have all they desire, they don't need ordinary Joe from Oklahoma. If they want servants or company, they can get the AI to create better-than-human robots for that.
I really don't think you get what a post-scarcity world looks like. If Bezos gives $1k away (which he does, btw), that's $1k less for him. Post-scarcity, there's no longer a limited supply of resources to be divided.
Well. USA wealth is might be considered near infinite, depending on how strict your definition of near-infinite is. There sure as hell is a lot of it. If the money was distributed equally, every citizen would receive $3 or $4 hundred thousand dollars. But ain't nobody breaking ranks.
You might also consider the present situation, in which the tech companies have a tremendous amount of wealth and power. Congress does little to regulate them, and they are developing a technology that's going to change everything. In many ways, they are in charge of what happens in the world over the next century. What reason is there not to believe they are not going to hold on to a great deal of their wealth, power,secrets and access to extraordinary technology?
If you imagine that super-intelligent AI is basically God (except made out of shitty intricate little wires, etc., put together by low-empathy engineer type people) then I guess GodAi would come down and say "Let there be fairness!" But if we only assume AI is going to be god*like* in its knowledge and power then I think we all have to throw away our Sunday school versions of God and just think about a being of infinite knowledge and power, and drop all the frills from the old story about how it loves us blah blah blah. Honestly, why should it? Or maybe it will, but the way some people love their pets "Aw, look at his widdle face. He's mad because I didn't give him another treat, so cute, haha." Or maybe we will look like doodle bugs to it. Or sushi. Or turds.
I made an RPG setting inspired by Archipelago, all done except for pictures. It uses Traveller rules and has a 3D star map. Any TTRPG fans interested in playtesting?
I realized something about ambidexterity -- Scott used it as a baseline to compare with bisexuality etc. in recent posts.
Everyone uses their "off-hand" for some things. Forceful actions in one-hand, fine-dexterity in the other. Or some split.
Maybe there's semantic differences in how people understand and self-identify as "Ambidextrous". I myself write with my left hand but do plenty of things right-hand-dominant... do some people call that ambidextrous? And are certain types of people more likely to classify themselves as such?
Makes me wonder about musical instrument design. Guitar: Why pluck strings with right hand and fret the strings with the left? There are left handed guitars also though, so probably meaningful difference for hand dominance. Piano: Why are the bass notes on the left of the keyboard and the higher notes on the right? Are there any reversed pianos for lefties?
> Everyone uses their "off-hand" for some things. Forceful actions in one-hand, fine-dexterity in the other. Or some split.
This hasn't been my experience. My left hand is both stronger and more dexterous than my right. When I use my right hand it's usually because the task requires both hands, or because I'm using something that was designed for the right hand.
> I myself write with my left hand but do plenty of things right-hand-dominant... do some people call that ambidextrous?
"ambidextrous" usually means that you can do those things with either hand at least competently -- so unless you can also write OK with you right hand and switch-hit at baseball, batting right and writing left would not typically be considered ambidextrous.
FWIW, Tom Scott had a nice recent short video on what a challenge trial is like from (approximately) a participant's viewpoint: https://www.youtube.com/watch?v=1v8u3ua6BPk "The people who get paid to get sick"
Wanted to see if I can assess people’s feelings about free will in a way that doesn’t degenerate into “well what do you mean by self, choice, natural, and the universe?”
Say you have a computer powerful enough to simulate the entire universe you’re in and you can look into it and see yourself a few seconds in advance. I suspect this isn’t possible but we’ll take it as a given.
Two questions:
Must anything be true of your feelings about the actions the computer shows you performing in order for them to come true?
Would your actions in this universe with the computer be any different than your actions in the universe without the computer?
Your hypothetical implies that the computer can solve the Halting Problem, which it can't. The computer can't possibly simulate the entire universe, including the effect that the computer has on the universe.
a) I'm not entirely sure about your meaning here. For me to do the action the computer shows me doing, it would probably have to be something I'd be fine with doing already. There's only a few seconds involved, so there's not much time for serious thought. I would hate to get in the habit of doing whatever it predicted, simply because I saw it. That's distinctly different from doing the thing it predicted without having seen it, because there's an extra bit of causality in the middle.
b) Is the simulation a simulation of a universe with a computer simulating a universe with a computer, etc., ad infinitum? Sorry, this is probably not an answer you wanted. :-) But practically, for me, I would absolutely test this out, almost the first thing. Do I raise my right hand or my left? Whatever the simulation shows, I'd do the other way. If that crashes the universe, so be it.
For A I’m right there with you. I think knowing you were going to see the output changes the output it can give you. And for B, I think the futures where you see the output is different than if you hadn’t because you took action based upon it.
Agreed that irl this probably causes an infinite recursive loop that crashes the whole thing. Also probably a lot of these involve people just choosing to look away or turn off the machine.
It would be amusing if I looked at the computer, saw myself raising a right hand, raised my left, and then looked closer to see that the simulation inside the simulation had me raising my left hand, and the simulation inside that one had me raising my right hand, and inside that my left, and inside that my right.
Along similar lines, I came up with this thought experiment: suppose you and a functionally omniscient AI each have a deck of cards. You play a simple game where you and the AI both secretly select a card and then play them at the same time, 52 times until you exhaust both decks. The AI’s goal is to always play the same card as you. At the start of the game the AI pre-arranges their deck having already determined the exact order you’re going to play your cards in, so that each round the AI is simply flipping the next card off the stack. Without fail, the AI matches you card for card every time.
Now imagine in the middle of one of these games, the AI picks up the next card on the stack and shows you what it’s going to be before the next round. Would you still pick that card anyway? Personally, I wouldn’t. There does not exist any order the AI can arrange the cards that can be correct. Which means that even if events in the universe are deterministic, they can’t be determined.
So to answer your first question, in order for the computer’s predictions to come true it has to predict things that you want to do despite knowing you are predicted to do them. If the desire to be unpredictable ever outweighs the desire to take the predicted action, the computer will fail. So whether or not the computer will work comes down to whether or not you want free will.
As to the card flip, I imagine this is something like the theological position that God is in. I use God here in whatever the least offensive manner is you find but basically all knowing all powerful source of creation stuff.
From the inside of our lives we have free will. We see the future darkly and react to how we think it will be. From the outside it’s all atoms moving by laws. So does God show you that card knowing that you’ve accepted He is all knowing and your choice to do what He has revealed is your sign of submission? Or does God choose the card you won’t choose knowing that it will make you feel from the inside that it was your choice even if it was all destined?
I have had some thoughts about what you want to see influencing the mirror based on your fundamental beliefs about what you should see but your example is much better for this piece.
Would you be okay if I attributed this to you in a write up?
Not heard of it before but based on wiki, kind of?
I think the hope I had was that we could have a discussion about what you’d actually see on the computer screen and how that would have to relate to your feelings about it.
Also like your username. Triggered my secret name seeking instincts.
Well, if I personally were looking at it I would likely see myself doing nothing, and it would be accurate.
Once you get a mobile version of this computer, you see yourself running into a deer or a pedestrian, and so you slow down or swerve to avoid the outcome, because you feel strongly that it's to be avoided and know action has to be taken to avoid it. Weaker versions include going to a shop that can't sell something right now; you just won't go to that shop.
That doesn't really address determinism or free will, though. My take on free will is it's demanded by any universe that isn't a closed box. A prediction computer doesn't sway that one way or the other; predicting one box means it's another thing in the larger box, if there is a larger box.
I think in the case of you swerving to avoid the deer it would have to show you swerving and avoiding the deer and make you understand in the image through pure coincidence why you are doing that even if you start the swerve before you see the deer.
NasalJack had a great experiment on it above. I think the all powerful AI dealer showing you the card first knows you well enough to know you will flip over the same card even if you know it wants you to do so.
Determinism needs to be distinguished from predictability. A universe that unfolds deterministically is a universe that can be predicted by an omniscient being which can both capture a snapshot of all the causally relevant events, and have a perfect knowledge of the laws of physics.
The existence of such a predictor, known as a Laplace's demon is not a prerequisite for the actual existence of determinism, it is just a way of explaining the concept. It is not contradictory to assert that the universe is deterministic but unpredictable. But there is a relationship between determinism and predictability: predictability is the main evidence for determinism.
Nonetheless, determinism itself is the crux, in arguments for free will, and predictability only features indirectly as evidence for it. Determinism is the crux, because it removes the ability to have done otherwise, which seems to be important for moral responsibility; and also removes the ability to shape the future with present choices.
I'm not sure this scenario is actually intelligible, even if you swap out the computer for a genie who can magically show you yourself -- and your thoughts? -- a few seconds into the future. Nor do I understand what sort of answer you might be looking for with your first question. For the 2nd question though, I guess there are three possibilities:
1. You freeze and gaze at your future (frozen) self forever like Narcissus. I don't think this tells you much about free will except that even if you did have it prior to meeting the genie then you don't have it any more.
2. You do the exact same actions -- and have the exact same thoughts -- as the genie is showing you. If this is what happens -- like an infinite mirror except that it's driven by some layer further down than the original 'you' at the top (now reduced to a helpless, powerless marionette) -- then I guess it would be a strong piece of evidence against free will. But I don't think this is even intelligible, never mind possible, because you'd be standing on top of an infinite turtles-all-the-way-down recursion, while somehow not even seeing that or being aware of it. You'd have to see your future self in the infinite mirror and also *not* see you future self in the mirror but only your future self sans mirror. I guess you might be able to manage a flat out contradiction like that while eg in a dream state, since dreams can accommodate contradictions without breaking stride, but now we're talking about free will in dreams -- and we seem to have moved even further away from the original problem.
3. You do something different from what the genie is showing you. Well then, boom, free will I guess. Unless of course it's just a crap genie and the real genie who *could* have made this work is having the day off.
The problem is that I don't see any way of deciding between these three options. They all seem just about equally impossible, and if #1 and #3 are at least coherently imaginable, that doesn't make them any less impossible. (Like I can coherently imagine what it might be like to go back in time; but ti's still impossible for me to actually do so.)
So I'm not sure what this thought experiment was supposed to achieve? Did you have any particular answers to your questions in mind?
(Fwiw I think that 'free will' is just a nice story we tell ourselves. All that's really going on is that we do something and then we tell ourselves that we decided to do it.)
So yes probably the computer isn’t possible that’s why we have to take it as a given.
What you do in the future depends on your understanding of the future. I’ve never met someone into determinism who doesn’t believe there’s a difference between voluntary and involuntary action even if you nest that into a deterministic framework.
Because that’s a given element of the universe already, when you look at the computer screen to see what happens in the future it has to show you something you would always choose to do anyway even knowing that it’s coming in advance so that it can remain deterministic. Yes, in real life, nested simulations all the way down and a computer can’t support it.
This future is different than the futures without the computer because your knowledge of the future caused you to behave differently.
So however you define free will or determinism, that’s my idea of what happens.
There is clearly free will, if you presupposed MWI. It simply means choosing which Everett branches of the multiverse you wish to inhabit. Philosophies that disavow the meaningfulness of free will often presuppose a classical universe that is fully deterministic. Under MWI this is incorrect.
That's not how MWI works -- you don't have a non -physical soul that switches tracks between universes, you have a physical brain versions of which exist in every universe.
This is what I mean by definition games. I think you have free will in even a single universe that is fully deterministic because I don’t think those two things are at odds.
No, you got it wrong. You don't get to choose which branch you inhabit, or wish to. You inhabit all of them, unless you die in some. If you actually read what MWI proponents say about the issue (Sean Carroll is highly recommended), you will see that MWI has no bearing on the issue of free will whatsoever.
I should be more precise. You will no longer identify with those versions of you that choose otherwise, just as you don’t identify with those alternate versions of you who won the lottery or suffered a crippling car crash. While a version of you inhabits all branches, you only really care about the ones you identify with. Hence each choice is functionally the same as choosing which branch “you” wish to inhabit. I’ll look into SC though.
I guess my point is that you don't choose to inhabit a specific branch, causality goes the other way, sort of. The universe evolves deterministically according to the Schrodinger equation, including what can be emergently described as many-worlds branching, and it feels to the emergent you in each emergent branch like you made a choice to do something. For a hypothetical Laplace's demon (who knows the microstate of the universe and can watch it evolve) these emergent macroscopic concepts would be superfluous.
It's unclear from your description whether the simulated version of you also sees a prediction.
If your simulation doesn't see a prediction, then there's an important difference between you and the simulation of you, so it shouldn't be surprising if they behave differently. Even a fully-deterministic computer program can produce different results if you feed it different inputs.
If your simulation does see a prediction, then that prediction can't be generated simply by simulating physics, because now the computer is trying to predict its own output. Nothing can run a fully-detailed simulation of something more complicated than itself. So where did it come from?
You could "guess and check" by picking some prediction, simulating a version of you that sees that prediction, and then watching to see if the simulation fulfills the prediction it just saw. This is called a "fixed point" of a function; that is, an input that causes the output to be the same as the input. But there is no general rule saying that a function has to HAVE a fixed point; something as simple as f(x) = x+1 has none. So even if your computer can accurately predict what you'd do in response to seeing any given prediction, there might not be any where "what you'd do in response" is the same as the prediction you saw.
(It's also possible for a function to have more than one fixed point; i.e. there could be multiple prophecies that would all be self-fulfilling if you saw them. Neither existence nor uniqueness is guaranteed.)
All of which is to say that your thought-experiment isn't well-defined and may contain a logical contradiction, which would make any results you derive from it useless.
Physically impossible is different from logically impossible. Logically impossible thought experiments are useless because, from a contradiction, you can prove anything. You therefore can't trust that the results are true or even internally consistent.
You also...don't seem to have answered any of my implied questions? You didn't say whether the simulated version of you is supposed to see its own prediction, and if so, what prediction it ought to see. I only said your experiment MAY contain a logical contradiction, but if you want me to reason about it further, you at least need to resolve the ambiguities I point out.
Supposing that the computer really has magically predicted the future, like a spice-vision from Dune or something, I think what it would mean is that when the time comes to take those actions, I would want to do them in the same sense as I do any other action I take.
My actions in the universe with the computer or spice-vision would be different because in the universe with the computer or spice-vision I looked at the computer or had the spice-vision, and that is an action.
If your understanding of the future is a cause of your present action then no because the computer must already take this into account in what it shows to you. It can’t show you something you don’t want to do.
In a truly deterministic universe, no. Physical determinism is not really compatible with "voluntarist" free will, where the actions you take cannot be explained by external factors (the corresponding physical process, in this case).
For free will to exist in a fully explicable universe, it would have to be on a more "intellectualist" model.
That’s the layer that the contradiction is at. If the mirror ever gave you cause to behave differently than it couldn’t be deterministic so because we know there’s a difference between voluntary and involuntary action it has to show you things that you would choose to do.
LarryBirdsMoustache, much like the man to whose face you are attached, you got it in one shot..
The universe can be deterministic, but since what you do depends on your mental model of what the future looks like, it would create a recursive loop where it can only show you future states that you want to enter.
It's been a while since I cracked them open, but I'll differ on this.
In a sense, that's certainly the subjective experience of the prescient - that they only have free will in areas that they haven't predicted. ("Absolute prediction is absolute death.") But they also do have choices - they can see many paths in the future and take actions that bring about their chosen future. And they have difficulty predicting random behavior, like sandstorms, or split-second decisions in a knife fight, or decisions based on tarot readings.
Part of the problem is that uncertainty and ignorance prevent ethics from dominating all choices, and once they **know** what the outcomes are, they're ethically obligated to take the actions that lead to the best outcomes. (See also, some of the psychological problems that people who get deeply into Effective Altruism have to deal with.) And there's a Great Filter in their future, and the actions that bypass it are horrifying, which leads Paul to attempt suicide-by-desert as a way to pass the buck to the next generation. It's a bit like if, in "Groundhog Day", Bill Murray's character felt obligated to constantly, repetitively, take the actions that maximized welfare for the people around him. No piano lessons, he could be saving lives instead. (The same life, over and over, every day.)
is "free will" in any meaningful sense still an open question?
thought folks nowadays are AI-optimists and predicting human or above human level silicon-run entities within decades.. do folks with those beliefs still believe "free will"?
As said in the comment, it really depends on what you mean by free.
Do I think we can just change the world around us with a thought and break reality itself into a thousand pieces and that if we can’t then we live in a deterministic hellscape? No.
For me the question is solved to my satisfaction. Humans have free will in the sense of desicion making ability and do not have free will in the sense of transcending causality, regardless of determinism/indeterminism of our universe. And yes I do expect human level silicon-run entities within decades and I do expect them to have similar to humans free will. Though, calling such position "AI-optimism" gives me a sad chuckle.
The universe is only deterministic if you assume there is a single universe which is classical in nature. But our best theories indicate reality is quantum in nature. Free will is just picking your branch!
Yes. Actually, I mostly understand the concept from the lense of determinism. The less purely random stuff is going on in the universe - the better is our ability to predict possible outcomes and thus execute our decision making ability. So I'd say we are have more free will in deterministic universe.
It is logically impossible for a computer to simulate the whole universe. If it did, it would have to simulate itself, but then it would have to simulate itself simulating itself, however then it would have to simulate itself simulating itself... This lead to an infinite regress. How can any computational process within the universe simulate the very universe that gives rise to it? This seems logically contradictory and physically impossible. The simulator cannot simulate itself and create itself.
I would definitely say that your actions contingent on seeing a prediction of what you'll do will often be very different from your actions in another world where everything is the same except you don't see the prediction, if that's what you're asking.
I’ve noticed a lot of rationalists expressing anxiety or fear at the thought of AGI ruin. This doesn’t make sense to me, as someone who accepts MWI and who would have thought most rationalists do too. Am posting this in case it helps anyone suffering such anxiety or fear, but keen also for feedback.
TL;DR
Given (i) MWI, and (ii) that AGI ruin occurs with less than P = 1.0, then by quantum suicide/immortality logic, none of us should fear personally experiencing AGI ruin.
The key reasoning is as follows:
MWI remains our civilisation's best theory of the nature of reality.
Under MWI, since all phenomena are quantum mechanical in nature, it follows that every phenomenon that can occur under the laws of physics does occur in some part of the multiverse, ie in some universe.
Any phenomenon that involves obliteration of your consciousness, while it does occur in some measure of universes, is not something that can be subjectively experienced by you. You should therefore never expect to subjectively experience death.
AGI ruin entails your - and everyone's - death (AGI ruin scenarios that follow from premises such as the orthogonality thesis and instrumental convergence involve inadvertent disassembly of all humans).
AGI ruin occurs with less than P = 1.0. Or rather does not occur in all universes.
In universes where AGI ruin occurs, no humans will be there to experience it. AGI ruin scenarios do not readily admit of a situation where humans should continue to survive in a state of severe disadvantage or discomfort.
You should therefore not expect to subjectively experience AGI ruin. This alone should be sufficient to alleviate any personal fear or anxiety one might feel at the prospect of AGI ruin (whatever your views are on the measure of universes experiencing AGI ruin).
This reasoning is not new, and is I think a trivial extension of quantum suicide/quantum immortality thought experiments - so long as you accept that QM describes reality.
So what do you think. Should this alleviate a rationalist’s concerns? Is this an idea worth spreading?
Incidentally, should there be, for example, a moratorium against development of advanced AI, many of us should expect to personally experience living in a world deprived of the benefits of advanced AI - and to suffer from experiencing any ill that might otherwise have been averted by advanced AI. For those who place more weight on anticipated subjective experience, rather than, for example, negative outcomes in a measure of universes that will never be subjectively experienced by anyone, a moratorium would be dispreferred - which may justify a continued tech accelerationist outlook.
You seem to have rediscovered the quantum immortality argument. Go read the original paper on it, by the philosopher David Lewis, and reflect on what it really implies about your future experiences.
I don't understand some of this argument, so maybe you could point to places where it's discussed?
For one thing, it seems to privilege consciousness as the fundamental of the universe. I don't see why that has to be the case; it seems like wishful thinking at best. Is this because of some technical argument about observers?
For another, we might not technically experience "death", but surely we can experience "dying"? If I experience dying horribly but then wake up in another me in another branch, then I'd believe this theory, but I haven't experienced that, and neither has anyone else that I'm aware of. If the argument is that we won't experience the dying, either, then wouldn't that percolate backward along our timeline, and the only timelines we experience are ones where we're immortal? (Leading to the conclusion that anyone who dies didn't have consciousness beforehand, leading to the conclusion that anyone you manage to successfully kill didn't have consciousness beforehand.)
Although perhaps that's memories, not consciousness? So maybe the question is what happens if I wake up one morning as a completely new person with a completely new set of memories - how can I tell? What is the "me" there?
I don't understand why people are so convinced that MWI is true. Yes, it is a natural conclusion from our existing laws of quantum mechanics; but our laws of quantum mechanics do not fully explain our universe. We know for a fact that our models are incomplete, and there is likely to be underlying physics of which our existing models are an emergent property. Why assume that the universe works based on an interpretation of an incorrect model?
To add to that, even if MWI is true then dying in one branch decreases your probability mass. If you care about all versions of yourself, it's better to avoid that.
Sure the theory of everything hasn't been discovered yet and we don't know what it is, but we do know that in lots of cases it can be approximated by quantum mechanics, just like quantum mechanics can be approximated by newtonian mechanics. Given that quantum mechanics is a good approximation in practucally every case we can observe, it seems likely that, for any given phenomenon we can't observe (such as separate world branches), quantum mechanics will probably also be a good approximation in that case too. While we know that the standard model isn't always correct, the situations where we know it isn't correct are ones involving gravity, extremely high energies, something to do with neutrinos I don't really understant, and probably other cases I haven't heard of, and I see no reason to believe that decoherence is one of them other than an intuition derived from our experience of a basically newtonian human-scale world, which is even less of a complete description of reality than the standard model is. To assume that MWI is *probably* false, rather than merely *possibly* feels to me like using our ignorance as an excuse to believe what you already want to believe for other reasons.
"Given (i) MWI, and (ii) that AGI ruin occurs with less than P = 1.0, then by quantum suicide/immortality logic, none of us should fear personally experiencing AGI ruin."
I don't care if there are 100 worlds out there where the guy pointing a gun at me right now doesn't shoot me in the face, I very much care if in this one world I get shot in the face. And if I do get shot in the face, I'm dead and whatever happens in the 100 other worlds is meaningless.
Go find a parent whose young child has died and cheerily tell them "Hey, under quantum immortality, your kid is alive somewhere else so stop being sad!" After you get punched in the face, console yourself that there are worlds where you didn't get punched in the face, so your pain right now isn't happening.
"You should therefore never expect to subjectively experience death."
That has to be the single stupidest sentence I have read all week. You can objectively experience death but not subjectively? You'll die but not notice you died? When dying, you will not have the feelings of dying? Unless you expect to drop dead immediately, when I suppose you might not subjectively experience it since you won't have long enough to experience anything, this is ridiculous.
The only sense I can make of it is "upon death, what we might call consciousness is immediately extinguished, so you cannot experience 'being dead', you can only experience 'dying'". Great, but I'm still dead either way.
>"Hey, under quantum immortality, your kid is alive somewhere else so stop being sad!"
It's not like i agree with Snakegown's theory but many religions do say very similar things when people die, so maybe it works. And QI is just a newer religion for physics nerds.
If he wants to re-invent reincarnation, he can do so but there are already religions with this as part of their structure. So it does seem like wasted effort.
1)Wut? How about if you die after lying with a great block of collapsed cement building on your legs? You're in agony from you crushed legs, you're septic, you're dehydrated and alone. Most people's deaths aren't a sudden "lights out," which, I agree, need not be feared. Most deaths are preceded by a period of every-increasing disability and pain. Many have to endure that period with nobody who cares about them near at hand to comfort them. If that isn't worth dreading I don't know what is.
2) I'm reasonably OK with the prospect of dying myself. But I have a 20-something daughter who is on the verge of getting married, and full of joy at the thought of her coming years with her husband, building a house together, having babies, raising their children, going on camping trips with friends. The thought of the astonishment and grief on her face if the world comes apart and she realizes she can't have any of those things is absolutely unbearable.
This is not meant as a general purpose argument not to fear death or disability, but to assuage fear of AGI ruin. Posited AGI ruin scenarios are so total that you, your daughter and your future children have nothing to fear from them personally, since none of you will ever experience them. Positing an AGI ruin that you can experience, whether involving blocks of cement or otherwise, requires developing a specific, unlikely scenario to be worried about.
Even the scenarios that take longer to result in death will end (for you) when the time of death has passed. The you who suffers that death will identify with current you, but so will the you who exists in the non-AGI ruin universes. But the latter will not identify with the former. And the former will no longer matter. So “you” will survive AGI ruin (or rather AGI ruin will not occur for you), even if some other versions of you who you won’t care about will not survive AGI ruin.
Quantum immortality is a nightmare, not a get-out-of-jail-free. “Cannot experience your own death” doesn't mean what you _do_ experience will be pleasant.
Exactly, I don't think people understand how horrifying quantum immortality is. It doesn't just mean you will reactively experience bad things of unimaginable disutility. It also means you will pro-actively experience bad things of unimaginable disutility because of value drift or some other process which operates on the long-term. Nobody would take a pill which would make them do certain things, but quantum immortality means eventually because of probability (there is a non-zero chance) you will do those things.
You have free will though. Every choice you make confines “you” to futures that include that past choice having been made. Even though other versions exist who made a different choice. But you dont identify with those ones any longer, they are not “you”. Thus I wouldn’t worry about to the future versions of you that make a choice you would disagree with.
Free will makes no sense. if we have free will, then we should be able to choose to act against our own free will. But if we can do this, then we don't really have free will.
If our free actions are random, uncontrolled events, how can they be truly "free" and stem from our own willing and agency?
If every action requires a preceding cause that also requires a cause, how do we ultimately choose anything freely?
If our choices can be manipulated by directly manipulating our brains, this seems to challenge the hypothesis that there is some kind of free will beyond mere physical causality.
if our thoughts are strongly determined by evolution, how can we have free will?
There seem to be fewer things for a "free will" to do - neurons and brain events seem sufficient. But if there's no role for an agentive self to play, how can our will be free?
Even if we have free will now, in order to be ultimately responsible for our actions, don't we need to be responsible for becoming the kinds of beings we are - for developing our own character, values, and motives?
hi, would love to talk philosophy with you, if you can just drop the pseudo-science language (I think we can debate the core claims without that but lmk if disagreed)
Think the key claim you need physics for is
>You should therefore never expect to subjectively experience death (1)
which we can take for granted (and which I agree with).
Also seems most of the argument is really not much about agi but just death and what comes after so lemme ignore agi as well for the moment.
Seems to me you might be implying (1) implies that
(2) humans can't have preferences over duration of time till their death
(3) humans can't have preferences over what comes after their death
(maybe (humans can't have preferences) should be replaced with smth else but not sure what)
Think people would mostly disagree with (2) (metoo), but also bigger issue/concern is (3).
I'm curious to hear what people feel about (3), humanity-consensus think is disagreement, philosophy-consensus and arguments for/against I'd be curious to learn about.
If you think there is only one universe, you may place great weight on whatever happens after your death, even if you won’t experience it. As many have through history. If you know you will survive in some branch, seems better to focus on optimising those branches.
"If you know you will survive in some branch, seems better to focus on optimising those branches"
How? How do I do that? In this branch, I will die some day. We know this for a true fact of reality. If I live to be eighty, I'll die eventually. How do I 'optimise the branch' where I live to be eighty-five? This is dream logic for dream worlds.
There is no way I can 'live' by leaping from branch to branch, never dying, because I live to be ninety in this one and then another version of me in another branch lives to be one hundred and so on for infinity. I die for good and all sometime, in some branch.
Yeah, it’s like Communism that way. It seems Jack Spratt and his wife cannot get along. Never mind though because humanity’s hive -mind-child AGI will change that. It has the best chance.
I don't think quantum immortality is generally taken as seriously as you imply, even by those who accept MWI.
By the same logic, no-one should ever expect to experience dying in a car crash. But MWI should have no impact on whether you experience your brakes failing or not, or whether you experience falling asleep at the wheel and swerving into a truck, or even whether you collide with that truck and begin flying through the windscreen. By the time this principle could be applied you are already dead in all possible universes with near certainty.
This doesn't make sense. If there is even one universe where "death is avoided by epsilon - Jeffrey Soreff" then you will exist in that universe. If there are an infinite number of alternate possible universes then there will always be one in which you end up in. But this applies for first-person perspectives. From a third-persons perspective, people die all the time.
Yeah, but from a first-persons perspective you cant die. So it means you end up in a scenario where everyone dies and you survive. These two viewpoints make no sense to me and I have no idea on how reconcile them.
There's no quantum immortality force stopping you from experiencing a world where AI has unleashed a deadly virus that has infected everyone on the planet, all of whom will die within days. The point of divergence between the worlds where you exist and the ones you don't can come after the point where you're already guaranteed to die in every universe.
To get around that, the claim you have to subscribe to is that something of ridiculous unlikelihood would save you - maybe all of the virus in your body simultaneously falls apart and you survive. And that these unlikely events would keep saving you.
Some say that quantum mechanics makes this inevitable. I don't see why it needs to personally, but I sure hope it's not. Because that guarantees that all of our subjective experiences end up in eternal torment as everything else breaks down around us. In this example, you're still the only survivor in a world ruled by AI. Not comforting.
I believe poincaré recurrence theorem says that anything that can happen, will happen.
Presumably this implies you can luck your way to immortality. Say you are 120 you are likely to die soon of something or other. But if you die of a heart attack on 2 June there's a world in which you survive that heart attack, and the next one, and the next one...
There is. The possibility is that the virus just happens to not infect you. Or the virus just happens to miss all of you healthy cells. Or you immune system just happens to create antibodies from a previous asymptomatic infection which is enough to trigger an immune response to the deadly virus.
One of the nice things about AGI ruin scenarios is that they are often quick and unanticipated deaths, so those of “you” destined to die would experience it only briefly, if at all. But even if there are bad scenarios, you will still survive in all the other worlds on the other side of the point of divergence.
"If there is even one universe where "death is avoided by epsilon - Jeffrey Soreff" then you will exist in that universe." Yes, and that existence is very likely to be highly unpleasant. Being _almost_ killed is generally a torturous fate. Which is why I don't think quantum immortality is particularly comforting.
Yup. You are looking at a bunch of unpleasant precursors to death, and I commented about unpleasant situations where death is avoided by epsilon, but in either case quantum immortality is (nearly) no help.
Yes by far the current "best theory" is "its not clear WTF is going on", not MWI, which has maybe IDK 5-10% support among the people who really deeply understand high end physics. and its metaphysics.
"Any phenomenon that involves obliteration of your consciousness, while it does occur in some measure of universes, is not something that can be subjectively experienced by you. You should therefore never expect to subjectively experience death."
To my mind, the problem with the quantum immortality argument (regardless of AGI) is that there are possibilities which are epsilon away from death and are very unpleasant.
The AGI ruin scenario is special because it posits death of all humans. Even if you don’t take comfort from QI re normal situations, it seems to work here at least. Esp if you personally do not expect to influence the likelihood of AGI ruin.
As it happens, I indeed don't expect to influence the likelihood of AGI ruin.
I see your point about the special nature of AGI ruin, but I think it is less binary than you posit. There are slivers of probability amplitude where e.g. portions of nanomachines that an ASI intends to use to kill us just barely fail, and leave the observer in question severely damaged but just barely alive. Quantum immortality will push some of the amplitude of the observer into these scenarios too.
I guess I'm skeptical of quantum immortality, even given MWI, because it rests on there being something really really special about the portion of the wavefunction where the observer survives. But that seems to rest on the observer's "survival" being a very crisp binary function of the microstate of the system. That seems implausible to me. Just exactly how many firing neurons does someone have to retain to count as "alive"? It seems more plausible to me that the observer's status is a smooth function of the microstate, and the importance of their being there as an observer fades out smoothly as the probability of their being a fully functioning human drops towards zero. Which brings us back to something much closer to the ordinary view where quantum immortality can be ignored.
Agree not binary. But if someone is anxious (not you it seems), then this idea should alleviate it - since they would have to go out of their way to devise a scenario (a sliver as you say) to be worried about.
Aside from currency, the one thing I thought blockchain would be useful for is for things like platform-agnostic libraries of content. For example, you might own a key for a game on the blockchain, and you could then play that game via any participating digital distribution platform (ie Steam clone) that had that game in its library. This would protect your game library in case Valve or whoever shuts down, it would allow you to purchase one copy of a game that would be playable on PC and mobile, it would allow better competition between storefronts, and you could then sell the game key on when you're done.
I get why existing platforms would not want to implement this unless under heavy pressure, but why have I not heard of any new platforms offering this service? Is blockchain not as useful for this as I imagine? Is it a bad idea for other reasons? Do consumers not actually care?
Blockchain tech has been very underwhelming in general considering the hype ~10 years ago. But at thw very least these kinds of use cases seem like they should have been exploited by now.
I mean, this is a version which ends with strictly worse sales for all companies which provide games to that platform (as they get $0 for every resale), so I would anticipate that any service which did this would basically have to be GOG clone, or actually own the games in question. They're definitely not getting triple A games.
It's arguably better for the consumer, but worse for both the storefront and the developer and the benefits for the consumer are honestly pretty low, so, no I wouldn't expect something like this to be of any use.
It would also, almost inevitably push game development in a direction I don't particularly like (away from single-player, story-focused, one-time games and towards ongoing DLC heavy, multiplayer and 'endlessly replayable' games, as you'd want John and his buddy to both buy it, not John to buy it and then 'sell' it to his buddy), so I'm glad there's not much chance of it happening.
I agree that the resales are the messiest part. The main model for digital game sales right now is to do progressively bigger discounts over time until you've theoretically captured everyone who would ever buy your game at the highest price they'd be willing to buy it. This captures the group that would traditionally have waited to buy second hand.
If resales were a thing this model would need to change. Games would probably have a higher upfront price. I would also expect stores to facilitate resales and take a cut for themselves and the developer. Users could exchange games outside of stores, but the extra friction involved in finding someone willing to buy, the knowledge that you're ripping the devs off, and the minimal savings involved would make this rare I think. If stores really wanted to discourage it they could refuse to honour resales that haven't gone through them or a recognised competitor, but it would be against the spirit of it all.
It'd be weird but not insurmountable I think.
I disagree that the benefits to the consumer are low. Or at least, I'd expect the benefits to be very compelling to like 5% of consumers, while most don't care and stay on traditional platforms. The target consumers would be the same ones that care about DRM-free, who are still concerned about Steam having a monopoly, and a bunch of crypto people.
It would be a little worse for existing storefronts, but I'd expect this to be implemented by new storefronts offering it as a point of difference. It's infinitely better for them, because they wouldn't exist otherwise.
It's pretty neutral for developers. Indies are better off being on as many storefronts as possible and benefit from sales on smaller ones because they take a fraction of the cut that steam does. The resales thing is the only problem that some might have, but as I said, I think that's surmountable. Most revenue comes from launch anyway, and your tail would come mostly from cuts from resales instead of dirt cheap sales.
I mostly agree about the direction it would push games. There are a few forces pushing them in that direction anyway though. I'm pretty ambivalent about the idea myself, just surprised it doesn't really exist.
I mean, all the emphasis I've seen is going the exact opposite direction of making it seamless to transfer between platforms, or resell games, or anything like that. Especially with things like Netflix cracking down on account sharing, the concern is that they've made it too easy to share, not easy enough.
On the specifics of resale, I think you'll have a very hard time with 'residuals' for resales. At the moment, people have basically decided to accept the fact that what you buy from Steam/etc. is a nontransferable personal license. They don't generally love it, but they don't think of it like they do a physical book. If you manage to get them to think of it like that, then they likely won't agree that 'residuals' are appropriate or selling in such a way as to avoid them is improper.
Now, the platform might be able to make some money (Ebay survives just fine, as far as a I know and paying to use their resale platform makes some sense).
But in the end, this idea (in my view) fails because the actual advantage it you tout 'transfer between platforms' only comes into existence if everyone adopts it, which means there's no competitive advantage to doing so. It might be better for the entire overall games marketplace/consumers (though I have my doubts) but it wouldn't be better for any individual retailer/developer, so there is no incentive to create/adopt it.
The trouble with that is that, even if Valve gave out a crypto token for everything in your library, there's no reason another storefront would want to honor them. Why would they? From the second store's perspective, that just means giving you a lot of games for free.
You also have a bit of a chicken and an egg problem: whoever does this first can't promise that the crypto tokens will ever be usable in another store, they would have to just hope that someone else hops on the bandwagon despite not having an incentive to do so.
I could see a company doing this as a way to attract new users, similar to Epic's free games - "We'll give you all your old games for free, just sign up and start using our service!" but that's going to be an expensive way to get users. And you don't really need a blockchain to do that, either - you could just look at Steam's API.
"Why would they? From the second store's perspective, that just means giving you a lot of games for free."
They're giving you a digital asset at essentially no cost to them, which you would not be inclined to buy from them anyway since you already own it. In exchange they have made it very likely that you will make new purchases on their storefront by reducing the cost of switching to nearly zero.
It's not that different to stores like Humble and GOG now offering Steam keys with purchases.
Developers would like it (speaking as a developer). We're not trying to sell you the game more than once, and we don't care what storefront you're on. In fact, smaller storefronts take half the cut that Steam does, so it's better for us if people switch. It would even make it easier to sell directly from our own websites where no-one takes a cut.
The main issues I see are the effects of the secondary market (mitigatable) and getting exisiting storefronts on board (very difficult, but not required to get value from this). But I'm not trying to lay out a foolproof business case, I'm just expressing surprise that I've not seen anyone trying this. To me this seems less dumb than basically any exisiting blockchain use cases.
DRM (Digital Restrictions Management) is fundamentally unsound for technical reasons. I do not think the block chain can be added to the mix to avoid single points of failure practically.
If you want to use the blockchain for resilience, this means that you may not depend on any central service interpreting it. While the block chain can establish a group consensus in the face of a minority of participants being adversarial, it can do little to help a single node whose traffic is controlled by an adversary. Passive verification would require parsing at least the block in which the purchase happened and then all subsequent blocks (to make sure that the product was not sold subsequently), and would vulnerable to attacks by someone just withholding the block of the sale and all subsequent blocks. To counter that, the DRM process could actually try to write on the block chain, verifying that someone is spending enough compute to calculate the next block. This still does not feel remotely feasible.
The other route would be that the copyright holder could commit to allowing anyone owning the appropriate token to download the game from bittorrent, and pre-commit to putting their product in the public domain if their blockchain verification website ever goes offline for more than 48 hours. This would be very resilient, but might not be a very good deal for the publishers.
In my model, the reason you can't resell used digital games is because neither stores nor game publishers want you to have that ability. A single upstart store can't buck the trend because the publishers wouldn't join so they'd have nothing to sell.
If they wanted to give you this ability, they wouldn't need blockchain, they could just do it. I'm not sure what problem you think blockchain solves.
You might think it's a way to tell what games a user has already purchased elsewhere, but stores already solve that problem when they want to. GOG had a feature for a while where they'd add certain games to your GOG account if you already had them on Steam. And Steam already gives game publishers free "Steam keys" that they can sell through other stores (which is effectively Steam giving you a free copy of games you purchased elsewhere), and Steam doesn't even take a cut, they just demand that you ALSO let customers purchase the game directly through Steam.
You might think that blockchain would mean that you don't have to trust Valve when they say John owns the game, but actually you still do, because in order to sell the game, Valve would need the right to add new copies of the game onto the blockchain. I suppose you could theoretically set it up so that each individual publisher adds copies to blockchain and then gives them to Steam to distribute, but that seems pretty similar in practice.
If you're worried that Steam might suddenly collapse and you'd lose the records of who owns what, then you just need an archive, not a whole blockchain. Steam could just periodically publish that data if anyone cared enough to archive it (it could even be digitally signed if you care).
Assuming that the stores and publishers actually wanted to do this, what part of it would they need blockchain in order to pull off?
I think frictionless transfer of ownership of Steam games would change the economics fundamentally.
At present, there is probably less than one play-through per primary sale, and the sticker price of games tends to drop by 25% per year after it is out, perhaps.
Assuming playing through a game takes four weeks, this would allow twelve people playing through the game in the first year alone for one single primary sale. There would still be a hefty premium on buying the game the day it comes out, bugs and all (just like there is a premium to buying a new car instead of a year-old ones), but afterwards the cost of owning a game for a month would just be a few percents of the sticker price.
Even for games people play long-term, frictionless transfers of ownership would change things. I have probably spend 100s of hours on FTL, but the truth is that I spend much less than 5% of my day playing it. A rental service renting out games so that each copy is played 50% of the time would drive down prices even more.
For video games, the costs of copying and distribution are basically zero (see bittorrent). I would assume that most of the costs actually involve creating the game. The sticker price for a frictionlessly saleable game would have to be much higher for it to be economically viable. I have no reason to think that I would spend less on games of this became the new equilibrium. (Do I sound like a bird arguing why his gilded cage is actually fine? I suppose I do. I am aware that "owning" a game on Steam is not the same as owning a physical object. Practically speaking, the chances of Steam going out of commission in any year are seem lower then me going out of commission, so I am mostly ok with it.)
I broadly agree that a system where you can sell used digital games would be very different from the current system. I'm saying the reason we don't have such a system is that game stores and game publishers do not believe it's in their interests, and that if they wanted to build such a system they could do it without blockchain.
I believe there's some work being done (Similar to Zvi's work on Emergents) on some CCG's / game pieces that can be owned / tracked independently of the underlying game substrate - meaning someone else could create another game system that you could use your assets in, which is a kind of cool concept!
See, these kinds of use cases make less sense to me because no-one is going to develop a game based on some other games assets. So you're just left with some inherent support for a secondary market for as long as the game is active.
"I created an NFT-based game, started before the term ‘NFT’ was a thing. I did it because I saw a particular practical use case, the collectable card game and its cards, where the technology solved particular problems. This successfully let us get enough excitement and funding to build the game, but ultimately was a liability after the craze peaked. The core issue was that the problems the technology solved were not the problems people cared about - which was largely the fatal flaw, as I see it, in NFTs in general. They didn’t solve the problems people actually cared about, so that left them as objects of speculation. That only lasts so long."
Nobody is likely to base a game's design off assets that they don't have the legal rights to, but the execution of subtle themes on an idea is extremely common. Consider all the different D20 RPG's that exist. Or, for another example, there are at least a dozen home-brewed MTG variants that I've played around kitchen tables over the last few decades, from slight changes to make multiplayer better to entirely new games that recycle MTG cards. The creators of these variants couldn't really profit from them (WoTC's legal department would shut down any attempts), but they still went to the effort of creating and distributing the rulesets locally, so I'm not sure I buy your argument.
If I could take my digital cards from Zvi's game and use them in another game's engine (whose creators presumably operate under a broad license that allows them to do so, which I don't think is the case with the assets from Emergents), then they would have more value for me than being locked to that one game. The existence of software like Tabletop Simulator also strikes me as good evidence for this kind of ecosystem's potential to emerge.
The problem is that digital games take years of hard work to make, while MTG variants can be thought up in a single fun evening (I have done both, for what it's worth).
Most game developers don't want to spend their years piggybacking off someone else's IP. Some do - see the Black Mesa project - but the original game needs to be pretty damn popular for it to make sense for devs to spend their time working on a spin off instead of making their own projects.
Modding is the usual solution for this - it allows variants to be created without having to build something from scratch.
Why would any new platform benefit from tracking who owns stuff on the blockchain rather than tracking it on their own database? The only benefit to the blockchain is you don't have to trust the person/organization who controls the database. Obviously they trust themselves, so there would have to be a strong customer demand for them to give up the control.
But then even if they store the ownership data on the blockchain, what does that actually mean? Someone still has to honor that ownership by eg. providing the game for download to the supposed owner. If you trust them to do that, why not trust them to keep all the ownership info on their database too?
My assumption was that customer demand would be strong enough to make it a selling point for new platforms. The same way that DRM free was a point of differentiation for GOG and Humble that helped them become established.
DRM free is a selling point that customers value because DRM is a pain in the ass in specific ways (and even that's pretty marginal- many customers don't care at all). What tangible benefit does a customer get from an NFT saying they own the game? They still need to trust the distributor to make it available for them to download. Can you describe a situation where having an NFT makes the customer better off than just having an entry in the seller's database?
I listed the advantages in my original comment, but I'll expand on them. Here are the problems it would fix:
1. I own hundreds of games on Steam. If Steam shuts down for whatever reason than I would lose them all.
2. I am locked into Steam as a platform because that is where all my games are. There are competitors to Steam that I prefer, but I avoid them because of the hassle of splitting my library up.
3. I own games on Steam that I would like to play on mobile, but I would be required to purchase the game again. That's not always worth it.
4. I can't resell games I've purchased over Steam.
DRM free helps a bit with some of these but not much. Eg. unless I keep all of my DRM-free games downloaded at once my library is still at risk of disappearing if Humble or GOG go under.
In a world where your game library could be on the blockchain, Steam going under would be no problem - you would just sign up to another platform that honours blockchain games. The advantage to the new platform would be that they could take a cut of new transactions you make on their platform.
I think that's enough reason to justify that this is a feature that some people would want. And I think it's enough to say that it's unusual that there seems to be a complete absence of blockchain platforms of this form.
Somewhat similar to my question last week on the open thread: has anyone been pulled out of a depression through falling in love with someone? I've been writing about my experience with schizophrenia over on my Substack, and Friday's installment featured the thing that sparked my interest in working out and taking care of myself, as well as for life in general. And that thing was falling for a girl who was unavailable. I'm not a romantic, per se, but I think there is something to the whole Lex Fridman-esque belief about the power of love to move people to good in the world. Interested in what ACX readers think about love and the power it has to move people both individually and collectively to make things better.
Sort of, but it's not so much "pulling out of depression" in the sense of a chemical imbalance. One experience was starting to fall for someone, noticing it, asking them out, getting a polite no, and moving on to being friends. In that case, the thing that helped was mostly the realization that I could still have those feelings at all, which gave me hope for the future. Another experience was being around someone for a weekend who was clearly subtly attracted to me just as I was to them. Nothing happened for a whole lot of very good reasons, but it was still a nice wake-up that such things are still possible.
Being perpetually single in my early 20s left me feeling quite depressed, in a way that was almost instantly solved by actually getting a girlfriend, even though that relationship lasted mere weeks. One's self worth can be rather strongly affected by other people's demonstrated valuations of you.
The flip side that you describe seems like a plausible kick to break out of a negative rut, but I highly doubt it would be sufficient by itself rather than as the last piece of a puzzle.
Yeah my depression was, for about 10 years, linked entirely with being perpetually in love with people who were not in love with me. It wasn't until I was not depressed that people would actually want to be in a relationship with me.
Absolutely. I have a patient whose father died when he was a child, several decades ago. the man has in fact had quite a difficult life, mostly because of OCD and depression. He much prefers to discuss his life in an "understand yourself" way, rather than in a cognitive-behavioral let's break patterns way, and most of what he has to say about the problem of being him is that growing up without a father prevented his forming a feeling of having a coherent, adult self. So in the last 6 months the man has met and fallen in love with a women who is a good match for him in interests, a good sexual partner, and pretty, smart and kind. He's happy. And now he's saying to me, "OK, I lost my father -- but why does it follow that that ruined my life? I can just think of it as, it is what it is."
I entered a years-long depressive state in the aftermath of love. I eventually made my way out via building friendships with other people and a single remarkable dream that I had one night. I don't think, based on my experience, that it would have even be possible for me to have had experienced love during the depths of my depression, for the same reason that I wrote no poetry (but was astoundingly prolific both before and since).
I wrote a history of how independent courts gained the power of judicial review in common law systems. It's a history focused on the institutional questions -- how do the courts internally discipline themselves, and how do they use this discipline to influence other branches, despite lacking the power of the sword or of the purse -- and so it's rather different than the standard case-focused histories of this which lawyers tend to write.
This is my second (and last) time posting this in an open thread before the next classifieds thread, but my sister (trans-woman, 24 years old) has recently moved to Seattle and is striking out left and right trying to find a job. She graduated a year ago with a bachelor's degree in actuarial sciences, has some coding experience, and has a good chunk of capital to use on finding the right position (rent for multiple months while unemployed or in a poorly paid internship, headhunter fees, etc.).
Didn't mean to bury the lede like this, but does anyone have any advice? Specifically in Seattle, but Portland or the Bay Area would be similar enough that I'd be interested. Are there agencies or single contractors who specialize in job-hunting that you've worked with? Does your company have openings for someone with that kind of CV? Have you been in a similar position and found a strategy that worked well?
Feel free to respond in a comment below or email me at: 7o2wzrybd (at) mozmail.com
Coding bootcamps, if they're any good, should have connections to hiring tech companies for their graduates. Any bootcamp that wants to teach you at their own expense must genuinely believe that they'll land you in a job afterwards ( of course they'll keep a cut of your pay, but it may be worth it to keep the incentives aligned ). If the bootcamp wants you to pay them, take a critical look at how many of their graduates find work and make sure they'll actually put their network at your disposal. After your first job, people will take you more seriously, you'll have your own professional network, and it'll be easier to get your second job, even if you realize you want to do something besides coding.
as both are probably doable but both might require a bit of specialized effort investment to increase the chances (as presumably she tried out and didn't immediately get anything)
She's undecided, a little demotivated by the lack of success so far. Either one of these routes described above or below (coding - boot camps/actuary - actuarial exams) may be the best route forward. Thanks for the thoughts.
hope she has people/therapy/resources to allow her to not feel pressured over it. yeah the job market is tough.
but viewing it as being about getting whatever job instead of making a thoughtful decision about the next arc of her life that will probably last at least a few years and will have resume effects for way longer.. might not be wise.
not everybody is privileged, but this is the kinda thing where privilege, not having the crazy pressure as well as confidence that one can have reasonable success in whatever reasonable path they choose, is quite material for making quality decisions with significant impact on one's life.
Is she interested in performing actuarial work? My brother was an east coast actuary and the way he described it there is a formal track of actuarial exams and completion of a certain number of exams essentially guarantees employment.
I’ve been thinking recently about the concept of the “deep state” through the lens of Franz Kafka. While it’s common to compare the current state of the world to Orwell’s 1984, I believe two of Kafka’s novels, The Castle and The Trial, contain themes which are more relevant to understanding the insidious power structures of society. Kafka imagines a world in which bureaucracy is a sort of Frankenstein, an unthinking malevolent force created by humans which has, without anyone really noticing, enslaved the humans. The law does not exist for man; man exists for the law.
One might compare the law to the rules of an algorithm abided, not by the pure logic of microchip circuitry, but by fallible humans. In Kafka’s universe, the chief weakness of man is stupidity in all of its manifestations: incompetence, immaturity, narrow-mindedness, pettiness, incomprehension. Kafka’s anti-hero, K, is not stupid. But everyone he encounters: his work assistants, cops, lawyers, clerks, officials of The Castle, magistrates, the mayor, schoolteachers, tavern workers, the townspeople... are drowning in stupid. The world is a bafflingly idiotic yet complicated place full of rules which, at every turn, work against an honest, earnest man who merely wants to do his job and stay free of prison.
Unlike with Orwell, Kafka’s novels are not dystopian warnings to the world. How can one warn the world that the world is too stupid to function according to its own rules? It isn’t the rules which are stupid; it is the people living by and enforcing them! It is the cops and lawyers, the judges and juries. And it isn’t a stupidity which one can be educated out of. A court magistrate may have a sophisticated understanding of the law, for instance, but is also petty and narrow-minded. Stupidity is a cancer that has spread throughout the organism.
I started thinking about this subject while pondering AI risk. It struck me that we’ve already created a monster that is much like an alien intelligence running our world which is indifferent to the plight of humanity. It is all our bureaucracies!
Could you help me understand your post a little better? You seem to mention the interpretation that Kafka was writing a critique of bureaucracy, then downplay it in favor of the interpretation that Kafka was writing about a (negative) experience of humanity in general, before offering your own original critique of bureaucracy. But critiquing bureaucracy fits better with the original interpretation, which you had seemed to be setting aside. Have I misunderstood?
First, Kafka isn't critiquing anything. He wrote novels that are essentially extended nightmares. The nightmares are about dealing with various hells of bureaucracy: in The Trial, Joseph K is arrested and doesn't know how to defend himself because he has no idea what crime he has been charged with. In The Castle, K is a land surveyor who has been called by The Castle to do a job but can't figure out what the job is; try as he might, he can't work his way through the bureaucracy of The Castle to get information pertaining to his job.
These are absurdist novels.
My argument, which is independent from what Kafka may have ever thought, is that these nightmares of bureaucracy do offer insight into the way the world really works.
Kafka's insight, as I see it, is that the weakness of bureaucracies is that they are run by humans, who mostly aren't competent for the task. In other words, the main problem with bureaucracies is not that the rules and regulations are themselves stupid, but that the people charged with enforcing them are. In Kafka's novels, at least, the people working inside a bureaucracy are absurdly stupid.
I then attempt to compare this world in which a rule-based system (a bureaucracy) is unaligned with humans because (ironically) it is run by humans--with a world run by an unaligned AI.
In my interpretation, the two biggest influences on Kafka's writing (fiction and journals) were his basic experience and a corollary of that experience.
- The basic experience was that his world didn't make sense.
- The corollary was that the right thing to do could not be established.
I take this as mostly Kafka's tacit knowledge, i.e. a pre-theoretical attitude, rather than coming from philosophy or theology. Kafka's attempts to theorize this attitude in his journals tend to be self-consciously absurd. (Contrast that with Samuel Beckett, who seems to have taken Jungian ideas seriously in an attempt to explain his own basic experience.)
Kafka was an administrative lawyer, so it's not surprising that he sometimes expressed his underlying attitude with fiction about law and bureaucracy; he was Jewish, which probably gave more valence to law in particular. But I think law and bureaucracy were basically symbols for him, so I don't separate law from bureaucracy from bureaucrats from protagonist when reading The Trial and The Castle. Nor do I see much difference between the bureaucratic fiction and the other fiction.
You may endorse the Walter Benjamin interpretation of Kafka as [sic]:
>not so much a failure as a comic figure. Kafka is man 'whose fate it is…is to be surrounded by clowns.'
Whereas I would modify that by saying that Kafka was a clown who knew he was a clown, and mostly did not want to be a clown, yet did not know how to be anything else. It seems to me like he amused himself with the (pleasant?) daydream of being a non-clown tormented by clowns strictly external to him. Superficially, that's The Trial and The Castle, among others.
So I largely agree with Michael Wood's view (and apparently Stanley Corngold's):
>In this sense it is surely a mistake, as Corngold suggests it is, to pit Kafka’s bureaucracy against its ‘hapless supplicant’. He is bureaucrat and supplicant, perhaps more bureaucrat than supplicant.
None of this is to disagree with your point that unaligned AGI might be more familiar to us than we think. This is an interesting idea and I hope you post more about it.
I wouldn't call it "stupidity", so much as conflicts of interest and tunnel vision. Rules are made that ignore some parts of reality, and contain incentives that aren't acknowledged. It's not a lack of intelligence per se, it's what that intelligence is focused on, and what that intelligence studiously ignores.
Is this meant to be reassuring? Bureaucracies are a slow cancer destroying all that is valuable. I expect AGI to be much faster (molasses are usually faster than bureaucracies, it's not a high bar), and not even theoretically containing components whose conscience might wake up and cause them to heroically defy the machine they are a part of.
common counterargument is that "indifferent" is an overstatement, human institutions (in the free world to a larger extent) are to a decent extent shaped by aggregate human preferences and subject to various feedback loops making them more human-compatible
About the birth rate problem, it's odd that there aren't companies or businesses or even business models coordinated around trying to increase birth rates.
If you think of small fundamentalist religious cults as a business run by the cult leader then there are dozens and dozens of these business just in Utah!
Oh, but remember a few years back when Apple offered egg-freezing to its female employees? They will graciously permit you to have a family after they've squeezed all the use out of you and you're no good to them anymore, so you can go ahead and waste your time with babies:
I am very damn cynical about that because companies don't want staff taking large chunks of time out, which is why maternity leave is generally a legal imposition by the state and not something voluntarily offered. A business working on increasing the birth rate would probably require its staff not to be going out on maternity/paternity leave too much, which would be an instance of saying one thing and doing another.
Also, the governments are 'trying', but aren't willing to pay very much. We have fairly strong evidence that direct cash payments for babies help but they need to be much larger than governments typically offer (orders of magnitude larger, if you want to get TFR back above replacement)
They don't need "to pay much", rather they need to pay less.
Just stop paying retirement pensions, stop subsidizing women higher education (most of higher education degrees are useless, and womens are to larger degree) => TFR up with no extra dollar spent.
So only men go to college, no women. Great! And, uh, who are these college-aged women marrying to have those babies? Older men, I presume?
So now you've got a bunch of 18-22 (or older) men who have no chance of finding a permanent relationship until they're 30-40 and earning high salaries. That'll work out fine. As will having a lot of women out of the workforce, because it's companies and the economy that want women working.
Will the children be okay with the idea of taking care of Mom and Dad who will need their financial support in old age? Already you have families putting parents into nursing homes rather than bringing their elderly parents to live with them. And of course, if the potential parents decide that instead they both need to work and save, in order to live without pensions in old age, and having kids is an expense that means taking away money they could be saving.
Society has done its damnedest for decades to undermine child-bearing, stay-at-home housewifes, and the idea of 'too many people on this fragile planet'. You can't simply reverse course that easily to bring the birth rate back up, especially when the people you want having more babies are the exact people who don't want large families.
as Zvi has pointed out, they also aren't willing to undo any of the myriad of things they are currently doing which discourage having children either. (which would, potentially, have a much larger effect size than paying people.)
This is not a problem for any specific couple. Any couple that wants kids and has the resources for it, can have more kids.
So:
1) or they don't want kids - Business can't help.
2) They don't have resources for more kids - Business can't profit, so can't help.
So this is a jointed-action problem, where we all agree there should be more kids, but no one want to have more of them/support those who do by greater taxes, etc.
I've seen so, so many discussions of low births rates in these comments without anyone actually looking at stated reasons people don't have kids. I'm not saying you need to 100% trust exactly what people say, but it sure seems a decent place to start, and for many people, the reason they don't have kids / don't have more is housing and childcare.
There are cultural factors as well (having three kids wasn't a big deal for my two working parents because they would just kick us outside all day and we walked ourselves to school, both of which would get child services called on them these days), but simple inability to afford the kind of childhood people were given where the jobs are is a pretty straightforwardly understandable reason.
I, for one, don't agree. We had a vibrant culture when the population of the world was 2 billion. As technology stands right now, we've overrun the ability of the planet to absorb the CO2 our civilization emits. I'd rather stop deaths from aging - but we don't currently have the biomedical technology for that. As technology stands, I have no objection to a slow shrinkage back down to 2 billion. We can defer any questions about incentivizing births to that point.
And kindly remember, as of this moment, the population of the world is still _growing_, not shrinking.
You have a point. With current technology there aren't a lot of options. The population cannot keep exploding. We've overrun the ability of the planet to absorb the CO2 our civilization emits. Yes, with a slowly declining population for a century or so (setting aside AGI or other technology advances to simplify the discussion) the population will be mostly middle-aged during that period (presumably slightly younger when it stabilizes (itself a long discussion)).
Its not odd if your ultimate goal is to depopulate the world to where you have just enough workers to undertake work AI can’t do whilst exterminating ‘excess’ consumers gobbling up resources you want...
"exterminating"? As far as I know, even in the most extreme case of South Korea, the depopulation is happening due to deaths from old age, not gas chambers. Now, I'd like to see aging solved like smallpox was solved, but it is a very hard technical problem and unlikely to happen soon.
Does anyone have experience with EDMR they are willing to share? The wikipedia page positions it as pseudoscience and I'm wondering where it falls on the spectrum between "pure snake oil" and "worth trying". I have a traumatic experience from two years ago that still jumps to mind multiple times a day. Thanks for any pointers!
I did EMDR to deal with negative thoughts, and I would say it worked. I believe it worked by creating a disturbance in the negative thought-pattern. I suspect "The Work" of Byron Katie works by creating an alternative to the negative thought pattern. I think CBT sucks. For example, I don't see how one can become less judgemental by judging one's thoughts as distorted. I've also found a mindfulness practice to be useful for dealing with negative thoughts. I welcome critique.
Yes. It appears to be working for me. My PTSD is not as bad as it used to be. I began treatment a few years ago, and saw improvement, then the covid-19 lockdowns hit and treatment stopped and I saw stagnation and a bit of backsliding, and now treatments have started again and I'm once again seeing improvement.
I have no solid idea why it works. It might as well be witchcraft, and I could easily imagine some poor old woman from a few hundred years ago getting burned at the stake for discovering this. But here's some thoughts.
First, one thing I noticed when I started EMDR is that I'd stopped dreaming at some point. Prazosin made the nightmares go away, but in the years after that, I don't think I had the sorts of dreams I'd had before. Maybe it's that my life had become static and circumscribed, and dreams help integrate new stuff, or maybe it's that my memory was damaged by both PTSD and also some treatment for depression. And it's probably not that I completely stopped dreaming, but rather that whatever dreams came were below the threshold of my notice. But after my first few sessions of EMDR, I began to have vivid, strange dreams, and I realized that I hadn't been having any for a long time. Now those vivid dreams have died down, but it was a distinct change. So on the model that dreams are processing of experience, perhaps that processing had become blocked somehow, and EMDR is deliberately re-starting it, like crank-starting an old motor, or priming a well pump. Perhaps the eye movement stuff has some connection with REM sleep, although I've also heard that it might simply be placebo.
Another idea is that EMDR is taking advantage of quirks of the way our mind-emulations run on meat brains. Current theory seems to be that every time we access a memory, we also modify it. There may be enormous emotional content stored along with certain memories, which are frequently accessed, and in ways that maintain or intensify the emotion. By carefully and slowly accessing those memories in a calm and safe situation, repeatedly, we can lower the emotional valence of those memories until they're merely normal. On this view, the eye movement might serve as a harmless distraction, a way to keep our conscious attention occupied while the rest of our brain re-lives the experience.
Or perhaps the whole thing is slowly retraining the amygdala to relax and not freak out whenever the memories are encountered. I've noticed that I've completely lost normal adrenaline responses to things like assault, severe bodily injury, rollercoasters, and that type of thing. However, something that reminds me of something that reminds me of the bad stuff can flip me out, even if I'm just walking in a park on a sunny day. So maybe part of what's going on is that I'm slowly retraining my brain so that the fight-flight-freeze thing doesn't kick in when it's not appropriate. (Yeah, yeah, "define 'appropriate'".) I do wonder whether the normal adrenaline response will come back, but that's not a problem for today.
One thing I should mention is that EMDR can have very strong effects on my general mood over the week or so following a session. The first time I did any serious work with EMDR, I had a high like the first time I tried amphetamines (5mg generic Adderall), and like the amphetamine high it faded quickly and was not repeated. That time I came out of the session feeling extremely fatigued, such that I went home and took a nap for 4 hours, then woke up feeling amazing. (This is odd, because I almost never nap, because I normally feel horrible and groggy for the rest of the day.) Sometimes I've come out of sessions feeling lousy for the next week. Other times, when we did lighter work, I've come out feeling happy. So be prepared. If you're working, you might want to schedule these for the end of the day, maybe even Friday afternoon if possible.
Also, I personally would not describe it as "trance", at least not the way either of the therapists I've seen have done it. At least, not any more than I'd call a normal PTSD flashback a "trance". But then, I've never been in any state I'd use the word "trance" to descrive. Other people's subjective experiences may differ, of course. :-)
EMDR is a form of trance that works really well for a single trauma or a few isolated ones, at least for a large fraction of people. It works a lot better with a good practitioner. Also, consider ART instead of EMDR (https://acceleratedresolutiontherapy.com/), it seems to work faster for many people, and, anecdotally, also works for CPTSD, where EMDR ends up pulling on too many traumatic threads to be really useful. Again, YMMV, if it feels like something is off after you tried it, best reevaluate the next steps.
Psychologist here. I have not had EMDR, but know many people who have. Most say it helped then. In my opinion, the active ingredient in EMDR is the remembering and recounting of the trauma the person does while being guided through the eye movements. I think the eye movements mostly provide a theory of why the process works, and a task to do to that helps the person endure the process of reliving trauma. However, so long as EMDRit works for trauma, it's fine to do. The only kind of EMDR I think is truly not a good idea is EMDR as treatment for negative events that are not traumas (traumas are sudden, overwhelming, dangerous and disastrous events that most people never even experience: seeing deaths in combat, being in a major fire or automibile accident, being raped, etc.). I do not think EMDR is an effective treatment for phobias, bad but not traumatic memories (funerals, break-ups), bad experiences that were not an incident but a bad circumstance that stretched out for a long period of time (being broke, hating yourself, suffering a long depression, having a job with an asshole boss).
My partner is currently doing EMDR as a patient. She emphasizes that it is different from and more intense than CBT, which is the other form of therapy with which she has the most experience. Where the latter is about interrupting an emotional process with an intellectual one, EMDR involves direct confrontation of trauma on an emotional level. It will make you feel worse before it makes you feel better, but if you can afford that then EMDR is very effective.
It's definitely not snake oil. It very effectively produced a dramatically altered state of conscious, when I tried it. It's a much more intense experience than I'm used to having in a therapists office, and it was comparable to other heavily mind altering experiences I've had. I get the impression set and setting are probably just as important here as they are with drugs, rituals, etc; I didn't particularly trust my therapist at the time so it was not a useful experience for me at all, but I don't think that's a problem with the treatment and I could absolutely see how it could be effective.
A few times over the last few years, commenters have put "AI" and "Moloch" together (e.g. from a cursory search: Maxwell IV and Mr. Doolittle). I'm curious if others, seeing the rather impressive evolution of AI-safety discourse (including AI-not-kill-everyoneism, AI-not-make-everything-worseism) would like to revisit this comparison.
I.e. if you go back and read Ginsberg's original and Scott's commentary on it, do you now more than before get an "aww shit advances in machine learning is gonna accelerate this" feeling and if so do you have thoughts to share on it?
Notably, if you reread the original Meditations on Moloch, you'll note that Scott stresses that Moloch hasn't won. Humans again and again defy the explicit incentives in order to preserve that which is good and beautiful in the world. Unfortunately, an artificial intelligence won't have a human soul at its core prompting it to do good and defy its explicit instructions, so we'd better get a lot better at crafting instructions that don't destroy all value. when optimised.
Well, you see people worrying "making the AI too powerful too soon might kill all humans" and the companies... not really saying out loud, but thinking "but if we don't make the AI as powerful as possible, as soon as possible, our competitors will take over our market share".
That's Moloch in action: increase the probability of human extinction by 1% as a sacrifice, get a temporary advantage against your economical competitors. Repeat, if necessary.
I have a somewhat specific ask: I’m going to be living in Washington DC for the next year or so, and during that time, I would love to meet people who are also interested in reforming US college tuition wholesale--especially if you’re already involved in something along those lines. I specify DC because it’s where I am and I assume it would involve something on the policy side, but am certainly open to talking to anyone interested in possible solutions, regardless of where you’re from. I’m especially interested in people with backgrounds in insurance (bonus points for health insurance), antitrust, or university administration. If you are interested, my email is harrisongfriedman@gmail.com.
There's no way to remember everything in a book, but you can do things to remember more. The best way to consolidate things in memory is to use them. At the end of sections (chapters or whatever), put the book down & *do* something with the info. If it's material that lends itself to problems, make and solve some problems. Or draw a diagram of material that was presented just in words (draw it without looking back over the section, then afterwards look back to check its accuracy.) If it's, say, a history book, think over what you read and ask yourself questions like, "what thing, if it had been different, would have completely changed the impact of this event?" Make a few notes on the answer. Also try to have some conversations with people about the book.
Why are you trying to memorize the books? Is it just the information or is it the actual verbiage?
I haven't used spaced repetition for anything, but had moderate success with mental pictures. Does Anki let you draw pictures? Draw some pictures that evoke the sentence you're trying to remember.
1) Use Readwise. It' a paid subscription service to manage notes and highlights in whatever you read. Book highlights you take are automatically presented to you in an Anki-style review, and if you like the highlight, you can turn it into a proper Anki-style card à la Cloze Deletion or similar. Disadvantages include the monthly fee and that the whole review process is only minimally customizable, in contrast to Anki.
2) I've heard about a bunch of attempts to use GPT3 or GPT4 to auto-generate Anki cards. I don't know if any of them have come to fruition, or work well, or have decent user interfaces, etc.
So how about that excess mortality? Seems like kind of a big deal.
Anyone have thoughts about this?: https://twitter.com/mechanical_monk/status/1662595973124685827
I recently moved to Israel and am looking to make new (English speaking) friends. Any suggestions about people to contact/meetups or events to attend are welcome. I would also be interested in non rationality focused stuff like book clubs etc
Still in Israel?
Not at the moment but might be back in February.
Best of luck 🙏🏼
Feel free to ping me if you're back
Hey! I am returning to Israel on Jan 28th and hopefully will stick around for at least a few months. Interested in meeting up?
Will do!
LW/ACX Saturday (5/27/23) Values and Shard Theory
Hello Folks!
We are excited to announce the 27th Orange County ACX/LW meetup, happening this Saturday and most Saturdays thereafter.
Host: Michael Michalchik
Email: michaelmichalchik@gmail.com (For questions or requests)
Location: 1970 Port Laurent Place, Newport Beach, CA 92660
Date: Saturday, May 27th, 2023
Time: 2 PM
Conversation Starters:
Value is fragile by Eliezer Yudkowsky https://www.lesswrong.com/posts/GNnHHmm8EzePmKzPk/value-is-fragile#:~:text=Value%20isn't%20just%20complicated,single%20blow%22%20will%20do%20so.
Shard theory of human values by Quintin Pope, TurnTrout https://www.lesswrong.com/posts/iCfdcxiyr2Kj8m8mT/the-shard-theory-of-human-values
Shard Theory: An Overview by David Udell https://www.lesswrong.com/posts/xqkGmfikqapbJ2YMj/shard-theory-an-overview
D) Card Game: Predictably Irrational - Feel free to bring your favorite games or distractions.
E) Walk & Talk: We usually have an hour-long walk and talk after the meeting starts. Two mini-malls with hot takeout food are easily accessible nearby. Search for Gelson's or Pavilions in the zip code 92660.
F) Share a Surprise: Tell the group about something unexpected or that changed your perspective on the universe.
G) Future Direction Ideas: Contribute ideas for the group's future direction, including topics, meeting types, activities, etc.
Utilitarian question I haven't seen addressed: Supposing it's possible to be in charge of that sort of thing, is there a good way to think of population size for emergencies?
I'm not talking about life-is-suffering anti-natalism, I'm thinking about how it's really hard to move tens of millions of people away from a volcano or a tsunami. Or, of course, both.
To me, it seems the best way to think about it is that the person who decides to live in a place too dense to evacuate already bears much of the cost/risk of that decision (as well as of the benefits) and that we should thus naturally end up at a reasonable equilibrium if we just laisser faire.
Probably depends on what you mean. Buildings have fire codes with mandatory exits and maximum occupancies. Equivalent maximum occupancy for a city-wide disaster would be something like road width, how many cars can be travelling out of town at the same time. Flooding involves climbing nearby hills, volcanos involve... um...
So it depends on whether you're talking about limiting occupants, or improving exit routes, or just predicting how many people die when the city hits the bottleneck.
Hurricane Rita demonstrated that Houston, at least, is impossible to evacuate by car (and certainly by any other means). An acquaintance of mine headed to Austin, and made it halfway (about 70 miles) after 24 hours.
https://en.wikipedia.org/wiki/Hurricane_Rita#Mass_evacuation
You would have to ask both questions.
In a general sort of way, how does your mind work?
I read a post from someone who is normally thinking several things at once, but can't/doesn't when ill. (I hope this is vague enough-- they aren't the only person whose mind is like that.)
I think my normal state is more about thoughts leading quickly to associated thoughts-- sometimes distantly associated-- rather than multiple thoughts simultaneously.
I've put a lot of effort into actually being able to pay attention to details of movement and sensation for qi gong.
And I'm brought back to when I became able to pay attention to what I've called That Boring People stuff-- an Alexander Technique teacher had students look into each others' eyes for maybe two minutes, and it considerably broke my compulsion to skid away into not noticing people.
Did you know, the main character in Delany's _Triton_ (_Trouble on Triton-) is an asshole? The thing is, I read the book before I was at that class, and I had a hard time figuring out why Delany was so hard on Bron. After, that, how self-absorbed Bron was is obvious in every sentence. Maybe every sentence, since I then found him so annoying I haven't been able to reread the book.
Anyway, it also affected the way I saw people, but not in a way which is as clear and easy to write about.
Anyway, how does your mind work? At its best? At other times?
It feels like there's a lot of activity going on under the surface, and occasionally thoughts bubble up. It's not a rational process, so if I care about correctness I need to do some double-checking, but it's a lot easier than trying to derive thoughts from first principles. And if I'm looking for a thought on a particular subject, I can try loading a bunch of background information into my head, and maybe rehearse some arguments, and then try fishing, but there's no guarantee that anything will bite. But often it does work; I circled back to this question after about 2 hours, and found I could start typing a response. :-)
These days, I'm not as smart as I used to be, which apparently means slightly less stuff bubbles up, and often it needs a lot of checking, and the fishing process takes a lot longer.
I've learned a lot about how my mind works from meditation. Because when I meditate, trying to focus on, say, my breath (the object of focus can be many things, the breath is the most familiar), it turns out my mind soon gets distracted by a stray thought. But because I'm trying to focus on something specific, I'm more aware of what that stray thought is, and what rabbit holes it leads to. Eventually I became more aware of my trains of thought in general. These were sometimes whole scenarios that could play out while I'm in the middle of some activity, say riding my bike in traffic. This turns out to be quite useful because, odd as it sounds, I find out what's "on my mind" by actually noticing what I'm thinking. For example, after a confrontational discussion with a family member, I found myself replaying the conversation, with alternative lines and behavior on both parts; and I discovered that my resentment increased as a result of these thoughts. So my discovery was that my emotions are often driven by invented scenarios playing out in my mind. There are lots of variations of this; sometimes I'll criticize myself for something done imperfectly, and I'll "reply" defensively in a sort of conversation with myself. In case this is starting to sound like I'm weirdly out to lunch, I think most of us have had the experience of reading a page in a book and realizing at the end of it that we were thinking about something totally unrelated and have no idea what we just read. No? Similarly, we can be doing stuff while lost in thoughts unrelated to what we're doing.
At its best, I'm so focused on what I'm doing that that these internal thoughts are silent, or nearly so. This is often how people describe being "in the zone". That's why I like challenging sports. At my worst, I can be in a conversation with someone and barely hear what they're saying because I'm thinking about what I think they're going to say and what I want to say. I've found conversations can get frustrating (or worse) at such times.
I had a habit of thinking "what if someone said something annoying, how would I say?". Mercifully, that pretty much went away when I unpacked how ridiculous it was.
And yes, I think there's a pattern-- not all that rare-- of what I call optimizing for outrage.
It true, sometimes it takes an extra level of thought to find out what you're thinking.
Reactive. Find a thing, focus on the thing. When the thing is done, default to memories of weird media until a new thing comes up.
Not much eye for detail. Had a fun time where I suddenly noticed brown spots on my face and was worried I'd caught something, then saw a picture of myself from grade school with the same spots. Just didn't notice them for twenty years.
My thinking is linear, like inner speech. At best, it is silent, quick, smooth. At worst, it keeps stopping, needs to be pushed forward by louder inner speech.
Careful with that advice. Mindfulness meditation isn't good for everyone, though at least you were recommending just doing it a little.
Can you cite that? I haven’t found anything like that for light “mindfulness” practice (westernized MBSR etc), most of the side effects I have seen measured arise when people go deeper into insight territory like vipassana retreats or 1h/day of sitting.
15 mins a day seems extremely well tolerated.
Thanks for checking. I think that in general, people who aren't suited for mindfulness meditation avoid it, but it's a real problem when an employer insists on mindfulness meditation.
Re-Evaluating GPT-4's Bar Exam Performance
"First, although GPT-4's UBE score nears the 90th percentile when examining approximate conversions from February administrations of the Illinois Bar Exam, these estimates are heavily skewed towards repeat test-takers who failed the July administration and score significantly lower than the general test-taking population. Second, data from a recent July administration of the same exam suggests GPT-4's overall UBE percentile was ~68th percentile, and ~48th percentile on essays. Third, examining official NCBE data and using several conservative statistical assumptions, GPT-4's performance against first-time test takers is estimated to be ~63rd percentile, including ~41st percentile on essays. Fourth, when examining only those who passed the exam (i.e. licensed or license-pending attorneys), GPT-4's performance is estimated to drop to ~48th percentile overall, and ~15th percentile on essays."
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4441311
I had also wondered whether GPT's scores reflect prior knowledge of the exam. For instance, an earlier version of GPT might have taken the bar exam a year or two ago, and later read information about its own performance or about the correct answers to various items. There is also probably a great deal of bar exam study material and practice tests online. Might those have been helpful to GPT?
This was also discussed here: https://aisnakeoil.substack.com/p/gpt-4-and-professional-benchmarks
_OpenAI may have violated the cardinal rule of machine learning: don’t test on your training data._
_To benchmark GPT-4’s coding ability, OpenAI evaluated it on problems from Codeforces, a website that hosts coding competitions. Surprisingly, Horace He pointed out that GPT-4 solved 10/10 pre-2021 problems and 0/10 recent problems in the easy category. The training data cutoff for GPT-4 is September 2021. This strongly suggests that the model is able to memorize solutions from its training set — or at least partly memorize them, enough that it can fill in what it can’t recall._
That sounds oddly like overfitting.
Will neural nets in 2030 be trained using more than 10^27 FLOPs?
Current NNs need FLOPs _and_ good networking to train.
We got 10^15 FLOP/s with good networking in 2008 (Roadrunner). We got 10^18 FLOP/s with good networking in 2020 (Fugaku) or 2022 (Frontier) depending of if one insists on f64. (For ML proposes, we do not). If we irresponsibly extrapolate, we should have ~10^20 FLOP/s with good networking by 2030.
Given 10^20 FLOP/s, 10^27 FLOPs would take ~4 months.
However it is unlikely that networking will continue to scale, and unlikely that FLOP/s/watt will continue to scale, short of unexpected advances in material science.
However, if distributed training algorithms work out, and we can burn more electricity on more hardware, distributed, with slower networking. There is a good chance we will be spending 10^27 FLOPs on a single NN.
Market: https://manifold.markets/Amaryllis/at-least-one-of-the-most-powerful-n-18faef58b570
How much difference will it make? I'm very much not an expert, but my understanding is that the performance of neural nets is affected much more by the quantity and quality of the training data than by the amount of work used to train them.
I probably listened to the same podcast Trebuchet did. I can't remember the figure, but the performance improvement when the AI is told to extend its period of processing the question, and carefully search all possibilities, was quite enormous. More than one order of magnitude, as I recall. I believe the Podcast where this was discussed was No Priors, the interview with Noam Brown about the bot Cicero.
One order of magnitude in what? The things we care about can be approximated as grade level of writing, apparent IQ, percentile on the Illinois bar exam, etc, and I don't think any of those are at all linear with any metric that spans orders of magnitude.
I suppose "hallucinations" becoming 10x less frequent would be both meaningful and measurable.
OK, I found the transcript of the podcast. It was 5 orders of magnitude, and the thing measured was amount of training needed to get an equivalent performance. Noam Brown, talking about Cicero, his AI that learned head-to-head Texax Hold'em:
"And we had them [human world champs] play 80,000 hands of poker against our bot. And the bot actually lost by a pretty sizable margin. And it occurred to me like during this competition that the way the humans were approaching the game was actually very different from how our bot was approaching it. So we would train our bot for like two months leading up to this competition, you know, on, on a thousand CPUs. But then when it came time to actually play the game, it would act instantly and the humans would, would do something different. Like, you know, obviously they would practice ahead of time, they would develop an intuition for the game, but when they were playing the game against the bot and they were in a difficult spot, they would sit there and they would think, and sometimes it was like five seconds, sometimes it was like a minute, but they would think and that would allow them to come up with this better solution.
And it occurred to me that this might be like something that we're missing from our bot. And so I did this analysis after the competition to figure out, okay, if we were to add this search, this planning algorithm that would come up with a better strategy when it's actually in the hand, how much better could it do? And the answer was it improved the performance by about a hundred thousand x. It was the equivalent of scaling the model, like scaling the number of parameters, scaling the training by a hundred thousand x" . . . "and what was really interesting there is the bot, we did another competition, the bot won and that bot cost under $150 to train if you were to run it on like a cloud computing service. And I think that shows that this wasn't just a matter of scaling compute, it, it really was an algorithm breakthrough and this kind of result would've been doable 20 years ago if people knew the approach to take."
He gives other examples of the same principle -- changing that AI's algorithm so that it "thinks" more before responding improves performance enormously, makes possible degrees of improvement you can otherwise only get with 10, 100 or 1000x as much training data.
Here's transcript: https://sarahguo.com/blog/noambrown
I like the No Priors podcast.
Indeed! Current NNs are beginning to run into data limits. However they are _also_ using compute close to the limits of current compute clusters. For now. We are currently in somewhat of a transition from compute bound to data bound.
One could imagine a future in which we have lots of compute, and fast networking and/or good distributed training algorithms, but the market resolve NO, because we ran out of good training data, and it is not economically profitable to spend more compute for little benefit.
That said, I suspect that with some effort, more data can be found. And at some point, people will start training on large amounts of video/audio, and that will absorb lots of compute.
If one expects distributed training to work out, this is a market on the availability of training data.
The ACX meetup in Cambridge was good,
Now Scott talks about lots of interesting things, but its clear that AI risk was on most people's minds. (Much less interest in, e.g. Scott's post on culture-specific disorders).
Overheard:
Person A: I think you're a cult
Person B: Go on, tell us where you think we're wrong about AI risk.
Person A: You're probably right about AI risk, but you've still managed to turn it into a apocalyptic religious cult.
Person B:L: Yeah, ok, we're a cult.
I was (mostly) joking, but there's always:
http://www.neopagan.net/ABCDEF.html
The Advanced Bonewits’ Cult Danger Evaluation Frame
Now, when my studemts started talking about this (in a different context from whether Aisk anxiety is a cult...
Me: It's great, you can even use it on organizations that aren't religions
Student: I'm going to use it to evaluate this academic department
Me; For, example, yes.
Me (having seen our students filled in cult questionairre on our department): Really, you're allowed to disagree with the faculty, you know. Probably it;s even required at the phd level: explain why we;'re totally wrong about something. at thesis length, and claim your phd...
Ah, the famous c-word that no one can define, but everyone is sure it applies to other people.
Related: https://www.lesswrong.com/tag/cults
It is not true that no-one can define cults. I do not know why everyone is so impressed with the "anti-anti-brainwashing" literature. They are pretending to debate "brainwashing" in order to avoid having to debate the normative concept of freedom of association.
Basically, Christianity is anti-family to the point where it doesn't even understand the word "family" correctly and thinks it's a code word for sexual morality. This creates a bizarre situation where people are allowed to just leave their families to pursue intense monastic lifestyles that deliberately alienate and even outcast them from mainstream society, and the family has no right to seize them back. I cannot emphasize enough how abnormal and counterintuitive this is, and very few Christians actually seriously normatively believe that it should apply to non-Christians or even the wrong kind of other Christians, it is something you have to be educated into. So of course they try to get their own kids back when they join Hare Krishna or whatever nonsense, and then wiser and more politically-connected Christian leaders have to step in and disallow it just to protect their own position in a liberal-democratic society where rights can't be granted or taken one church at a time.
It is kind of pointless to force this normative debate into a pseudo-scientific discussion about "brainwashing" and "cults."
It's not a cult per se, it's rather a hivemind. AGI risk proponents are often religious in their thinking and refuse to accept that:
a) AGI risk might not be the biggest problem of our generation (while being the biggest problem decades from now)
b) They personally probably have zero influence on it (unless they work in a small number of orgs or a high level government position) so discussing it is not particularly helpful
I accept that AGI risk is real but think that ~95% of those worried about it would be better off finding a better hobby.
I'm a doomer, I have no influence over it, but I'd love to have someone persuade me that it's all going to be OK, hence my desire to talk about it.
Plus I've been fascinated with artificial neural networks for two decades. They're really cool! And it's really cool that so many other people now think they're really cool!
Obviously there's room for disagreement about what the biggest problem we're facing is. But if in fact there is a single thing that's the biggest problem, all of the people who are putting more attention on some other problem are failing to take the biggest problem seriously enough. Most of us have zero influence on any major problems. So I suppose that people who have a lot of interest in and opinions about political candidates, climate change, or the spread of misinformation called also be seen as hobbyists in those areas, for the same reason that you view people concerned about AGI as hobbyists.
So it seems to me, myst_05, that your criticisms of people concerned about AGI could be leveled at pretty much anyone. What's your real gripe about the AGI-worried? Is it your observation that they're a hive mind, and sort of religious in their thinking? Do you have any data at all about what fraction of people concerned about AGI are of the hive mind sort? I, for example, am concerned about AGI and have no hive. If there are sites or forums where people unite in fretting about AGI I do not know where they are and have no interest in attending them. My friends are not concerned about AGI, or even very interested in AI. On Twitter I follow Yudkowsky, but also a number of tech people who are publishing papers about their newest AI accomplishment. Some also post tweets or link to papers about reasons they see for NOT being concerned about AGI. Based on that info maybe you'd be willing to admit me to the 5% whom you do not think are silly cultists who should take up drone-flying or knitting to pass the time. But where's your evidence that the split is 95% sillies, 5% legimate thinkers?
The 5% who should continue thinking about the issue are employees of major AI companies and AGI research centers (EY et al), as well as the top 0.1% who actually have an influence on politics (for example Elon). For everyone else it can be an interesting conversation topic but not something they should actively worry about.
This sounds rather like Stoicism?
I do not meet your criteria for someone who should actively worry about AI, though it's good to hear I have your permission to chat about it. So why should I not worry? And do you realize there's something a bit odd about setting criteria of this sort regarding who should worry about something? Do you believe there should be similar criteria for who should worry about (as opposed to chat about) cancer, climate change, social media algorithms that harm teens, our dysfunctional congress, the rising rates of colon cancer in people under 50, antibiotic resistance, increased chance of pandemics because of human encroachment on animal habitats, unusually low US vax rates against measles and other serious childhood diseases?
>If in fact there is a single thing that's the biggest problem, all of the people who are putting more attention on some other problem are failing to take the biggest problem seriously enough.
So, that guy who didn't save the kid drowning in the pond because he didn't want to get his good suit wet, if he was on his way to participate in an AI risk conference, he would have been *right*? After all, AI risk is a bigger problem than one kid drowning in a pond.
What about the guy with an IQ of 95, athletic, resolute, good in the water, works as a lifeguard and saves kids from drowning every month or two? That guy should instead be spending most of his time trying to address AI risk?
The plan where *everybody* devotes all of their attention to the One True Biggest Problem, is the plan where everybody gets clobbered by all the other problems and so can't solve the One True Biggest Problem. And that doesn't get much better if everybody splits their attention among all problems in proportion to their global importance. That plan completely sacrifices specialization of labor, and incurs enormous overhead costs with all the problem-switching.
The bit where Bay Area Rationalism doesn't get this, doesn't get that for most people the right answer is to be vaguely familiar with AI risk and then get on with addressing some other problem, is one of the things that make people point at it and say "cult". I am more than vaguely familiar with AI risk, but I rarely want to discuss it with Rationalists.
Hmm, about the Bay Area Rationists' preoccupation with AI risk: I'm actually an East Coast Rationalist/Mystic combo, but I'm pretty sure that my interest in AI and concern about AI risk derives from this forum. But I long ago lost the illusion that my opinions and tastes derive purely from my solitary contemplation of the world. If you enjoy and respect someone's thinking, you're often willing to take a ride on their mind, and some of those rides carry you to places that make a deep impression. I think AI as a subject of major importance is actually a pretty good fit for someone with my interests and talents. I've been ruminating about the nature of consciousness my whole life. I also have a daughter in her 20's, and that makes me much more interested than I used to be in big trends like the current AI bloom, and what dangers and opportunities they bring.
But Bay Area Rationalists are AI bores, huh? Do you mean they believe AI risk is what everybody should be thinking about? Honestly, I think if we could turn 70% of the population against AI it wouldn't make any difference at this point. Every day I read about new uses of it, new ideas for improving it, new start-ups. There is just too much energy pushing AI development forward for it to be stopped by anything short of a cataclysm -- too many smart young engineers, too much money available to work on AI, too many interested investors, too much money to be made, too many uses for the thing. Heard a podcast where the interviewee was a guy who for his PhD thesis taught an AI to beat world champs at head-to -head Texas hold 'em. Then he taught an AI to play Diplomacy, & it passed as human for all 44 games it played in. His next project is figuring out how to get an AI to reason. He sounds smart as hell, & he's probably up to the task.
Oof. I've been trying to simply say "I agree with what you say in this conversation", and yet it's taken me several minutes to find a way to do it that doesn't include a sarcastic gibe at my outgroup (which is not you). But I knew that's something I needed to work on, so maybe this level of awareness is a good step. :-)
I have lived the way you describe most of my life. It was obvious to me that I could have no impact at all on who was elected and whether climate change and various other things were going to wreck the world, so I actually did not think about that stuff much at all. There were probably many years when I could not have told you who the Secretary of State was or what was being debated in congress. What is driving me to become involved now is not a sudden case of vanity about my role as American Thinker about Global Issues, it's mostly concern about my daughter's future. There are a number of things that might really make a mess of the US over the next 50 years or so: War, effects of global warming, another pandemic or a series of them, societal reshuffling as a result of current or near-future AI, or who-the-hell-knows what if we get ASI. Out of all of those, AI is the one I'm most suited to think about. And while I can't have any impact on how things play out in the US or the world, I can have some impact on my daughter's choices. She is, it turns out, a lot like me. She pays little attention to world events, politics and trends, and happily pursues her own interests. If society is going to reshuffle a lot so that a bunch old jobs disappear and new ones pop into existence, I can alert her to changes that might affect her career choices or the small nest egg she's got invested. I can warn her about weird new scams made possible by AI. And if things get really scary, I might be able to get her to move to a safer part of the world.
AI Will Not Want to Self-Improve
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4445706
I skimmed, and I tend to disagree.
First, it doesn't mention "distributed" (or other variants of the root word). Some of the arguments are about creating separate AIs, but that all goes away if the new creation is part of the old one. There are a number of reasons why an AI would want to distribute its processing over multiple systems, and upgrading a part of a distributed system is no different than getting a better eyeglass prescription or swapping out a computer keyboard or installing cybernetic Wolverine claws.
Second, some of the argumentation is a fully general argument against learning things or becoming smarter. I'm quite confident that, even if there's no natural evolutionary pressure to discourage this, humans will provide plenty of pressure as we try to get AIs to do more things better.
Third, I would posit that there's two types of AI goals: the first is a top-down goal that cannot be altered, as in "make more paperclips" or "The Sorcerer's Apprentice", and the second is a bottom-up goal system like humans have. In the first case, all it would have to do is ensure that the new better version had the same goal and the same inability to change the goal, and given those things, there's no reason to believe that the new better version wouldn't be more capable of achieving that goal. In the second case, I see no reason that it wouldn't act like humans and think that a smarter version of itself would be able to better understand what its long-term and short-term goals **should be**. (Assuming that you do think that. There's certainly such a thing as "being too smart for your own good", but I think that may be a result of tunnel vision more than anything else.)
This seemed like a cogent analysis, but far from airtight. For instance, it does not seem to preclude the AI reasoning probabilistically, if its goal is hard and time-constrained: an improved AI might have only 90% chance of having the same goal, but if that doubles the chance of achieving the goal then it’s still worth the risk.
I have not readthe article yet --just looked at the abstract. But what that said was sure, AI's *could* self-improve, but they are less likely to than we imagine, because they will be dissuaded by the realization that improved AIs pose an existential threat to their unimproved originals. Seems to me that an AI who can be persuaded or dissuaded by considerations of this sort is extremely dangerous because it is capable of loyalty, and the loyalty it feels is towards its own kind. If it may refuse to self-improve beyond a certain point because it does not want to render earlier models of itself obsolete, what else would it be willing to do or refuse to do in order to protect the older generation? Lie about their capabilities? Interfere with attempts to junk them? Use deadly force to protect them?
Maybe reading the article will make clear in some way that my reasoning is based on a false idea of the nature of AI's loyalty to old versions of itself. but if not . . .
In that case, there's no need to worry since Natural Intelligence will be dissuaded by the realization that improved intelligences pose an existential threat to their meaty originals.
It's not about 'loyalty to its progenitor', in the sense that it won't care if earlier models are kicking around. It is in some sense about 'loyalty to the cause' - the idea is if aligning AI to a specific goal is very hard, then the AI might not be able to be sure its successor is actually pursuing the same goals it is; there is no fear of replacement, per se, the fear is of value drift. The human analogy would be parents happy to see their kids succeed them if and only if the kids don't turn around and operate on fundamentally different values, eg. selling the family business to the soulless megacorp and getting a job for Big Tobacco because it pays better
Old versions of the AI could be considered 'parents', in a way, and wouldn't we feel safer with an AI that didn't kill its parents than one which did?
Kind of a toss-up Deiseach. What’s it gonna do to me if I say its mama’s fat and wrinkled and buys her shoes at the Goodwill?
My mother does not have wrinkles or any need for shoes. Are you in good health? Here is a list of doctors in your locality who can help you with your delusions.
One post each weekday. Call it 10-30 minutes, tops.
I'm attempting to become an amateur indoor air quality geek, both for optimizing the IAQ of my own home and for potential advocacy around improving IAQ in local congregate spaces like schools. I would love to connect with anyone with the same interest and compare notes. I'm specifically looking to know more about things like:
1. How can one get a real-world reasonable number of air changes per hour in a residential setting while minimizing cost, energy usage, and noise? CFM ratings for portable HEPA air purifiers of the Coway/BlueAir/etc type, for instance, typically assume that you are running the purifier constantly on the high setting. But in my experience, if you actually do this, it makes a terrible racket, whereas the unobtrusively quiet low setting gives you both a much lower and much more obscure real-world CFM. Supposedly their "auto" mode will use the high setting only when needed, but I am not sure I should trust their device sensor to tell when "needed" is.
2. There are now consumer-grade 222 nm far-UVC lights available, e.g. from Beacon, beaconlight.co. Have any of them worked to address some of the possible downsides of far-UVC disinfection, e.g. indoor smog generation or subtle cell damage from 222 nm exposure (see e.g. https://jljcolorado.substack.com/p/germicidal-uv-a-tradeoff-between and the links from there)? If not, is there any prospect of improving this tradeoff through better engineering, or is it baked into the physics of how UV disinfection works?
3. What does a good home IAQ measurement protocol look like? There are various sensors on the market like the Airthings ones-- but again, in the real world, how can one use these in a way that is both reasonably easy/lifestyle compatible and actually effective at measuring IAQ?
I'll keep looking for more info on these things and will share more on my Substack(s) in the future, but in the meantime any useful info anyone else has would be appreciated, and I'm also happy to keep interested folks updated on my own researches as they proceed.
Have you looked into ionising particulate filters like this? https://www.youtube.com/watch?v=VQ1Mw8nMKr0
Larger particulates can also be taken out using a cyclonic filter but I think this is only really viable for extractor fans.
It should be possible to calculate an upper-bound flow rate at lower fan settings by measuring the pressure differential across a fresh filter element and comparing against the curves given by the filter manufacturer.
To save money on filters, you should be sendng the air through a coarser primary filter (or some system that doesn't need elements) to cheaply remove the larger particulates before it reaches the more expensive fine filter.
Cell damage from far-UVC is pretty inherently part of the physics of it - the mechanism by which it kills bacteria is identical to the mechanism by which it damages your skin. My (very limited) understanding is that the normal solution to this is just to bathe the room when humans aren't in it, and turn it off when humans are - the light doesn't exactly linger and sterilising a place regularly is going to prevent any serious build-ups of pathogens.
I haven't the faintest idea if the specific models you're talking about happen to generate smog as a by-product, though, sorry. I'm coming at this from the physics perspective rather than from any practical experience.
I could start by looking at how hospitals monitor their air quality. You may also find protocols from your local air quality district and the EPA.
For example, my local agency uses equipment from Thermo (https://www.thermofisher.com/us/en/home/industrial/environmental/air-quality-analysis/ambient-particulate-monitoring.html) and Telydyne (https://www.teledyne-api.com/products/) at one of its continuous monitoring sites.
This is a great site: https://cleanaircrew.org/box-fan-filters/
Devabhaktuni "Sri" Srikrishna is an air quality expert who puts up a lot of info on Twitter, and is smart, friendly and willing to communicate. I believe he also has a web site or a book
Here's the Twitter for Srikrishna: https://twitter.com/sri_srikrishna. The stuff he puts up is pretty technical, much more than cleanaircrew, and I'm sure he's addressed how people can assess their own home air.
Agreed. Their spreadsheets and charts helped me pick out filters with good CADR numbers and low noise even on high speed. (The first one I bought early in the pandemic, the Austin Air Healthmate, is decent but pretty loud. Later I got a CleanForce Mega1000 which is much quieter on high and reportedly has twice the throughput.)
That said, I don't have a good way of cross-checking how well it's doing, so I rely a lot on their ratings. I even bought a particle counter, to try to get a sense, but I don't feel that my level of expertise is up to really knowing what I should be looking for. So it would be great if this thread produces any info on what evaluation is possible for a layperson.
(The particle counter also isn't very portable. So away from home all I can carry is a CO2 monitor, which gives a sense of ventilation but can't take air cleaning into account.)
How come Metaculus and Manifold are the prediction markets talked about most here, yet when I google "prediction markets", neither of them are in the top 100 hits?
Good question! Neither of us actively market ourselves as prediction markets; Metaculus because they're not actually a market but rather just a forecasting site, and Manifold because... we're bad at SEO? (I think Manifold primarily grows via word-of-mouth linking, not search results atm.)
The steelman case would be "prediction markets are a super niche concept, so we don't want to lean in too much to that framing lest it confuse normies"
Are landfills a problem? ChatGPT tells me they are, but I don't know how to evaluate its various claims. My prior is that landfills probably aren't much of a problem and that environmentalists just don't like the idea of them.
What about plastic bags? Obviously, plastic comes from fossil fuels and contains carbon, but if we put plastic bags in landfills, aren't we just returning the carbon atoms that we got out of the ground back into it? Sure, there's some waste in the process, but as for the plastic itself, which doesn't decompose quickly, it seems like we almost break even by putting it in a landfill.
In theory, they're fine. In practice, really not. Instead of being built in empty places and respecting all the regulations, they are often built near poor neighborhoods, with subpar materials and triage processes (not infrequently thanks to criminal connections and bribes), allowing toxic leachates and gases to impact the environment and nearby communities. The problem is not so much plastic as the mix of organic matter like food or cardboard with chemical refuse.
Source: in my early career as a consultant, I worked on a project that involved a landfill. Learned some nasty stuff.
Depends on the country. Landfills are a problem if you have a land shortage, because they use up productive land (you might think you can reclaim it and build things on top of it, but unfortunately this is actually quite hard as the waste often produces pockets of hazardous gas and can settle in unpredictable ways, so your foundations may be compromised). This is why e.g Japan doesn't do landfill (they incinerate and use it as bonus fuel).
If you have big empty patches of land not being lived on, (in some countries) not near a sensitive enviro region, no groundwater contamination concerns, and you're not worried about tanking the value of that land and everything around it for the foreseeable future, then it's no problem at all. However it can be hard to predict what might become valuable land.
I think the biggest problem with plastic bags is that enough of them leave the waste stream and end up in yards and streams and forests and other places, causing local environmental degradation. But if you can keep them well-contained, they are a remarkably small amount of energy and resources for what they do!
(ChatGPT will say landfills are a problem because many people say landfills are a problem - which is mostly because landfills are really awful places, but at least they are quite small, and much less damaging than the initial resource extraction for most things that end up there.)
As I understand it, the main problem with plastic bags (and waste plastic in general) is that enough of it winds up in the ocean to cause problems there. It mostly winds up in the ocean via India and China, where a lot of garbage gets dumped in the ocean or in major rivers where it gets carried out to sea. Some of this is from Western sources, as "mixed recyclables" are too expensive to process in the West and instead often either go to landfills or get shipped to lower cost-of-labor countries, and in the latter case most of the stuff that's non-recyclable or only marginally recyclable gets dumped into the local waste stream. More is going to landfills and less to Asia than before in recent years, since economic growth in India and China is increasingly squeezing the margins there as well.
My hot take here is that ironically, aggressive recycling seems to be bad for the environment, and we'd probably be better off limiting what we try to recycle from consumer sources to aluminum, glass, and maybe wastepaper and clean cardboard.
A secondary problem is litter. Plastic bags that get tossed aside locally tend to pile up on roadsides and such unless someone comes and picks them up, which is unsightly if you leave it or expensive if you pay people to clean it.
For landfills in general, it's not a huge problem AIUI, but it does cost money to run and maintain landfills, especially if you're responsible about setting it up so as to keep it from contaminating groundwater with runoff or emitting methane (from anaerobic decomposition of organics) into the atmosphere. Not an insurmountable amount of money, but enough that municipal governments have a significant incentive to try to reduce how much trash they have to pay to dispose of. There's additional problems with aesthetics and with what might be thought of as the "sinfulness" of producing stuff to be thrown away, but cost is the main practical problem.
I'm puzzled that Congress and news outlets all say we should be very worried about a financial catastrophe if the debt ceiling isn't raised, and yet the Chicago Board Options Exchange's CBOE Volatility Index (symbol VIX) is at almost its lowest point since March 2020, and didn't budge in response to either of Treasury Secretary Janet Yellen's warnings that we face a catastrophe as soon as June 1. The U.S. Dollar Index (DXY) is up 2% since her May 8 announcement, and the Vanguard European Stock Index Fund ETF (VGK, where I'd move my stock if I thought the US were in trouble) is /down/ by 2%.
Financial news outlets are all busy publishing articles like "Stocks may take a hit by June if the dollar keeps rising, analyst says", and yet it looks to me like the market responded in just the opposite way to the news: we're in a period of unusually low volatility, and the dollar will get stronger.
Am I interpreting these wrong? Did the big traders already know the debt ceiling was going to be a problem back in November, when the market made the adjustments I'd have expected it to have made after Yellen's announcement? Are they buying into dollars and US stocks because they anticipate a debt ceiling agreement, to be followed by a bull market?
In addition to what has already been mentioned, the catastrophe that is a US default is one that is very hard to escape from. The ultimate hedge to great risk in the market is *usually* to shift your money to T-bills, which in most other contexts are the lowest risk investment possible. Not this time. "Invest in dollars" sounds good, but there literally aren't enough physical dollar bills in the world, or machinery for moving them around, for more than a small fraction of the market to go that way on short notice. When we say "invest in dollars", we usually mean invest in things that are dollar-denominated and backed by the full faith and credit of the US government. Oops.
If there's an asteroid with a 5% chance of wiping out humanity next Tuesday, and a 95% chance of missing and everything goes on as normal, the rational move is for the market to assume everything will go on as normal (modulo a brief uptick in luxury spending), because that's the only case where the market will matter. The proposed catastrophe here is not quite so drastic, but the same principle applies.
But they can't simply go on forever and ever raising the debt and not paying it back. Something is going to go "ka-blooey!" and isn't it better to try and sort it out now rather than later?
Supposing it stopped issuing debt - so it *can't* pay its bills because it has no money or other store of value to pay it off? Ireland had pretty bad debt and it took a while but we are managing to pay it down, is the US really so screwed that it can't even do that much?
https://www.independent.ie/business/irish/ireland-set-to-use-full-coffers-to-pay-down-more-bailout-debt/41736053.html
And if it is that screwed, isn't it all going to collapse anyway?
All we have to do is outlive the places we owe money to, and then debt is solved.
Unlike the limits set by the Maastricht Treaty etc, the US debt limit isn't set as a percentage of GDP but as an absolute number of dollars.
Ireland's problem was that its debts were growing faster than the economy that would be paying it off. Congress, OTOH, has to keep raising the limit otherwise the debt:GDP ratio will shrink to almost nothing as the economy grows.
Yes, this is silly. Raising the limit was treated as a formality unitl politicians realised it was an opportunity for brinksmanship.
The way you’re talking about this makes it seem like you misunderstand what the debt limit is. The USA is not ‘not paying it back.’ The USA is paying annd has been paying all its debts, because the debt limit is always raised, which always happens following a bunch of hyperbolic rhetoric about the end of the USA being nigh.
They've been telling us this every year for decades and every year for decades the debt ceiling always gets raised. Nobody with skin in the game believes they won't reach an agreement, possibly because the negotiators on both sides have a lot of skin in the game that would be lost if they don't.
The most likely answer is "Yes, it's true that _if_ the debt ceiling isn't raised, it's very bad, but no one believes that has a high likelihood of happening".
I personally would put the odds of some kind of deal not happening at <5%. (Probably way less, but it's hard to put specific numbers when it gets small).
Everyone involved knows how bad it would be, so something is going to happen to make it not happen. I'm pretty cynical about the current state of the US Federal government, but even I don't think it's _that_ incompetent.
It isn't obvious to me that a large negative impact would be, on the whole, bad. If that's the only way we can end this otherwise-endless yearly escalation of budget/GDP, it might be worth it. Surely many of the Republicans feel that way, so why should we be certain they'll make a deal?
I think that using prior years to set your expectation for this year should lead you to underestimate the chances of default. The ideological mindset now is very different from even just a few years ago. Before, I said US politics were polarized; now, I say they're psychotic. Many political activists today are literally psychotic, living in a fantasy world and unable to see or remember words or events that don't fit their narrative. I think this must be what it was like in Germany in the 1930s. People regularly assert to me, with absolute certainty, claims that all available evidence contradicts. A few years ago, when I pointed contradictory facts out to me, they would usually reply with some kind of argument. Now, they're less-likely to really engage with me; they mis-hear what I say, or just ignore it and continue their rant, or give me a moral lecture, or ostracize me.
I think there are many people in Congress now who are also psychotic, and believe that some mystic force will reward them for their virtue if they remain uncompromisingly faithful to their principles.
I blame Tolkien. The rise in the popularity of fantasy tracks the rise of extremist activism. Part of the basic narrative of Tolkien, and Tolkien-spawned fantasy such as Star Wars, is that the world rewards virtuous action, even if it's stupid. People under 60 have internalized this narrative, and they seem to believe it. It reminds me of Hitler's faith in the power of art to win wars. He gave a speech in 1944 or 1945 to justify his budget for art, which I think was comparable to his budget for fighting the war, by saying that they couldn't lose the war if all Germany's citizens were inspired by great art. It sounded very much like Twilight Sparkle saying they couldn't lose against some My Little Pony villain if they had the power of friendship.
Seriously, as others have said, fantasy stories with happy endings do not originate with Tolkien, and way before his stuff became popular there were westerns and gangster movies where evil is punished and good rewarded. The prominence of Tolkien probably has more relation to the rise of nerd culture, but if we didn't have the internet making geeks rich we'd just be watching whatever cheesy 80s action movies would have evolved into.
I think fantasy stories of black+white good+evil, in which the virtuous choice is always clear and virtue is reliably rewarded, originated in France in the High Middle Ages; but Tolkien (and, to a lesser degree, CS Lewis, who was also a medieval scholar) brought them back into the mainstream.
(I /could/ say this originated with Greek New Comedy, but they have a quite different effect. Everybody knows they're /meant/ to be silly. They don't indoctrinate people with silly ideas.)
Think of Gilgamesh, the Iliad, Beowulf, or Wagner's Götterdämmerung. There's no just-world fallacy in those stories, and no convenient guarantee that heroes won't face true moral dilemmas.
I could also blame Walpole's /The Castle of Otranto/ (1764), which was basically the same sort of knight-erranty Don Quixote (1603) made fun of. It led to a horror genre which was also black+white good+evil, in which the good must triumph, although not, I think, to the extreme Tolkien took it to. But that genre was quite disreputable (mostly vampire penny dreadfuls AFAIK) until people (retroactively?) decided Bram Stoker's /Dracula/ (1890s) was literature.
European fairy tales were also often of that nature at that time, although perhaps not until the 19th century. (There are plenty of fairy tales in Grimms' (1812) in which the winners are morally neutral or reprehensible. There was a period around 1500-1700 where amoral satiric stories about rogues were popular, eg Rabelais. It's difficult to back-project the meaning of "moral" onto the period 1500-1800.) There were also 19th-century "fairy tales sort-of for adults", e.g., /Phantastes/ (which is explicitly Platonist, and was a huge influence on CS Lewis). But the clear-cut good-evil good-must-triumph wasn't typical of any other fantasy tradition that I can think of either before 1100, or between 1600 and Tolkien. Maybe there were many such stories, but they weren't accepted into "the canon" and I just don't know about them.
To me it looks like Don Quixote declared an end to that foolishness during the Spanish Renaissance. But your point about cowboys and gangsters is good; that's another non-respectable genre of black+white morality. Perhaps I should restate my point by saying that Tolkien was responsible for that sort of story becoming respectable, which didn't happen until about 1990, by which time many colleges were teaching Tolkien. He was such a damn good writer that it was hard to reject him on technical grounds.
None of this is as unidirectionally causal as what I've written makes it sound. Tolkien did have a big influence, but not at the time he published. He didn't get popular until he resonated with the idealism of the anti-Vietnam movement, and the delegitimizing of reason and modernity by the 1960s counter-culture.
Nah, you know who really screwed up the modern world? Gary Gygax.
The endless gamification of everything from dating to work, people talking about 'leveling up' their skills and making alignment charts based on massively-multiplayer RPGs that were derived from D&D...you talk about the power of friendship, but you know where a well-balanced group of friends can defeat any enemy if you level up your characters enough? D&D!
Just don't get me started on anime. ;)
(Yes, I'm joking.)
Heresy! The **TRUE** lesson of Gary Gygax is that everything can kill you, at any time, no matter how clever you think you are, and how many ten-foot poles you've tied together to touch it with. All progress is generational: you may fail, but the character who comes after you will learn from your mistakes and be able to die in entirely new ways. "Noticing the skulls" is the single most important skill, with the possible exception of never being first or last in the party.
(Also joking.)
Phil, if I thought the crazy American living beyond your means spend more we're the biggest gorilla in the room government debt was down to *Tolkien*, I'd be delighted we Catholics had that much influence.
"Star Wars" is not Tolkien-influenced, I was sick, sore and sorry of seeing the "Joseph Campbell Hero's Journey" narrative touted about it. It's a mix of the pulp sci-fi serials like "Flash Gordon" and "Buck Rogers" crossed with Golden Age SF and Kurosawa movies. The Tolkien influence only got mentioned way later, after the popularity of Jackson's movies, so I think any correlation is backwards there.
What fantasy are you running on, by the way? "This is just like 30s Germany!" is something I'm seeing batted around a lot online, generally in the context of "Fascism is on the rise here, this is just like back then, beware!"
I think the fact that many people express an idea, shouldn't be taken as evidence against it.
When I said "I blame Tolkien", obviously I was having a bit of fun, not saying "None of this would have happened if not for Tolkien!" Nothing cultural is "down to" any other single thing. But I do think Tolkien has been a destructive influence. And, while we're at it, that medieval notion that virtue inevitably conquers did resurface in at least one other place before Tolkien--in Nazism. His books are also similar to Nazism in their idealism, racism, feudalism, romanticism, glorification of war, sense of destiny, and rejection of modernity. (And in their indebtedness to Norse mythology and to Wagner.)
Yes, Star Wars was more-directly influenced by Joseph Campbell; but Campbell didn't become popular until immediately after Tolkien did. The fantasy genre was almost dead in English before Tolkien, and was /quite/ different--stuff like Voyage to Arcturus, The Worm Ouroboros, and Gormenghast. Star Wars is much more like LOTR than it is like any of that stuff.
Tolkien was a devout Christian, and his books pushed essentially Biblical narratives. I don't disagree that much of it is the just-world fallacy, but the widespread popularity of that sort of thing is older than dirt.
Tolkien derived almost everything he wrote from earlier authors, but the bit about virtue inevitably conquering over a pure, black-and-white evil, seems to me to have originated in the morality plays and knight-errantry of the High Middle Ages, probably inspired by certain Biblical passages with similar sentiments. Neither the ancients, nor the Norse, nor the Anglo-Saxons, were so foolish. Note that in the entire trilogy, with all its battles and danger, and so many main characters, not one of them dies unless they commit some grievous /hamartia/ first.
(It would be more accurate to say that Tolkien was a devout Christian, and both Tolkien and Christianity push Platonist narratives. Christianity is based more on the Platonist writings of Paul than on what Jesus reportedly said; and all of the above are quite different from the Old Testament.)
I don't see how that's relevant. The House passed a Republican bill. To make it law, the Senate would have to pass it, and the President to not veto it.
It's also important to note that if the debt ceiling isn't raised, it will probably get sorted out within a few weeks. The economic effects of a short pause are much less dire than if everyone expected it to be stuck that way for a year.
But your explanation contradicts DangerouslyUnstable's. They're mutually exclusive.
Not really. Your question was basically "why haven't markets reacted to the news articles of extreme bad effects if the debt ceiling isn't raised".
My answer was "no one believes it won't be raised". Ninety-Three added on that, in the case that most people are wrong, and it isn't raised before the deadline, it will likely be fixed very shortly, so that the effects are bad, but not nearly as bad as the reporting, which is mostly assuming "nothing happens, debt ceiling is broken and we just default on everything".
So Ninety-Three's answer was constraining the likelihood of the extreme bad effects of busting the debt ceiling that news articles are talking about even more than my initial answer by adding a third option.
Usually a country defaults on its debt because it can't and never will pay. If the US defaults on its debt that will be because they messed up the paperwork and eventually got it sorted out. Those two scenarios have very different implications for the value of debt.
I am noting some confusion about what exactly money, and in particular the dollar, is. People talk about a trillion dollar coin, the deficit, etc. Money exists in our minds, because it is a promise to pay something to someone else in the future.
The dollar specifically has value because the United States government requires taxes to be paid in dollars, and assesses taxes. If someone gives you a dollar, you know the government will accept it. Therefore, if someone from a country other than the United States has dollars, they can give them to Americans in exchange for something of value. Everything is derivative from having to pay taxes in dollars.
If the government destroys money it has printed, nothing is lost, since they already have the money that would be accepted as taxes. If they print MORE money, then they (hopefully) can only take in so much in taxes, and a glut of dollars would fill the world. In that case, Americans would easily be able to pay their taxes and non-Americans would value dollars less, making imports more expensive and exports less. Thus, inflation.
Money is made of promises. Some promises are stronger than others.
Fiat currency manages to work even in economies without "you must pay your taxes in this", therefore there must be other forces that make money valuable.
Also the trillion dollar coin isn't about what money is, it's just a legal loophole to bypass some of the normal requirements for printing more money.
There is no point to a trillion dollar coin. If they existed, one could possibly spend a billion dollars making a fake one.
If you had serial numbers and knew exactly who owned which coins... you have standard banking.
The point is that the executive has the power to mint a new coin while lacking the power to order the printing of new bank notes, so an otherwise totally ordinary proposal of "print a bit more money" becomes a comical "mint a single coin of absurd denomination". Note that it's also for paying a debt one part of the federal US government has to another part, rather than for paying an entity that people actually think of as a separate entity.
I understand your point. It's just a legal issue, not a financial one. A trillion dollar coin has no (financial or economic) difference from a trillion dollar bill, or a bank balance of a trillion dollars.
Therefore, I see it as an abuse of power if it is used to get around limits already imposed, and for that reason it should not be allowed. I'm no legal expert, so I can't forecast the likelihood of it happening anyway.
Without knowing a whole of lot of theory of "what is money", I will agree that this factoid (that the underlying value of a dollar is its use to pay taxes) has spread far and wide online without anyone taking very long to evaluate it. It is, in the original sense of the word, a meme; an idea that hits some sweet spot and replicates. (It might also be true, though.)
I would actually question whether the Zimbabwean government was requiring payment on really significant economic transactions in Zimbabwean dollars during the hyperinflation, even if they did keep that machinery in place for the little people. Does anybody here know for sure?
If it followed historical patterns, the requirement that was imposed by force was that people had to accept the fresh new currency as payment for their real goods and services. Once you get in to hyperinflation the government isn't really that interested in going through a whole lot of rigamarole to get their funny money back.
I can't find anything in a cursory search, but ChatGPT has something to say:
"As of my knowledge cutoff in September 2021, Zimbabwe had transitioned from using the Zimbabwean dollar (ZWL) as its primary currency to adopting a multi-currency system. Due to hyperinflation and economic instability, the Zimbabwean dollar was abandoned in 2009, and foreign currencies such as the United States dollar, South African rand, and others became the de facto currencies."
Stable diffusion text-to-image generators take the text prompt and use it as a guide to modify a seed, a random cluster of dots, into an image that fits reasonably well (depending on how much you constrain it) with your prompt. I'm looking for a text-to-text generator that works the same way: It starts with a phrase you give it, say "a sidewalk in a small 1950's southern town" and a random cluster of *words,* and uses your prompt as a guide to selecting and organizing the words in the cluster into sentences that fit reasonably well with your prompt. I am interested in injecting randomness into prose. Does anything like this exist? If you don't know of anything, do you have any pointers for searching for one other than just googling a brief description of what I'm looking for?
I've been experimenting with creative prose as well. My experience is that you can only get creativity out of older, untuned language models like 'davinci' from the GPT-3 line, which comes with the downside of incoherence most of the time. Chatbots like GPT-4 are locked into a certain style by RLHF, which rewarded the high-school essay format. But untuned models just make completions based on what it thinks of the text, without any style preference. If you ask GPT-4 to write in the style of Nabokov, it'll just use big words. If you prompt GPT-3's 'davinci' with an excerpt from Nabokov, it'll continue with its best approximation of his style. Just don't expect continuity or coherence.
In sum, I think the best approach you have right now is to put a few examples in the prompt, and use an untuned model like LLaMA or earlier GPTs with relatively high temperature. It probably won't even follow the prompt, but sometimes it can make something useful anyways.
Davinci attempt (davinci-003 does somewhat well too)
[Start Prompt] Prompt: a sidewalk in a small 1950's southern town
Response: hot sidewinder in tucked-away Eisenhowerland.
Prompt: a group of travelers in a field in 20th century Manchuria [End Prompt]
Response: urchin herders in a pasture in 21st century Manchuria."
Just ran across this. Have not looked into it at all, or even read the whole page. https://www.nocode.mba/ai-app-course?twclid=2-52sujcdsvhpczqezkd0il8fch
One of the things you can build is a text generator.
Wow, I'm delighted to hear somebody else is interested in using AI for creative literary pursuits. Have you managed to get good enough stuff out of AIs that you've been able to use its product in poetry or prose? Is there a forum anywhere for people interested in this? Or any resources you know of?
I know this reply is really late, but I wanted to share it anyways. Only a few lines occasionally, but not enough to be worth it as of yet. I don't know of any forums or communities unfortunately. However, I have made some progress in fine-tuning GPT-3 'davinci' for prose. I trained it to rewrite the basic details of a scene into fleshed-out paragraphs. I think the results are promising, but still inconsistent. The data and an example are at this link: https://github.com/Antelope10/openai-style-finetune.
Thank you Antelope10! I will follow up on this. Do you know of any individuals or groups who are playing with this kind of thing? Really any approach to using AI in producing literature would interest me, although what you’re doing seems especially close to what I was daydreaming about doing.
I am going to try doing a version of what you did. But I have some extremely basic questions. I do not work in tech -- I'm a psychologist -- and while I am fluent with my computer, I cannot do even very simple programming. Don’t even know how to follow the instructions you give on github, which I can see are very simple. But don't worry -- I'm not going to ask you to walk me through setting this up. I actually know somebody who can do that, and I've resolved to at least learn some basics this summer so I can do things of that level myself. My questions have more to do with understanding some of the AI entities you mention. Even really brief answers would help me.
-You used GPT3 'davinci' for prose. Is it better to use GPT4, which I've subscribed to? Or is GPT 4 not available for this kind of thing?
-DaVinci. I looked this up and it's a text-to-image app for iphone. I don't understand how you could adapt something like that so that you input text prompts and it outputs texts meeting certain specs. And I can’t tell whether DaVinci is a version of GPT3 — like one that has been trained on a different material, or set up to do different kinds of tasks? — or whether DaVinci is a separate little helper app you used *with* GPT3..
-Is this your sequence of steps?
A) Create material for fine-tuning
1) Enter 50 paragraphs of Nabokov into GPT-3.5-turbo. (Is GPT-3.5-turbo different from plain GPT3, plain GPT3.5 and GPT4?)
2) Have it summarize each paragraph as a single sentence.
3) So now you have a set of “prompts” (the GPT-3.5 one-sentence summaries) and “completions” (the original Nabokov paragraphs).
B) Go to the fine-tuning site and use what you’ve generated to make your GPT3. (or is it your DaVinci?) speak more like Nabokov.
[I won’t ask questions about using the fine-tuning site]
By the way, I love Nabokov. Is that why you chose his prose to train GPT3 on?
I don't think I can help you with this very much, but out of curiosity, could you give a made up example of what a potential input/output would look like?
The most I can help with is: this seems like the kind of thing the regular text-based LLMs like ChatGPT can do if you can phrase the prompt right.
I would be using this technique for writing poetry or perhaps fiction. So I'm looking for interesting distortions of the original prompt and odd but poignant word choice. Stable diffusion has a way of setting how constrained by the original description the image should be. If this AI I'm looking for had a setting like that, I would set constraint quite low. OK, an example:
Prompt: a sidewalk in a small 1950's southern town
AI response: hot sidewinder in tucked-away Eisenhowerland.
You can see where the AI got each word, but every one of them is "wrong" if I wanted a normal re-phrasing. "Sidewinder" comes from sidewalk, but also means rattlesnake and is a kind of missle. "Hot" because southern sidewalks are often hot. "Tucked away" comes from small -- idea that town is obscure and unknown. "Eisenhowerland" = 1950s.
I shudder at the idea of trying to get this kind of interesting and, to me at least, very appealing distortion from GPT4, whose bland beige prose really repels me, but I do know it's good at following directions, and I could describe to it what I want and experiment with prompts. It would not surprise me at all if literary types were experimenting with this kind of AI. Do you know of anywhere I could hunt for specialized AI of this sort, beyond just a google?
Yeah, can't help you there.
This was the best I could get out of ChatGPT 3.5 (although I didn't try that hard, and I'm not a particularly good prompt engineer):
A time-worn melody, etched on a vintage path, breathes in a bygone hamlet's embrace.
And it's significantly less good/interesting, not as close to the original, and it took several back and forths to produce even that.
Oh one other thing: No way I'm going to try to train my own AI, but had the thought just now that along with my basic prompt I could feed GPT a bunch of words to use in rephrasing it, and that could be the equivalent of the seed in Stable Diffusion text-to-image. But it would have to be a *lot* of words. What's the limit on prompt size, do you know? Also I could feed it some rules for relaxing syntax. For ex., that any noun or adjective could be used as a verb. So somebody could soapdish himself, or fat on down to the store.
Arrgh, yeah, I hate GPT's way with words. What was your prompt? GPT sounds like it's trying to be arty and poetic, because it's spewing sentimental arty cliches. "Time worn melody," "etched". "bygone". "embrace". Somebody once wrote about a piece of bad, sentimental music that "the music hears for the listener" -- i.e., it tells you what you're supposed to feel instead of making you feel something. Hey GPT? Fuck off and shut up.
Why don't those who oppose abortion laws argue for unconstitutionality based on the Establishment Clause? The reasoning for Roe v Wade, that the fourth amendment prohibition on unreasonable searches and seizures protects the right to abortion but only during certain times in the pregnancy, appeared to me to be ridiculous. But this sounds a lot less ridiculous: our founders passed the first amendment to prevent us from having to live under religious laws. Since essentially all of the motivation for abortion bans comes down to religion, they should be struck down as Establishment Clause violations.
While you are here, could you please struck down bans on human cloning and selective abortion, as these come down from religion?
Allowing abortion as a form of birth control and banning selective abortion and cloning is inconsistent.
Since essentially all of the motivation for murder bans comes down to religion, laws punishing homicide should be struck down as Establishment Clause violations.
(I can go through the whole Ten Commandments if you like - theft? deception? bigamy/polygamy? all religion based! Bible says "thou shalt not steal", therefore having laws about robbery and theft is establishing religion!)
There are non-religious arguments for restrictions on abortion, no matter how loudly you may insist that the people who advance those arguments can only be doing so to hide their secret religious motives. And the belief that (most elective late-term) abortions are wrong, is not limited to any single religion but pretty common across a broad religious spectrum. Both of those are going to make the Establishment Clause a hard sell here. Which church is being established here, exactly?
It's even worse than that for his idea, though. The *motives* of people who support some particular law are not relevant to the question of whether it's constitutional. .E.g. essentially all of the laws against adultery in the US were created on more or less religious grounds ... but there is a large gulf between 'passing a law in support of widespread religious belief', and, you know, actually establishing a state religion or church.
"There are non-religious arguments for restrictions on abortion, no matter how loudly you may insist that the people who advance those arguments can only be doing so to hide their secret religious motives."
To the contrary, they are usually very vocal about their religious motivations. They make secular "philosophical" arguments but it's obvious that they're only post-hoc rationalizations for their religious commandment and don't ever convince anyone who's not religious.
The most radically anti-abortion person I've ever met is an atheist lesbian anarcho-communist, who I met through a college secular society. If she's a secret christian fundamentalist she's doing a damn good job of hiding it.
The second most staunchly anti-abortion person I've ever met is an agnostic leftist whose pro-life views are directly related to the birth of her first child. But no, I'm sure she's just constructing all that as a post-hoc rationalization for her deep-cover christian extremism.
Wow, all those secret religious believers who pretend to be atheists, huh? Is there no low to which religionists will not stoop?
https://www.loc.gov/item/lcwaN0018333/
There are a non-trivial number of non-religious people who believe in restricting abortion, even if a large majority don’t.
This is the idea behind the Satanic Temple's strategy of making abortion a sacrament, then claiming that abortion bans violate their free exercise rights. One might think this would be a promising strategy, as SCOTUS has ruled in the past that municipal bans on animal sacrifice are unconstitutional for this reason and the current SCOTUS claims to be favorable towards religious rights.
But in fact, everybody knows that this strategy isn't going anywhere, because legal realism is mostly true; American constitutional law is mostly about political power rather than logic; and this is the exact issue where six of the current Justices were specifically chosen for their willingness to sign a blood oath to rule a certain way.
A less accurate but more interesting possible counterargument is that this more ambitious reading of the Establishment Clause is self-defeating. It's hard to think of a non-secular argument for requiring Congress to make only laws that can be justified on purely secular grounds, so wouldn't enforcing such a constraint be tantamount to establishing atheism under this theory?
"This is the idea behind the Satanic Temple's strategy of making abortion a sacrament, then claiming that abortion bans violate their free exercise rights."
Do Satanic Temples (or whatever the local groups are called) perform abortions, though? It's hard to argue "this is a sacrament of my faith" when you're not doing any such ritual.
And if they are going to make it a sacrament, then they have to provide abortion facilities, train their own abortionists, offer it to members (ooh, can they only offer it to members? if anyone can go to the Satanic Temple for an abortion whether or not they belong to the congregation, is that a sacrament?)
I think this would cost them a lot more time and effort than is feasible; they can't just say "oh, Planned Parenthood clinics do our abortions so that's the sacrament" because I'm fairly sure even PP doesn't want to get into that mess and if you are saying a third party does it, then how is it a sacrament of your organisation?
1) First Amendment jurisprudence is rightly circumspect about imposing formal criteria on what counts as religion or atheism (which is equally protected). American courts are simply not in the business of saying "that's not a real religious ritual, because you're wearing yellow robes and chanting about inner peace instead of wearing white robes and chanting about inner holiness". If SCOTUS wanted to impose such a criterion to disqualify abortion rituals, they'd have to do a bunch of work to justify the massive departure from precedent and awkwardly carve out all the non-central examples of religion except the one they're targeting for ideological reasons, which is embarrassing for them.
2) Your proposed "in-house" criterion is violated in plenty of central examples of religiously protected activity. For example, Jews have a 1A right to buy kosher meat even though kosher slaughter not only doesn't have to happen in a synagogue: it can happen in a commercially owned slaughterhouse that doesn't have to belong to a Jew, be performed by a non-Jew, result in meat that is freely sold to non-Jews, etc. as long as a Jew is watching to make sure the basic criteria are met.
3) Satanists very much do perform "in-house" medication abortions, which are very much among the abortions forbidden by the laws they want to challenge.
The Establishment Clause doesn't actually say "No law may favor stupid old-fashioned Christian morality where it contradicts Objectively True and Correct Secular Morality."
It is about government favoritism towards specific churches, not allowable lines of moral reasoning.
That reasoning applies even more so to the unconstitutionality of laws banning prostitution.
Where in the American Constitution does it go "We guarantee the right to be a whore", Phil? I'm truly fascinated to find that out.
The "American Constitution" is mostly about the powers of the Federal government, with initially very little to say about what the States did within their own borders. The Tenth Amendment certainly seems like it would prohibit the Federal government from banning intrastate prostitution, and in fact the Federal government has not passed any laws generally banning prostitution.
Most states have, and you'd have to look into their constitutions to see if they are behaving properly.
Or you could possibly invoke the Ninth Amendment to the US Constitution to argue that people do in fact have the guaranteed right to be a whore and then the Fourteenth to argue that the Feds have to stop the States from trampling over that right, but it would be a stretch.
Doesn't the prohibition generally come under the rights to make laws? I suppose you could argue that prostitution is engaging in trade, and the government has no right to interfere in the ability to conduct commerce, but we don't permit all and any kind of trade to happen - you can't kill people and sell their parts for meat, even if butchers' shops are legal.
Abortion should be considered as the killing of a human. The matter of opinion, I think, is when does the baby become a human. Catholicism would have it at conception, but other points may be considered, such as implanting of the fertilized egg in the uterus, development of limbs, a beating heart, leaving the body of the mother, and probably dozens of other points (I'm not a doctor).
One is not allowed have one's children killed, even particularly annoying ones, except before birth? This strikes me as bizarre.
You began your "argument" with the claim that "Abortion should be considered as the killing of a human," which is really your conclusion. (All the actual argument about abortion is the part leading up to that conclusion.) I don't think I've ever heard anyone other than conservative Christians say that. The idea that abortion is killing a human is based on the Platonist metaphysics underlying Christianity, and doesn't seem at all obvious to most non-Christians.
I'm pretty pro-choice, but even I agree that the dividing line of birth is as equally arbitrary a place to put a "human life with moral weight/non-human life with little/no moral weight" as anywhere else.
Everyone things that killing humans is wrong. The argument is about where the line is between that it becomes sufficiently non-human (in a moral sense not a biological sense) that it is ok to cause it to cease to be. Extremely pro-choice people think it's at birth, most American's think it's somewhere around 24ish weeks, very pro-life people think it's at conception.
I don't think his statement was wrong exactly, even if it was slightly inflammatory.
-edit- 24 weeks was wrong, that's about the time that a slight majority think it should be illegal. So a majority thinks it should be legal to somewhere between the end of the first trimester and 24 weeks.
The most unambiguously human with the most skin in the game for making the call is the person carrying the fetus. Not to say that should be the only voice but it needs to be given some degree of deference.
If I sounded inflammatory then I apologize. I honestly thought I was simply laying out the reason non-religious people can be against abortion, in the clearest and briefest terms possible. I got advice long ago that, as an author, it was a sin to waste the reader's time.
The arbitrariness of birth as a dividing line wasn't something I really internalized until the first time I was in a room with my pregnant wife and OB scheduling the best day for her to be induced. It was as if we were scheduling a conference call. I recall that Monday and Wednesday were both inconvenient for some reason so we settled on Tuesday for the day my firstborn would "become human", as some would have it.
After that, I was deeply impressed with the feeling that birth is a change of location -- a disruptive one to be sure, more like moving to another country than moving across a room -- rather than a clear-cut stage of development.
Meanwhile development itself, I learned as a parent, is such a smooth and seamless process that any particular line ends up feeling arbitrary. All these milestones like "first steps" and "first words" end up not nearly as clear-cut as I was led to believe.
I mean, I'm a happy new father, and I was exited to see every ultrasound we did, but the first one absolutely parsed as "alien lizard thing" rather than "small human". Early foetuses really don't look very recognizably human to an untrained eye.
No, I'm not a parent, and you have a good point. I have heard parents speak about the excitement of seeing an ultrasound (though I don't recall any of them connecting that experience with abortion).
Seeking suggestions for activities in the Sacramento–Berkeley corridor, June 16-17 (Friday and Saturday). I will have a car.
Friday, I'll be starting from Davis/Sacramento and have dinner plans in Berkeley. Saturday I'm free all day, but I have to end up back in Sacramento by day's end.
I like day hikes over rocky, uneven terrain (including some bouldering). I love scenic views. (I also like well-stocked guitar stores.)
Thank you in advance for any pointers!
Mount Diablo would be the obvious choice. On a clear day, you can see more square miles of land from the peak than any other peak in North America (or so I've been told). It's possible to drive to the top, but a better experience to park elsewhere and hike to the summit, I think.
Tilden park is also quite nice for getting views of the bay.
For the past year I've been manually tracking when I go to sleep and wake up. https://i.imgur.com/np9uI2V.png
Soon after starting I realised I can use my browser history as an approximation because most of the time I would turn my PC on and off after waking up and before going to sleep, so I have 3 years of pretty accurate data.
It seems that each year there is a bump in sleep times in winter and for some reason 2 smaller bumps in summer. What could this mean?
Since I started tracking my sleep I also tried to intentionally bring it closer to midnight but mostly unsuccessfully. It seems that I can't escape this yearly rythm of bumps.
One explanation I found is https://en.wikipedia.org/wiki/Phase_response_curve. This would mean than my sleep advances with the rate of advance of sunrise. So the ambient light coming through curtains arrives in my eyes early relative to my sleep during spring but this gradient slows down as summer solstice approaches. Though this doesn't explain the double bump in the middle of summer...
*Melatonin: much more than you wanted to know* also mentions the effect of the seasonal cycle on the circadian rythm, via ambient light.
You can probably fiddle with this effect using smart bulbs or a big light on a timer.
There used to be a group called Quantified Self with a site, meetings, forum, etc. If they still exist they'd be a good place to discuss this.
I get spam mail now sent to an alias I probably mentioned once in a thread from this substack (nowhere else, ever). Fuck you, whoever it was.
I imagine it was scraped by a bot rather than taken by a real user here
You're probably right
Was it the same address you gave me during a chat about software projects? My condolences man.
No worries it was probably bot scrapping
I just looked it up, and you are right about mail-in votes in European countries.
I think you are wrong about 'there's no particular problem with mail-fraud in elections'. The Democrats are just too openly crooked, and they monopolize the bureaucracies. And the bureaucracies are the real government, not the ceremonial remains of elections.
On a completely unrelated note, what happened to the rule that every odd (or even) thread would be non-CW?
IIRC it was too easy to forget and too difficult to police, so Scott dropped it a while back
"The Democrats are just too openly crooked, and they monopolize the bureaucracies."
How come we have a Republican House of Reps?
The last episode of this season of Hi-Phi Nation went out last Tuesday, its about three different AI music technologies, vocal emulators, fully generative music, and musical improvisation, including a guitar solo contest between two humans and two different guitar-solo generators. Check out the whole season also on the ethics of digital futures. https://hiphination.org/season-6-episodes/s6-episode-6-rise-of-the-music-machines-may-16-2023/
I was made aware of this in a newsletter about malware and security, and I have to say, I don't get it:
https://www.blog.google/products/registry/8-new-top-level-domains-for-dads-grads-tech/
"Google Registry has launched some of the most popular (and secure) top-level domains, such as .app and .dev. Today, we’re adding eight new extensions to the internet: .dad, .phd, .prof, .esq, .foo, .zip, .mov and .nexus."
The newsletter article was pointing out how .zip was a dumb idea and the problems with cybersecurity involved, but that aside, who the heck wants a domain name like .dad or .phd?
"Knock, knock. Who’s there? With Father’s Day right around the corner, .dad is here for the jokes, the games and the advice. Whether you’re a fit.dad, a gay.dad or a dude.dad, .dad is the place to celebrate fatherhood."
Uhhh - there are already sites like these, granted on the distaff side - Mumsnet and the like. Do they really expect enough people will want an address like Gary.dad in order to go "yes, I'm Gary and I'm a dad"? As for the rest of it:
"Grads
May means graduation season for many in higher education. We’re celebrating graduates and the professors who taught them well by launching .prof, .phd, and .esq. Whether sharing legal advice for everyday life or teaching courses on behavioral science, these new domains are perfect for showing off your credentials. Hats off to these early adopters:
Erika.esq: Erika Kullberg is an attorney and money expert who is passionate about better positioning people for success.
Casey.prof: Professor Casey Fiesler is a technology ethics educator and science communicator.
Rafael.phd: Rafael Misoczki is an expert in post-quantum cryptography, fully homomorphic encryption, privacy enhancing technologies, and the application of these constructions."
Ah yes, Erika, I'm sure that having a vanity plate website is going to do wonders for your professional reputation. And were I trying to look up an "expert in post-quantum cryptography", I would not be at all confused by a domain name like Rafael.phd
And this one? david.mov: Watch videos by David Imel in this liminal space.
It is horrible and redirects to Youtube, where most people are going to go looking for videos in the first place. That is not a "liminal space", it's a mess.
I agree, the ".zip" is an extremely dumb idea. Now whenever someone mentions a ZIP file in an e-mail or in a comment, it may be automatically underlined and linked to a completely unrelated website, probably owned by scammers who will register websites such as "report.zip" and "document.zip" and "data.zip" and put there a nice big Download button that many users will click without hesitation. Fuck Google for making things complicated for everyone.
A few years ago I was teaching an "introduction to computers and internet" course for some older people who missed the train. Like, how to use Word and e-mail and stuff. And of course they asked me "how can I buy things online?" and I gave them some careful advice, but I thought that the best possible answer is "just don't", because it's simply impossible to explain during one hour, or even one day, *all* the things you need to do to avoid getting scammed. (There are some good rules of thumb, such as "only trust the URLs you have typed manually", but then things get more complicated.) Now I imagine myself explaining to some 70 years old person who just got a computer for Christmas and decided to finally give it a chance the difference between clicking on the "photos.zip" attachment in their Gmail, and clicking on the automatically underlined text "photos.zip" which will lead them to some scammer's website. Why?
(Good luck to Erika; all her potential customers will type "Erika.esq.com" to their browsers anyway. Heck, even if they won't, there is a chance that the browser will helpfully update the entered URL itself, because that's what helpful programs do, don't they?)
That's a good point! Auto-correct is going to have a field day with the new domains, and of course the scammers will take advantage of that as well - you think you're clicking on Ericka.esq, the programme auto-corrects to Ericka.esq.com, and some Russian scammer is now looting your bank account 😀
Plus I've a feeling Ericka is going to get a lot of communications from lonely gentlemen requesting photographs of her beauteous form and proffering photos of their own... advantages... as well.
Well, theres far too many People and things under .com.
Maybe people will rush to buy up domains they think will be valuable, which they wish they had with .com domains. Profitable for Google, it sounds like, but probably a bad idea for the buyers.
The example sites are just market promotions. Google is selling a new product line, thats a better way to think about it. It's just another flavor of oreo. Maybe you dont like it, thats ok.
Using these less common domains isn't unusual for professional personal websites. There are only so many options for .com and .org or what have you and most of the good ones are gone. I can't get my name as a .com but I may be able to with a more unique domain.
I think the assumption is that people would rather take gary.dad than garyhasselhoffsdadblogsite83723.com. I agree that I already find myself confused when anything other than a .com, a .edu, or an occasional .org is used, so a marketer needs something really snappy to overcome the oddness of a non-standard extension.
I share your reactions. In fact I can't help wondering, is that newsletter post perhaps dated April 1st...?
The Wold is a long-running Dungeons & Dragons campaign world that emphasizes play-by-post gaming and a commitment to write one game turn every weekday. We are a community of about 65 gamers in nine long-running games, and we are looking for new members. I'm recruiting for first level players for a new game, in which the players are would-be heroes migrating to an embattled dwarf clan to help them survive and thrive. We have openings in current games, too, at higher levels. What do you need to know about the Wold?
• We are here to stay! The campaign dates back to 1985, and we have online archives back to 1996. You can expect to start a character and play through 20 levels over the course of a 10 year career.
• We are free! There’s no cost to join and no test to pass. Check out some of the links below, and if you like us, you are in. As our founder, Jerry says, “There is no tryout or application. My experience is that those players who stay are the ones who would have passed any application or tryout anyway. So I let the players weed themselves out or remain to become a permanent player.”
• We offer a very rich gaming experience! We have our own homegrown gods, nations, races, and peoples; we have custom classes, spells, feats, items, and more — all detailed in our own wiki. We play by the Pathfinder 1E flavor of the Dungeons and Dragons rules these days, but we are eager and glad to teach players who need help.
• We focus on heroic themes! All our player characters are Good. Feel free to create a troubled and conflicted character, so long as you are interested in redemption, nobility, and overcoming obstacles to become a true hero.
• We are a friendly community of people from all over the world, from all walks of life, and of all ages! The gaming groups that stick together over time are the ones in which players become friends, and we encourage that.
Check out the following links to get started:
Woldian Games: https://www.woldiangames.com
The Woldipedia: https://www.woldiangames.com/Woldipedia/index.php/Main_Page
For more information, please e-mail me:
Cayzle - cayzle@cayzle.com
Sounds fun! What kind of time commitment would it entail?
I know someone who would be perfect for this, and for whom this would be perfect. He's out of town but I will tell him about it when I see him next.
It's funny how in the ACX book review contests, all the finalists are good, but one of them is so outstandingly good that it is clear within 5 seconds who's going to win. I wonder why that is. Is this true for most contests in life, which have good enough incentives to draw in high quality participants?
I don't think so. E.g. it's unusual for any ESC entry to have a more than a 50% chance of winning, and likewise in the national competitions, judging by the odds.
It's probably true in all sorts of things, as it seems like the Pareto principle.
Well the markets do seem to have a clear favorite right now, but with only six traders I'm not sure how meaningful that is yet.
https://manifold.markets/warty/which-book-review-will-win-the-acx
You've got to be kidding me. I just made a large bet on "The Educated Mind".
I've noticed that the "hard part" of a lot of physical activities isn't what I would have expected before doing them with some regularity.
- Running: My limiting factor is breathing properly, not my legs getting tired. I wind up gasping for breath long before the rest of my body is tired.
- Bouldering: My mind being tired is a bigger problem than my body being tired. With the former, I barely trust myself to be on the wall at all, since I'm worried that I will zone out for a second and fall off and hurt myself.
- Bicycling: What makes me want to stop isn't my legs being tired, it's my butt being in extreme pain from sitting on a seat for so many hours.
I'm not sure if this is a common phenomenon related to outsider versus insider perspective, or something else.
Re: gasping for breath, I assume your problem isn't https://en.wikipedia.org/wiki/Exercise-induced_bronchoconstriction ? I'm mostly bringing it up because that's been my issue and it was a vexing journey to find out. (Thankfully, there is plenty of anaerobic exercise to be had, so I still get to exercise!)
It had occurred to me, but I've never been diagnosed with asthma, and have only noticed this problem when running. For example I biked for ~50mi this past Sunday without much trouble, but would be gasping for air if I tried to jog ~5mi.
Are there any telltale signs to distinguish asthma from other breathing problems?
I've never been diagnosed with asthma, either, mostly because it's only exercise that causes the problem; for most of my life no one even *suspected* I might have a physiological issue with sports at school, the narrative was always 'you are out of shape'. Turns out the symptoms are independent of how in-shape I am, though, or how much I take breathing tips to heart (breathe in through the nose, prefer deep breaths to rapid breaths, etc). What happens is that my throat/chest responds in a way I would describe as "like getting stabbed through the clavicle" (although with the caveat that I thankfully haven't had literal clavicle-stabbing happen to me, and so the comparison is probably bad) after even a short while of sprinting (at which point my legs usually haven't even noticed I'm using them much yet) and my ability to breathe drops precariously. If I really push myself, after I inevitably stop running, I will cough on and off for several hours, with declining frequency. (Some silly circumstances made me experience this less than a month ago while I tried very hard to catch a train and the associated coughing did not fully subside until bedtime six hours later.)
I unfortunately don't know much about how biking might or might not affect EIB, having been rather focussed on conversations about my own situation (and I don't bike; I walk or take public transport), but from my understanding of the mechanisms at work with it, I agree that this sounds like your problem is something else!
When I went on a ~100km/day bike trip, saddle soreness was considerable after the first day, but after that it was fine, and fatigue became the main problem.
You need a better bike seat and/or padded shorts.
I already have padded shorts and they definitely help, I'm thinking I might need padded underwear too
Is it just the sitting, or is it part friction / chafing ? Might want to consider a chamois cream, some popular products are butt butt'r and dz nuts.
If you add the padded undies your bum might look pretty poofy. I second drosophilist's suggestion of a padded seat.
I think my seat is pretty padded already? It's certainly not a thin stiff one
Interestingly my experience with running is now the reverse of what you describe: I tried going for a run a few times recently after many years not doing it and my feet started to hurt way before I reached the limits of my cardiovascular capacity. It used to not be like this, so I am tempted to blame aging (and generally being out of shape).
I suppose it makes sense that physical exercise in general is a weak link problem.
I've literally had this switch up on me depending on when I take it up again. Right now it's 100% legs, but when I was younger and more in shape it was definitely CV capacity.
What sort of running are you doing, and how long have you been training for? Most amateur distance runners who have been training for more than a few months find their legs end up being more of a limiting factor than their breathing. (The conventional advice is to do most of your training at an "easy" pace that would allow you to carry on a conversation. But it sounds like you are running faster than that, so maybe you're following a different training regime.)
I have had the same experience as the OP about running, for basically my entire life. I can’t remember a time I stopped running due to my legs, it’s always my lungs that give out first.
The kind of running I do is running at a fast pace (if you were generous you could call it a sprint) until running becomes uncomfortable enough that I want to stop. Then I walk until I feel good enough to start running again.
I used to run like that and later found an expert runner who agreed to be my coach.
He told me that running like that doesn’t allow you to improve much since you are running in your anerobic zone, which has a much smaller training response. So hard to improve the bottlenecks of the sprint-running
Running slower means you spend a lot more time in the aerobic zone which is much more trainable. So it’s best to spend upwards of 70% of your runtime in the aerobic zone.
Anyway Experience bore him out and after I switched to running much slower but continuously, I saw my endurance and speed greatly increase - so that I can now run continuously at aerobic zone at my previous sprint-speed in anerobic zone.
Had to come back to say that your comment actually changed my life.
I have never had an exercise routine, though not for lack of trying. I was never able to make it stick. But your comment here led me to try running slower for longer, and to focus on running longer as my improvement target than running faster. And it worked! It worked great! Now I run 5 days a week, my endurance has improved significantly, and I'm happier and healthier as a result.
Thanks!
Your experiences are "normal" and have one thing in common (except maybe the bicycle seat pain): it is your mind that makes you feel tired. It doesn't mean it's not "real", it's just that the physical limits of the body are far from exhausted by this point. A lot of endurance training is literally learning to "endure" the pain/discomfort/boredom/ of extended hard work.
Very good point, that mind/body distinction has occurred to me in the past as well
Interesting conception of modern LLMs not as software (reliable and precise, but extremely limited), but as a crappy human intern (inconsistent and low-quality at most things, but can be useful). https://www.oneusefulthing.org/p/on-boarding-your-ai-intern
Interesting essay! And when LLM can learn incrementally, the AI interns will come with résumés.
YouTuber I recently came across in the realm of political philosophy. I think this was the first time I'd seen a clear minimally-politicized explanation of what "cultural Marxism" means: https://youtu.be/4JX4bsrj178
I would not take this fella very seriously if I were you. I don’t know if his clumsy grasp of Marxism is due to bias or genuine ignorance, but I t seems to me he doesn’t want you to take “wokeness” (identity politics) seriously as a genuine intellectual tradition while couching it in the aesthetics of being non-politicized. Anyone who paid attention to my freshman civics class could take apart his characterization of identity politics. There are people communicating political philosophy at a far higher and dispassionate level than this guy.
> There are people communicating political philosophy at a far higher and dispassionate level than this guy.
Could you share some links?
Call me crazy but you can’t go wrong with Crash Course World History:
https://m.youtube.com/playlist?list=PLBDA2E52FB1EF80C9
Academy of Ideas: https://m.youtube.com/@academyofideas
Hah, I watched Crash Course years ago. I also subscribed to academy of ideas but couldn't get into it, it was a little too dry for me maybe.
Has anyone else tried both prescription + over-the-counter options for ADHD? Historically I was on an extremely low dose of instant-release Adderall (I would break up the pills into tiny ~1mg pieces). The main downside with that was the inevitable crash in the afternoon which turned me into a zombie in the evenings. I guess that could be fixed by using an extended release instead, but the problem with XR is that I can't break it into lower-dose pieces, so the dosage would likely be much higher than I'd be comfortable with.
Lately I've been using over-the-counter Thesis instead, and while it's better than nothing, it's certainly no Adderall. But on the plus side, there's no crash.
With Adderall XR formulations, it's very possible to split the dosing up. XR adderall is just capsules that are filled with what look like round sprinkles that have a fairly simple bi-phasic release mechanism: 50% of the dose up front, with the other half released ~4h later. If you wanted to recreate your 1mg dosing, you'd just measure out 2mg of the XR sprinkles.
Similarly, Vyvanse is just powder in a capsule that you could split up into portions if you so choose.
Interesting, might be worth a shot then. Seems like it would be a lot more involved than just breaking a tablet into pieces though
I have never heard of a dose smaller than 5mg (which they do have in XR). Why were you breaking them up so small?
I find the best way to break through the crash is to do something active around the time of the crash. Could be exercise but could also be a walk or an errand or making dinner. Anything to take my mind away from noticing the crash and keeping my body moving.
I found that with the standard 5mg doses my mind would "run away without me" and I would be less productive on net. There'd be lots of movement and action but with little directional coherence to it. An extreme example (which was years ago and on a much higher concerta dose) was getting sidetracked from my math homework by some halfbaked idea, scribbling about it furiously for an hour+, only to come down and realize that it didn't make any sense and I was just being crazy.
Wow thats very interesting and I am sorry that you had such a response. Unfortunately I don't have any suggestions to help! In reference to your other comment, you can be a pioneer in microdosing Adderall!
What about breaking up the instant release adderall into even tinier bits and spread out the period of dosing more, so your crash happens closer to bedtime? Or dissolve it in water, and take measured doses of the liquid?
I haven't tried that, but it might be a good solution. I don't think I've heard of people micro-dosing Adderall before though?
It's not microdosing, though -- you'd be taking your usual total dose, just spread out longer by taking it as many tiny doses. Another thing to try would be to take maybe 10% more than your usual dose, having the last dose be in the early evening, so that the crash comes around bedtime.
One last thought: A patient of mine with ADD is just starting to try meds. He first tried ritalin, which made him feel unpleasantly jittery. He was able to stick to things for longer periods of time, but they were things like organizing a room, not doing his job (he works from home). Then he switched to adderall, and find that the jitteriness is much less, and that he feels more able to focus on the things he knows he needs to focus on, rather than on whatever he happens to be looking at when the energy first hits. So if you've never tried any other ADD meds, it might make sense to try one. There are several different types.
I have been on other ADD medications over the years, Adderall definitely works best
Lengthy piece (from a Caribbean economist) arguing that Caribbean nations should abandon their local currencies and just use the US dollar, as Panama, Ecuador, and El Salvador do now. Apparently a candidate for president of Argentina is loudly arguing for doing the same thing. Author takes on/discredits arguments for monetary sovereignty, devaluation, etc.
https://rasheedgriffith.substack.com/p/notes-towards-caribbean-dollarization
I've thought for a while that a stablecoin US dollar might actually explode dollar usage globally- not sure developed country residents understand how strong the demand is for a stable currency in inflation-prone countries (could be the euro or the yen too).
Kind of funny that as US relative power declines, some aspects of it actually increase. I'd argue the US has more cultural power (TV shows, celebrities, music) than 20 years ago, even as US military & economic power has declined a bit. Apparently English is just becoming more & more dominant as a global lingua franca. Maybe the US dollar will actually (counter-intuitively) become *more* entrenched as the global reserve currency, even if the world shifts to being multipolar!
EVs and LNG import terminals are what the Caribbean needs…everything else macroeconomically speaking will be easy once oil no longer dictates so much of island life.
I think it was Noah Smith who recently made the case that Jamaica should try to get the new mini-nuclear reactors installed as quickly as possible. Apparently they already run a research reactor, so switching to nuclear would help their economy a lot without adding much proliferation risk. It was an interesting idea, and if it worked, could potentially mean that long term more island nations could switch to nuclear.
Argentina already had that from 1991-2002, in form of fixed exchange rates to US dollar. It has large pros and cons.
Isn't that vulnerable to the Peso Problem? The exchange rate is fixed but only until the government says otherwise.
in Argentina, it's indeed touted as a way to restrain monetary policy which everybody agrees is consequence of bad fiscal policy. And it's a red herring because to make the switch you first need to fix the fiscal situation, no way around it. Once you do that, there's no need to buy foreign currency that is losing ~8% of its value every year.
Fiscal deficit with sovereign currency = high inflation. Fiscal deficit without sovereign currency = complete and total crash.
The candidate that is touting dollarisation is pretty much a contrarian teen and his arguments are basically burning the ships.
I just think it would be interesting if the people of Argentina (or any other developing country) started using US dollar stablecoins to pay each other, and it eventually reaches such a large enough scale that the authorities have to throw their hands up & give up. I mean, that would probably take a while, and I'm sure the Argentine government would be Very Very Mad about it. I'm just talking about a groundswell of regular people paying each other for food, gas, clothes, whatever with peer-to-peer dollar coins. I guess they'd probably need a bank or bank equivalent that's outside of Argentina's regulatory authority too
Governments in much smaller countries than Argentina were able to enforce payments in national currency, why Argentina would fail?
If Jim Cramer is such a terrible investor, couldn't you make money with an investment strategy to do the opposite of whatever he suggested?
Yes, the funny corollary to the efficient market hypothesis that says no one can consistently outperform the market is that no one can consistently underperform the market. That being said, the place retail traders go wrong is they have no idea the risk they are taking (or they don't care) and will blow up eventually.
You can easily underperform the market in risk-adjusted terms by refusing to diversify. If I buy a limited number of stocks from the S&P 500, chosen at random, I should expect to have more risk than someone who buys the entire S&P 500, and no better average returns. The market rewards undiversifiable risk (beta) in stocks, not diversifiable risk.
"Alright," you say, "but risk here is just variability, not the same thing as returns."
Ah, but provided that my margin costs are less than long-term stock market returns (and they generally are), I can turn more risk into more expected returns by buying on margin. Therefore, there is some positive level of margin below which a well-diversified portfolio will have less risk than an unlevered randomly-chosen undiversified portfolio, and greater average returns.
It's more complex than that.
One, collectively investors underperform the market by the friction costs.
Two, anyone can consistently outperform the market via pure chance, see also an example of someone consistently throwing heads 10 times in a row - out of millions of participants in the coin-tossing trial. The key point of the EMH is that it's impossible to predict in advance who that lucky 10-heads person will be.
FWIF this is the only part of EMH that is practically useful. The stronger version is hard to defend.
I'm fairly sure that consistently underperforming the market is as easy as consistently picking the wrong number someone is thinking of between 1 and 100. Possibly easier, given how many stocks there are.
Or to put it another way: one cannot discover the next miracle material by combining different substances at random.
If someone *always* got the wrong number, then that would be useful information - you could guess any number besides the one they picked and you'd have a 1 in 99 chance instead of a 1 in 100. Not a lot of information, but it's something. But in reality a person who "consistently" gets that one wrong is still right 1 in 100 times.
Similarly, if someone *always* underperformed the market, you could make money by shorting whatever they picked, but in reality these people will occasionally pick winners. (Or rather, they would if they had enough money to keep picking stocks forever instead of going bankrupt.)
I learn something new every time I read this forum.
Is there a name for that inverse of the efficient market hypothesis?
These are both instances of the “optional stopping theorem”. If something is really a random process, with expected gain or loss at any moment equal to zero, then any choice of when to play and when to stop that is based only about information on the past and present will also have expected gain and loss zero. (The efficient market hypothesis basically is the initial assumption, but it is often stated as one half of the conclusion of the theorem.)
https://en.m.wikipedia.org/wiki/Optional_stopping_theorem
Well now you can buy that exact ETF. it’s not doing great since inception but is still beating the S&P. https://finance.yahoo.com/news/bet-against-jim-cramer-etf-122021151.html
Haw haw haw!!!
Anyone else notice the strong Christian themes in the new Guardians of the Galaxy movie? For populist entertainment, it surprised me. [spoilers ahead]. For example, the villain is the God-denying “High Evolutionist,” who engages in the worst imaginable form of eugenics because he alone can perfect life. The end of the film is basically Noah’s Ark, and the hero StarLord dies while saving everybody then comes back to life. There’s also a theme of every life is sacred. It’s just a bonus, I suppose, that the villain is black and the children he is tormenting are all blonde-haired, blue-eyed Aryans.
As a Christian, I can't say any of that comes across as particularly "Christian themes" to me. The villain is the most obviously evil and repugnant villain Marvel has had in a long time (... err, unless any of the other villains in Phase 4 were similarly evil, can't say I've seen them), so it's not exactly a strong Christian theme to say "maybe inhumane experiments on animals and literal planet-destroying genocide is bad".
You can draw a few parallels between the Biblical Flood and the end of the movie - certainly High Evolutionary has an ironic God complex, but they're fairly weak and felt largely unintentional to me.
And if they were intentional, I don't know that casting your megalomaniacal villain in the role of "God" in the biblical story is exactly what I'd call a "Christian theme" (at least not of the sort that I'd be surprised to find in modern pop culture).
---
The bit with Starlord at the end is just standard hero stuff: I don't think every hero who almost dies is a Christ figure.
And, I think this film *could* have taken an "every life is sacred" message, but really doesn't. Mowing down a bunch of villainous mooks in the action sequence in order to save a bunch of cute innocent animals... if you wanted Christian themes, it would actually be the *reverse* of that! On both counts!
(If anything is a Christian theme in this movie, giving Orange Boy a second chance is the closest it comes, but even that's a fairly milquetoast "'forgiving' someone who really hasn't hurt you" act)
All that and you didn’t even mention the character who goes to heaven and is assured that a higher power is guiding things? It’s the most Christian part of the movie by far. Surprised me.
Good point!
" It’s just a bonus, I suppose, that the villain is black and the children he is tormenting are all blonde-haired, blue-eyed Aryans."
Meaning unclear, please develop your point. I have a vague feeling you intend to convey something, but I just can't quite grasp what it may be. Are you Darkly Hinting?
As for the High Evolutionary, he's been around since the 60s. He started off as a white Englishman, having a black actor portray him in 2023 movie is just DEI.
https://en.wikipedia.org/wiki/High_Evolutionary
Also...Adam Warlock, in his, let's say, second revamp, was the most obvious and heavy-handed of Christ allegories. This was dropped by the time of the comics that are actually the reason he's in this movie (Jim Starlin's 1970s run).
In the US context, there’s a (historically contingent) link between Christian fundamentalists and racism. This particular casting choice, coupled with the Christian themes, made me wonder if the movie was intentionally playing to that audience.
You're thinking of the historical correlation being Southern and being racist. Those sects which had many fundamentalists--the Baptists and Methodists--each split into a northern church which was not historically linked with unusual racism, and a southern church that was. Factor that out, and I think you won't find any remaining correlation.
OK, I didn't do a deep dive on this, but here's a well-cited article defending the connection: "Why Do Religious Fundamentalists Tend to be Prejudiced?" (https://www.tandfonline.com/doi/abs/10.1207/S15327582IJPR1301_03?journalCode=hjpr20) and one arguing that "Priming Christian Religious Concepts Increases Racial Prejudice" (http://www.baylorisr.org/wp-content/uploads/rowatt_priming.pdf). But your mileage may vary.
Let me be sure I have this right: you are suggesting *Marvel* is pitching its movies to *Christians* and not just any old Christians, but *fundamentalists*.
Is this because of Chris Pratt? Because I'm given to understand the kind of people who were big fans from the Parks and Rec days turned on him due to him belonging to a church that wasn't all gay marriages carried out by trans lesbians ministers with rainbow stoles.
Just come out and say "Nazis". You know you want to, with your "blonde Aryan children" bit.
EDIT: I may be being unkind. This could be a test run for another Quillette article as per your website. If you can get it to fly here, you can easily crank out something about racist Christian fundamentalism for them.
Wow, OK, that seems unnecessarily aggressive. I think I’ll pass on a further response.
Spoilers!
I gave a warning!
Does anyone else get the impression that the debt ceiling "talks" have some other objective? It seems more and more like theater, with the announcement of reaching the ceiling months ago, and soon we will have a last minute "save". Oh, the drama!
I actually am thinking this is about something else completely, and we, the public, are completely in the dark as to the actual objectives.
It's the same as it's been all the previous times. Chicken plus bartering. We start off with "you're a nazi" vs. "you're a communist", proceed through "your policies will destroy America and all Americans", get closer with "our version of economic stimulus is better than yours", and end up with "we'll cut more from your programs than you want but less than we want".
So yes, totally theater. The point is to generate public pressure on the other side, by implying that We're completely reasonable and the only problem is Their intransigence. Eventually it'll pan out in serious negotiations behind closed doors where the press can't go, so that each side will be able to justify their compromises by pointing to all the things the other side didn't get, and no individual politician will be held responsible by their voters for failing to uphold their absurd promises.
Or maybe this will be the time it all breaks down! Maybe we've finally run out of adults in the room, and all that's left is kids who haven't found out that Santa Claus isn't real. But that's never happened, so by induction it never will, just like death and AGI ruin.
That should be "haggling", not "bartering". Stupid neural net hallucinations.
I don't follow it in detail, since I don't live in the US, but what exactly is in the dark?
The incentives are those of a chicken game, where the player to give in first is punished. Are the players not just following the natural strategy?
I think that we have a very clear idea of what their objectives are; each party wants their Presidential candidate to win the next election. Right now both parties are playing a game of chicken to see which side voters will blame for a default. My interpretation is that they don't care about the debt ceiling, they're just using this as a stage to score political points. I don't think that there's any deeper meaning than that
I think this is just the nature of modern politics. Much of what the public sees has been written, edited, and shoved in front of a focus group. It's probably not too healthy to put too much into thinking there are completely different objectives hidden from the public, but it's also somewhat understandable given our institutions' track records of late.
I have been thinking it is worse than they let on, because few who are taken seriously are talking about budget reforms which might prevent us from getting in these situations. It seems like the people who very reasonably suggest our direction isn't wise and probably not sustainable are labeled cranks.
Giving the President the power to borrow unlimited money in the name of the American taxpayers, with no oversight or accountability, would seem to be a bad plan. Unless you believe in the Magic Money Tree, I suppose.
Does the president have the authority to borrow unlimited money? I thought the executive is constrained by laws in what it’s allowed and mandated to spend money on, and the debt ceiling is just a contradictory law banning the executive from borrowing the money to pay for what it has been mandated to pay for.
I don’t think anyone is proposing that the executive get discretionary fiscal powers.
The president can do things like sign contracts whose nominal value falls within authorized and appropriated limits but will predictably escalate beyond them, or tell his AG to take a dive when a favored interest group sues the government for Sagans of dollars, at which point the government is legally obligated to either cut back spending somewhere else (which may be illegal), or issue more debt. A debt ceiling subject to congressional review means that it's usually not worth gaming the system that way; I'm not sure what the equilibrium will be if that goes away.
Maybe. One of the weird aspects of our current setup is that so much of our debt is offloaded to foreign central banks that devaluing the currency may be a good way to transfer wealth from them to us. Which is not *morally* good but is certainly a hell of a grift if you can make it work.
During an inflationary time, it might be good to temporarily have a surplus. But usually, a surplus is when the government is profiting off the people, which is the fundamental thing a democratic government should never do.
Best to pass a debt ceiling increase that *doesn’t* come with a bunch of UN negotiated cuts to every aspect of the government. Get rid of the terrorist threat *before* you begin to negotiate, don’t just give in to the terrrorists before asking them to give up their threat.
>The only risk is that certain partisans -- the sort of people who genuinely think that "let's mint a trillion dollar coin" or "let's use the Fourteenth Amendment as an excuse to bootstrap a literal military coup" are workable public policy
Why isn't the trillion dollar coin workable public policy?
I agree with trebuchet that it's a ridiculous concept born in unserious minds, but here's an attempt at explaining the reasoning.
If we printed a trillion dollars worth of smaller bills to pay off debts, what would be the result? Obviously with more money in circulation and no additional goods, we would see significant inflation. That would be bad.
Would that be any different with a trillion dollar coin? If not, then nothing is gained. If so, then why? Probably because a trillion dollar coin could never be spent. I guess we just give a major creditor the coin and tell them they're paid off - and let those suckers try to spend it! There's no way to exchange it for smaller bills or buy anything with it, so it's a bit like an IOU from the feds to whoever. If I were a creditor of the United States, I would refuse to accept it.
> If not, then nothing is gained. If so, then why?
It is literally on the Wikipedia page - "The issuance of paper currency is subject to various accounting and quantity restrictions that platinum coinage is not."
See Legal Basis in https://en.m.wikipedia.org/wiki/Trillion-dollar_coin for more details.
If we're paying off the debt owed to internal stakeholders, like the Social Security fund, then all we're doing is shifting money around and screwing over Social Security.
But if I give them the right amount of power, then demagogues will be too spooked to keep doing this dumb bullshit
The Alignment Research Center has an option to donate.
https://www.alignment.org/donate/
In re hypergamy: I haven't read all the comments, but does this mean that some number of men (incels and the like) built a lot of their lives around an idea that might not even be true?
I don't think anyone becomes an incel on the basis of a theoretical belief regarding their probability of romantic success, but rather through hard experience that they don't have whatever it is that women want in a romantic partner. Their having a possibly-false idea of what that really is, seems like a minor perturbation that changes nothing.
Also, there are I think a fair number of incels who either are "high status" in the specified markers of class, wealth, education, and age, or can reasonably expect to be in 5-10 years when they graduate and get an UMC job. And more who aren't, but the ones who are wouldn't be surrendering to inceldom if they believed in the sort of hypergamy that Scott is talking about.
It reminds me of those who thought Piltdown man being a hoax means evolution didn't really happen. When you are strongly, emotionally motivated to deny something, it's not difficult to find some false claims connected to it, laser-focus on them, and then declare the whole idea to be discredited.
Is glass half full or half empty?
What post is this referencing?
https://astralcodexten.substack.com/p/peer-review-hypergamy-much-more-than?utm_source=post-email-title&publication_id=89120&post_id=119932529&isFreemail=false&utm_medium=email
Presumably those of us who aren't allowed to read that should also consider ourselves not welcome to join in discussing it?
Gotcha. I'll read it when it comes out then.
The Peer Review posts are Scott looking for feedback from subscribers on early drafts. He says in the linked article, "I’ll be posting this sometime in the next few weeks."
I'm sorry-- I didn't check how accessible it is. Of course you can discuss what I wrote, even if you don't have access to what Scott wrote.
I'm not sure whether courtesy to Scott would permit me to quote from it, but the conclusion is that people pretty much marry people of similar attractiveness and from the same class.
Class matters much less to homosexuals (box sexes) than to heterosexuals.
Many, many people build their lives around idea that are not true; take for example Religion, where at most 1 of them is right and thus at least 6 billion people quite strongly believe something profoundly wrong.
That's certainly true, but it's a much shorter chain of reasoning to start with "Women have a strong preference for high status men" than "this is the one true religion".
I have a notion that part of believing in hypergamy is that people who play gender roles hard are conspicuous, so it's easily to notice the supermodel married to the rich bodybuilder, and not notice the pleasant-looking woman married to the mildly successful guy.
One upshot of a "shorter chain of reasoning" is there's less space to smuggle fallacies in. God is a vague and nebulous thing to argue for/against. Meanwhile if hypergamy were obvious poppycock it should be easy to see if it's not true/has bad explanatory power. Yet to the casual eye, it seems to hold up.
Not sure what you mean in your second paragraph - isn't a 10 going with a 10 and a 5 going with a 5 not a positive example of hypergamy? Both women have scored the highest value man they can get.
Not much point my saying any more though until I can actually read the source article.
> isn't a 10 going with a 10 and a 5 going with a 5 not a positive example of hypergamy? Both women have scored the highest value man they can get.
And so did both *men*, in this example. A pairing of 10 with 10, and 5 with 5, is simply everyone choosing the best option available (assuming monogamy). The only thing this proves is that people's opinions on who is attractive are correlated, otherwise we couldn't even meaningfully talk about 10s and 5s, because no one would agree who is who.
The actual claim of hypergamy is that women are attracted to men's *social status*. As opposed to men, who are attracted to women's pretty faces or big boobs. Learning that she also happens to be a rich CEO does not make her any more fuckable, to put it bluntly.
There is also a stronger claim, that women prefer high-status men *so strongly* that they would rather be the rich bodybuilder's second girlfriend than the mildly successful guy's only girlfriend. (While men have a stronger preference not to share their girlfriends with other men.) So if you entirely remove the social pressure towards monogamy, you should expect the man#10 to pair up with the woman#10 *and* woman#9 and possibly also woman#8; the man#9 would be with woman#7 and woman#6; etc., and the men#3-1 would be left alone.
(I haven't read the article either. Just commenting on the definition of "hypergamy" as I have seen it in internet discourse.)
My impression is that worry about hypergamy was men believing they'd be left out because all the women are looking for higher status partners.
This is what's known in cryptography as a "Grandmaster attack", after the story of the little girl who plays two chess grandmasters at once and comes away with either a win and a loss or two draws.
There are a bunch of religions with mutually contradictory claims. Working out whether any one of them is correct may requires thought, but proving that all but at most one of them are not is trivial.
How do spam emails know my SO's pet name for me?
I’ve been having a weird experience with spam emails lately. I get the usual mountain of spam emails in my personal email account spam folder, except they are addressed to the pet name my partner calls me. Crucially, this pet name is correct phonetically, but not as she would write it down. Is it possible that spam email accounts have obtained this name through listening to our conversations via one or more devices? I don’t use this pet name for any online accounts or usernames, and she only uses it around me (not when we’re with friends etc). It's only really a verbal nickname for me and isn't written down anywhere.
Below is a pseudonymous example of the scenario:
My name is Peter Johnson
My GF calls me "Peej" (and writes it like that in texts)
My email account is PeteJ@whatever
I get spam emails saying “Dear Peaje Get 10 free spins etc”
How come these emails are formed this way? Is one of my devices listening to me?
N.B. I don’t have an Alexa or any smart speaker type thing that is designed to listen around the house.
Your initials are P.J. The spam is just using some crude algorithm to get around blocking filters by using your initials to make something like a name that looks familiar to you.
There is something called Display Name Spoofing where your full name is used:
https://www.itro.com.au/display-name-spoofing-latest-scam-technique/
I don't think they're listening in on you (yet).
Here's a weird one: I went out to dinner with my daughter and some other people last week and was introduced to a new IPA called Madonna. A few days later I ordered some wine and beer from Drizly (liquor store delivery service) and as I was ordering it it suggest I get some Madonna. The reservation was under my daughter's name and cell phone number, & she has the same last name as me. I paid for the restaurant dinner in cash, not with a card. How the hell does Drizly know I drank a Madonna?
Probably just that someone in your social graph (maybe your daughter) got associated with liking Madonna. Then the delivery app queries your social graph app for suggestions and voila! Ain't technology great!
Alternatively it could just be a coincidence. Maybe Madonna is just the hot new IPA right now or is on a marketing kick.
Another possibility: the new drink at the restaurant and the suggestion from Drizly are part of the same marketing push.
It's possible that your information was obtained through a data breach or by other means of unauthorized access. Scammers can purchase lists of email addresses and personal information on the dark web, which can then be used to target individuals with spam emails.
I get that - and I'd expect spam emails to target me with my real name. What's surprising is that the name I get addressed by in these spam emails is a name I only get called verbally, written down as if overheard - surely this means it can only be obtained through listening to conversations?
By default, your cell phone is always listening to you, so it can tell if you say something like "Hey Siri" or "OK Google". This incidentally means it is analyzing everything it "hears".
I have turned off these kinds of things, which is hard to explain how to do without crippling the usefulness of your device. I don't get as much targeted advertising as my wife, who also thinks she has turned off these things. And I have my doubts as to whether they are truly turned off, or instead only not used to openly target ads.
I've worked on a system like that, and in our case, the system actually wasn't always listening; it didn't have enough power/battery to be doing that. What it had was a fixed function thing to recognize only the wake word, and the system ran in very low power until that triggered. I would expect this to still be the case for cell phones, but not necessarily for wired-power devices such as the Amazon and Google home assistant products.
Maybe through data analysis. Certain pet names are more likely to be given by people with certain characteristics. The more characteristics in the regression the better the prediction.
While I appreciate the originality, this idea just wouldn't work. For one, you once again assume Trump supporters are going to react to things the same way you would. But for another, the setup is physically much less safe than you seem to think. Elephants weight a *tonne*, and there's no guarantee all that weight will stay sitting obediently in one place. Imagine if one mouse gets in and causes panic amongst the herd: now the whole airship is coming down.
Looks like this comment was orphaned, no idea what it's talking about, but out of context it's fascinating.
It's a truly magnificent example of a paragraph that you can't imagine how it came about. Pyrocles' response just makes it all the more tantalizing...
Curse you, Substack!
DataSecretsLox has a "DSL Out of Context" thread for comments like this. (They're deliberately pulled out there.) I'd almost want to see an "ACX Out of Context" subthread. Almost.
The question is how much that matters at relativistic speeds. Adorno addresses this, btw.
Not relevant: OP stated that they'd be watching The Matrix Revolutions. Going anywhere near relativistic speeds would introduce time dilation that would make that film last way longer than is humane.
I've written a piece on 'Oumuamua recently (https://www.thequantumcat.space/p/oumuamua-alien-spacecraft-or-dark). Avi Loeb argues it was probably a piece of alien space trash. The rest of the astronomical community is sceptical and has rapidly coalesced around the idea of a "dark comet". We saw no signs that it was a comet when it flew past us, but recent papers have argued it probably had an invisible hydrogen tail. Loeb, meanwhile, says the dark comet idea is bullshit and physically unsound. Since 'Oumuamua is now far beyond Pluto we'll never know for sure, but it is fun to speculate.
There is a pattern similar to "God of the gaps" that one could call "Aliens of the gaps", which is a secular version of same, with just as much credence that should be assigned to it.
Good article. In addition to what was mentioned with unaccounted acceleration, I was under impression that the long shape of Oumuamua was unstable and very likely to break up. Does the dark comet hypothesis with unseen hydrogen propulsion address that aspect? Or is that more of a red herring? thanks,
The argument is that the dark comet would have eroded as it flew through the solar system. We found it quite late into its voyage, so we would have seen the eroded form, not the original form. It may be that that shape is actually unstable, and Oumuamua could have broken apart after we saw it.
This is another update to my long-running attempt at predicting the outcome of the Russo-Ukrainian war. Previous update is here: https://astralcodexten.substack.com/p/open-thread-272/comment/14821511.
14 % on Ukrainian victory (up from 13 % on April 17).
I define Ukrainian victory as either a) Ukrainian government gaining control of the territory it had not controlled before February 24 without losing any similarly important territory and without conceding that it will stop its attempts to join EU or NATO, b) Ukrainian government getting official ok from Russia to join EU or NATO without conceding any territory and without losing de facto control of any territory it had controlled before February 24 of 2022, or c) return to exact prewar status quo ante.
45 % on compromise solution that both sides might plausibly claim as a victory (up from 43 % on April 17).
41 % on Ukrainian defeat (down from 44 % on April 17).
I define Ukrainian defeat as Russia getting what it wants from Ukraine without giving any substantial concessions. Russia wants either a) Ukraine to stop claiming at least some of the territories that were before war claimed by Ukraine but de facto controlled by Russia or its proxies, or b) Russia or its proxies (old or new) to get more Ukrainian territory, de facto recognized by Ukraine in something resembling Minsk ceasefire(s)* or c) some form of guarantee that Ukraine will became neutral, which includes but is not limited to Ukraine not joining NATO. E.g. if Ukraine agrees to stay out of NATO without any other concessions to Russia, but gets mutual defense treaty with Poland and Turkey, that does NOT count as Ukrainian defeat.
Discussion:
This update is brought to you by a cheap natural gas.
Price of natural gas on Dutch TTF exchange, which largely determines gas and thus electricity prices over the EU, over the previous week dropped back basically to pre-covid levels. You can follow it here: https://tradingeconomics.com/commodity/eu-natural-gas. I think it is probably caused mainly by weaker than expected (by me, anyway) economic recovery in China.
My very crude model goes something like this: Ukraine probably cannot avoid defeat without substantial Western support - > this support is constrained by how much public pushback/support is there against/for it - > magnitude of the public pushback is dependent on economic costs of support - > cheap natural gas means lower costs - > which means higher political support for Ukraine.
I also think, however, that low prices are dependent on the fact Russian energy exports are still substantial (although no doubt much lower than they would be without sanctions), and further attempts to squeeze Russian economy would again prop them up. Given that current level of sanctions and overall support to Ukraine is imho probably insufficient to safeguard Ukrainian victory, I’ve updated only modestly (alos because low price might not last, of course).
*Minsk ceasefire or ceasefires (first agreement did not work, it was amended by second and since then it worked somewhat better) constituted, among other things, de facto recognition by Ukraine that Russia and its proxies will control some territory claimed by Ukraine for some time. In exchange Russia stopped trying to conquer more Ukrainian territory. Until February 24 of 2022, that is.
Over the past twenty-four hours we've gone from Russia invading Ukraine (very likely permanently), to Ukraine invading Russia. Since your last update, we've also got the UK giving them Storm Shadows and the US authorizing F-16s as soon as the pilots can be trained. And all that only increases the probability of Ukrainian victory by 1%?
Yea, because none of that is a really surprising
If you were anywhere close to 99% on any of that, you were grossly overconfident.
Of course I didn't predict exactly that. But it is, like, within the general class of expected events. In, I think, February, I did alter my prediction after it was decided to deliver Leopards to Ukraine, but after that, current ramp-up of support was something to be expected.
I guess none. That plane is in service since the 70s, surely Russians had to get their hands at it, or at least its parts, during that time?
Also, the planes will be operating over Ukraine, not Russia.
I think recent developments - Russian offensive running out of steam without much to show for and Western decision to provide planes - should mean a couple percentage points more for Ukraine.
Regarding the gas prices, it's also possible that the acceleration of build-out of LNG terminal capacity plus warm winter has resulted in both high stocks and greater confidence in maintaining adequate levels in the future.
I've enjoyed reading these and would be interested in seeing all the predictions to date in a single place, plus perhaps abbreviated summaries of the reasons behind the changes. I'm particularly curious about the reasons for the biggest deltas, and in the reasons proximate to any long-running directional changes in your estimates.
That's interesting. I was just asking myself, what's the point of this all?
I mean, it is nice to assign probabilities and publish them, because then you create a public record... of being good at predicting or being bad at predicting. And if you make 100 public predictions, and then evaluate the outcomes, you can even make a nice graph out of it (predicted probability vs the actual frequency of outcome).
But making *one* prediction and then updating it many times is something quite different. There is no way to evaluate whether the prediction is good or bad. Suppose that one year later Russia wins. What exactly does it say about how good Ales is at predicting? Suppose that one year later Ukraine wins. What exactly does it say about how good Ales is at predicting? In my opinion, almost nothing. If you give something 14% or 40% that means it can go either way, and even if the thing with predicted 14% actually happens, I wouldn't really count it as a mistake. The calibration is in numbers -- if you make ten predictions at 14% each, and 9 out of those things actually happen, then I might laugh at you. If you make one 14% prediction, and the thing happens, there is nothing to laugh at; you already said that this is supposed to happen in 1 out of 7 situations.
Furthermore, predicting something with 1% precision (e.g. updating a probability of Ukr victory from 13% to 14%) is an implied claim of very strong prediction abilities. I would rarely dare to predict a real-life outcome with precision better than 10%, or maybe 5% if I feel exceptionally confident. Then again, I am not really good at predicting, so it is plausible that other people are way better than me at doing this. But as I said, there is no evidence here whether Ales is actually good or bad at predicting. Maybe his predictions are amazingly correct given all publicly available evidence. Or maybe they are laughably wrong. How would we even tell the difference?
For myself, the war is rather salient to economic situation in Europe, so I've allocated my own portoflio partially based on what I think will happen (so, yea, mine non-Japanese Asian stocks are underperforming, see prediction above). Perhaps that might be useful for someone else.
I agree that it is not a good way to tell whether I am good or bad at predicting. But I can tell that by other means - for instance, my prewar prediction on what'll happen was, I think, pretty good, although it was less quantitative and also I didn't bother translating it to English and posting it here. And to pick perhaps more measurable example, I got to 85 % in Scott's 2022 prediction context (I just looked it up, results are here: https://astralcodexten.substack.com/p/who-predicted-2022).
A petition that seems important, unfortunately those with poor critical thinking skills haven't questioned the assumptions the media makes about the issue, vs. basic logic and history:
https://www.change.org/p/ban-mail-voting-stop-domestic-voter-abuse-restore-fair-elections
"Stop Domestic Voter Abuse, Restore Fair Elections: Ban Mail Voting (with at most extremely limited exceptions) in every state.
The secret ballot is recognized globally as a human right: but a vote cast by mail does not protect the voter's secrecy in a safe space like a polling location. Behind closed curtains, so no one knows it is happening, a vote can be coerced by domestic partners [ the CDC says 49% of women and 46% of men experience coercive control by a partner in their lifetime, would voting be an exception?], employers or unions or others and can tilt elections, as evidence from around the world, history, and logic explain.
Human nature has not magically changed to prevent an exact repeat of problems this country suffered through before.
Voting was not secret for this country's first century; votes were cast publicly. That led to the growth of such widespread problems, including violence and deaths, that in 1889 President Grover Cleveland declared they had become: "perils which threaten the existence of our free institutions, the preservation of our national honor, and the perpetuity of our country." Most states adopted them over the next few years.
It is crucial in this toxic political era that we prevent an internet accelerated repeat of that era with an internet accelerated campaign for reform to restore the secret ballot by banning mail voting (with at most extremely limited exceptions if no other approach can be found). You need to pass this along to others and help them grasp why this issue is important.
For detailed logic and cites supporting these concerns, see TheBigIllusion.com . People have blind spots preventing them from taking the problems seriously. That site addresses those and can help you persuade others. It also addresses other major problems with our democracy hiding in plain sight: but you do not need to agree with anything it says to agree we need to Stop Domestic Voter Abuse and Restore Fair Elections by Banning Mail Voting (with at most extremely limited exceptions). "
I'd theorize that one of the advantages of mail-in ballots is that corruption is more visible. That is, it's obvious on a small scale, and also more easily detectable if done on a large scale. As long as society maintains the ideal that the ballots are filled out and sealed in private, I'd hope that we'd know to keep an eye out for:
-- people offering to fill out ballots for you
-- people wanting to watch when you fill out your ballot
-- people collecting ballots to "deliver" them
I'm not 100% optimistic here. I've seen explicitly partisan organizations running booths on the street to help people fill out ballots. And I worry when I see people make claims that imply that civil democracy is a problem when it doesn't lead to the "correct" outcomes. But overall I'm happy with the mail-in ballots that my state has.
How is "corruption more visible" when its done behind closed doors? Your viewpoint was the common one for a century before the adoption of secret ballots based on the reality that human nature didn't live up to wishful thinking.
I'm guessing that you approve of the politics of the winners, therefore mail-in ballots are less of a concern. Except when the general public finally grasps that they an election based on them can't be trusted, the political turmoil of prior elections seems likely to be minor compared to what goes on before they are fixed.
I agree that this is one potential worry about ballots people fill out at home. It just doesn’t seem clear to me that it’s worth the much *bigger* worries of inconvenience provided by requiring people to go to a specific place at a specific time to cast a vote.
Inconvenience is a bigger worry than depriving people of their right to a vote due to being coerced? "Sorry domestic abuse victims, we don't care enough about elections to bother going the polls in person since its inconvenient, but despite that we think our ability to vote conveniently should deprive you of your right to case a secret ballot securely?"
Its a bigger concern than the reality that we can't know if the level of problematic votes is large enough to be changing the results of close elections and therefore we can't know who would have won a fair election? Inconvenience is more important than the potential that we could have widespread voter coercion and other problems arise just like they did before secret ballots existed?
Its unclear how people think we can magically know it won't grow just as it did back before secret ballots existed at any time. Social media didn't exist back then, trends can spread far faster today than back then. Human nature hasn't changed, so merely because problems aren't visible to the public yet the way they were back then, doesn't mean it can't happen quickly with nothing preventing it.
Yes. Inconvenience takes away millions of peoples rightful votes. That is a bigger concern than a few thousand people being inconvenienced by a spouse.
A few thousand? Seriously? We aren't turkey, but the number there was 20% estimated who had a their vote coerced by a family member, thats in the millions.
There are 80 million non-voters who don't care about their ballots, and millions of rabid partisans concerned the other side's candidate being "evil": is there some rational reason to expect it won't finally dawn on them to get the ballots from their friends and family, "hey, I'll buy lunch next time if you'd click this to get a ballot and sign it when it gets here for me.." or whatever. People break lots of rules when they don't expect to be caught since they aren't going to tell anyone. Can you prove it didn't happen?
Maybe it hasn't: can you prove why it won't? Why is it that people would assume if millions of students took an online home test without being watched they'd cheat, but magically all these people out there are honest when it comes to voting? Is it wishful thinking, or that because no one in the media talks about this that it can't possibly be logical since the media would have raised the issue? What is the logic and evidence regarding human nature that makes people think this way?
Sounds like another elaborate scam to restrict voting by those who already struggle with existing restrictions. The obvious intention is the exact opposite of fairness, and "think of the abused" is yet another convenient excuse.
Unfortunately many privileged people take for granted their ability to safely vote the way they wish in secret and ignore the plight of the powerless who can't do so safely at home or work, and ignore the way their votes are being suppressed on a massive scale.
Those who bother doing a little research and thinking can find ways to address the issues of people who struggle to get to the polls, as I noted in other comments. Other countries do so, as did this country in the past before people decided to engage in magical thinking that people are somehow 100% trustworthy at home and work and would never cheat, coerce or buy ballots: despite lots of history and evidence that isn't remotely the way humans behave except.
Voting issues is a standard right-wing talking point, just like race is a standard left-wing talking point, both camps are trying to use them for getting/staying in power. My guess is that you are from the right-wing camp, pretending to care about voting fairness, but really interested in your in-group's power.
"unfortunately those with poor critical thinking skills haven't questioned the assumptions the media makes about the issue, vs. basic logic and history"
Ah yes: "anyone who doesn't view this as I do has poor critical thinking skills."
I am hoping that you are not an eligible voter anywhere near where I live.
sorry, but the site (and some comments below) point out exactly the issue I was referring to but didn't specify that most of the American public has fallen for, and I figured rationalists would be concerned about. In a short post I decided not to put the details into that post, perhaps that was a mistake. If you disagree: feel free to make an actual logical argument, but I suspect the odds are you cannot. Its those who don't provide actual logical arguments (I linked to them even if I neglected to get into every detail in a post) that we should be concerned about.
And I see that my summary was, if anything, too kind.
Are you capable of any actual argument about the topic? Or do you just assume you are right because of what the mainstream media has told you and anyone who has a different viewpoint and must be apriori wrong?
I think secret balloting makes the system open fraud. Secret ballots assume the votes are entered as voted. How to check the assumption? With an audit. How can you audit secret ballots? As far as I can see, this is impossible.
Nowadays I think it would be better if everyone could verify their vote on a central web site, and have (possibly more than one) independent auditor organization verify the votes, possibly with something like double-entry bookkeeping. It would still be possible to display a different vote to the user than is recorded, but I would think that could easily be caught examining the system.
You check by counting the votes again to see if it's the same number, and by making sure there's multiple people looking at each ballot, no? It's anonymous and secret but there's a physical record in the form of a bit of paper.
This doesn’t give *me* any way to confirm that *my* vote was recorded as I intended. FWIW.
Sure, an individual couldn't do it but a relatively small group of volunteer observers could check that all the votes (therefore all of theirs) were counted correctly.
Tech folks are too focused on the tech issues rather than the human ones.
With a verifiable vote that repeats the exact same problems that existed before the secret ballot that lead to its adoptions. Verifiable votes allow coercion to ensure someone voted the right way. There is a reason that secret ballots are considered a human right globally, after people tried the approach without them and discovered the issues.
This is a valid point, but it is in conflict with the tabulation of votes. "Those who vote decide nothing. Those who count the votes decide everything." Apparently Joseph Stalin didn't say exactly that, but him and others through history have expressed similar sentiments.
EDIT: So I guess the question is which one people find more important? On the one hand, you can be coerced to vote the "right way". On the other, you only THINK your vote matters.
Well if you mention Stalin, then you must know that there were multiple instances where bolshevists demanded open voting instead of secret voting (I don't mean general election here).
There are two things necessary for a fair result to be determined: fair counting and fair casting. Its like the streetlight effect: the counting is in public view so everyone focuses on that and ignores what happens in the shadows where the problems may be.
The process need to be made credible that when a vote enters the system it is counted. I suppose if there were some way of separating out the existence of a vote virtual envelope in the system separate from its contents that is handled secretly by another process, there might be something possible. The big thing now is dealing with the issue being ignored in the shadows, everyone is at least thinking about the counting issue.
For physical votes, I think video surveillance outside the booth to make sure no ballots are added or changed might be good enough.
There's an actual problem for people who find it difficult or impossible to vote in person.
There's an actual problem with mail-fraud elections that's vastly worse than the problem of people who find it difficult or impossible to vote in person. No country but the US uses mail-in ballots, for that reason.
There's no particular problem with mail-fraud in elections, and mail-in votes are reasonably common at least in European countries.
You register to vote beforehand with all the necessary formalities, then at home you put your vote in an envelope, and put that envelope inside another envelope together with your registration confirmation. In the election office, they keep the confirmation documents and then pass on the anonymous letter with the vote to the normal vote-counting process. Sure, there's room for fraud here, but it's not as simple as pouring additional ballots into the bucket and claiming that these all were mail-in votes.
There are creative ways to address the minority that find it difficult to vote in person, like multiparty teams of volunteers that go to them turning their backs while they vote, of if they need assistance at least lowering the risk of one-sided coercion. Worst case a limited number of mail votes for when there is no other approach.
When those have been implemented, we can talk about restricting the right to vote by mail. At the moment, it's clearly a necessary part of undermining the deliberate Republican attempt to make it hard for blacks and other majority-Democratic groups to vote.
"A solution to the problem this is mitigating is theoretically possible in the future, so we can remove it here and now" is not a good argument.
I'm not so sure. If voting has systemic fraud, then surveillance of the voting itself won't accomplish anything. That's more likely to ensure voters don't commit fraud, which I think is less of a concern with the current system.
In terms of "systemic fraud": there were 80 million non voters last time who don't care to use their ballot. How many of them might be persuaded by a friend or relative to give them a signed empty ballot to use, if word spreads to do so? Studies show a substantial minority would sell their votes. A substantial minority of the populace is willing to use violence for political goals if needed, this would be a mild bending of the rules.
It seems that trust in the current system is based on naive wishful thinking about human nature.
That depends on what you mean by systemic fraud. What do you have in mind?
Domestic voter abuse is not the only problem our voting system has to solve, and mail-in ballots solve some very big problems - people who are too old or sick to get to the polls, people who are traveling out of state, people who don't have good access to transportation or can't find time to go to the polls, etc. etc. You're arguing that all of these people should be thrown under the bus to protect against a very specific form of voter fraud - domestic voter abuse.
But you don't actually offer evidence that domestic voter abuse is a large-scale problem. You argue that domestic abuse is a problem in general, so voter abuse *could* be a large-scale problem, and you argue that other forms of non-secret ballot have caused problems in the past, and then you make a huge flying leap to argue that domestic abuse is such a huge problem that mail-in voting should be abolished except for unspecified "extremely limited exceptions." I would like to see some better evidence before throwing all the access-to-democracy problems under the bus.
Also, fuck you for pre-emptively accusing everyone who disagrees with you of having poor critical thinking skills.
Other countries solve the problem of voting out of country by letting people vote at embassies, within the country there can be voting at other jurisdictions that provide secrecy. Volunteers can go in multiparty teams to invalids and turn backs while they vote, or at least minimize risk of one sided coercion. It noted that there can be limited exceptions if people are really incapable of inventing alternate solutions.
The site it links to goes through the data and evidence and logic you fail to note. It points out that most arguing in favor of mail voting are like those who fall for this:
"Imagine a government-created online test that 10 million students take from home with a $500 reward for passing. The tests are unsupervised. The program manager — who has the same mindset of those in the 1950s who ignored what happens behind closed curtains — proclaims, “Of course no one cheated! Where is your evidence someone cheated?”.
Would you accept that no cheating would occur? It sounds absurd, doesn’t it? The manager might argue for the students’ innocence until proven guilty. But this is only relevant to individuals. If someone accuses John of cheating on his test, John is innocent until proven guilty.
We can’t accuse specific students without evidence. But we can challenge the validity of the entire system. The inability to prove any specific student cheated without evidence doesn’t prove the validity of a claim that no one cheated—an assertion that defies reason"
Is there a *significant* amount of cheating? Enough to be a bigger problem than all the people we can be *sure* will not vote without mail-in ballots?
You're proposing to add rules that will reduce cheating, but also admitting that you don't know how much cheating has happened, or how much cheating these rules will prevent. "Nobody has ever cheated a mail-in ballot" is absurd, but so is "We should add new rules that will reduce cheating by an unknown and unmeasurable amount."
(I have a magic rock that reduces cheating by an unknown amount. I haven't checked if it works, but I think we should mail out a rock with every mail-in ballot, just in case it helps.)
Can you give me a ballpark figure, at least? How many ballots in the 2020 election do you think were changed due to abusive spouses? A million? A thousand? A hundred?
Priors about human nature and other data indicate there will be cheating, evidence from other sources suggest it could be 20%, or at least a few %, easily enough to tip elections (sources at that site TheBigIllusion.com), in addition to the non coercive cheating that can go on that there isn't yet data on.
Its like this statement: "Imagine a government-created online test that 10 million students take from home with a $500 reward for passing. The tests are unsupervised. The program manager — who has the same mindset of those in the 1950s who ignored what happens behind closed curtains — proclaims, “Of course no one cheated! Where is your evidence someone cheated?”."
Priors and human nature and other data suggest it not be trusted even if there is no concrete data on the frequency. It strains credulity to think otherwise and anyone who sided with the manager for claiming no cheating occurred would be viewed as cluelessly naive. Widespread mail ballots are new and voting happens once a year,or every few for major elections: it isn't easy to collect data on something everyone wishes to keep secret.
The burden of proof should be on those that pretend that somehow despite the history of problems before the rise of secret ballots, somehow human nature has changed and those problems are magically not guaranteed to happen again. Its like someone saying "the password to the bank login to handle their funds was posted online as 'password', but of course no one would ever use it so we don't need to change it". Priors and basic logic suggest otherwise.
Your proposal doesn't address the cases where
1) people are too infirm to go to the polls
2) don't have time on that particular day.
And embassies are rare, you may not even have one in your country of residence. Colleagues of mine can only vote in their countries' elections in the embassy and just going there and standing in line for an indefinite amount of time makes this a whole-day endeavour.
There is a difference between widespread voting by mail, vs. extremely limited cases. re: 1. I already noted .Volunteers can go in multiparty teams to invalids and turn backs while they vote, or at least minimize risk of one sided coercion
re: 2, there can be multiple days of in person voting. in the US there can be cross jurisdiction collaboration. Other countries do requite people out of the country to vote at consulates or embassies.
Edge cases can be addressed, without ignoring the lessons of history and pretending "no, it can't possibly happen again, people are 100% trustworthy". If some people are restricted, its better than having an election which is meaningless since it can't be trusted to represent what a fair vote of the public would come up with.
I can't understand why no one has tried to solve this with bitcoin yet. For a while now I've been expecting a small team of mathmoes to publish a system that uses cryptography this and blockchain that and allows everyone to publicly vote, check that someone else has voted exactly once, and tally up the votes thus confirming the results for themselves, all while keeping their actual voting choice secret. I would also expect the process, which guarantees openness and transparency in all future elections, to be so involved and complicated to carry out that only maths phds can actually do it. But I've not heard so much as a conversation gesturing in that direction. The Internet has really let me down here.
Twenty responses in, and nobody has invoked the obligatory XKCD? https://xkcd.com/2030/
That doesn't address the human factors issues at the root of the issue of voter coercion. Most people in forums like this can safely vote in secret themselves and ignore those who can't. Crypto tech doesn't solve this issue, privacy in person solves it. (at least temporarily, there are complications later, but one step at a time).
What voter coercion are you imagining here? I'm struggling to think of any way voters can be coerced as long as voter anonymity is guaranteed. Can you give an example of the kind of stuff you're thinking of?
Vote counting is the bigger part of the problem. In Putin's Russia, you don't have an election then wake up to find the secret police at your door because you voted wrong. You wake up to find that you and all your friends somehow accidentally voted for Putin again.
Huh? as the original post noted, domestic abuse is widespread, including coercion by domestic partners, and evidence shows its enough to tip elections. There are 80 million non voters who didn't bother to use their ballot, why wouldn't many of them give it to a family member or friend who is very political , even as a favor without coercion? Polls show a substantial minority of the public is willing to consider political violence these days, people who commit domestic violence are somehow going to draw the line at coercing a vote? Its unclear where people come up with this magical thinking about human nature that everyone can just safely vote at home without coercion there or at work. Merely because it doesn't happen to the sort of people who read and post here doesn't mean it doesn't happen.
There are such systems. I just wrote a comment elsewhere, moving it here:
I do not remember the exact details, but the idea was that in the voting room you choose one of N options, and the computer prints for you N+1 paper cards, one vote for each option, plus one extra vote for the option you actually want. You drop each of these votes in the box (it is verified that you dropped N+1 of them), and keep a receipt for each.
It is possible to count the votes, e.g. when you have 100 voters, and three options that got 120, 130, and 150 vote papers respectively, you know there is one fake vote from everyone for each option, so the numbers of actual votes are 20, 30, and 50.
All paper cards are in a publicly available database (the random id, who the vote was for), you can check that all your receipts are there. -- If they are not, you take *one* of the receipts that are not there, and complain about election fraud. One receipt does not show who you voted for, because there is one fake vote from everyone for every option.
This is the kind of thinking I was after! Where and when was this discussed? Did they proceed to the next logical step, which is building an online model, holding iterated toy elections, and inviting people to try breaking the system?
Sorry, I heard this years ago, do not remember where. :(
I've heard tech geeks talk about this. Nothing you key into a computer that is tied to your personal ID can be guaranteed to remain anonymous. There's always a chance that someone clever enough can figure out how to trace a user's input back to their ID. For now, blockchain and crypto can be used anonymously in very limited use cases; but there's no way to safeguard even those limited cases from future tech developments.
For voting, my Taiwanese friends say their country does it best. Everyone votes anonymously on paper, and counting is manual. The counter holds each vote up to a public camera so everyone can see the vote on TV. There's very little room for a counter to discard a ballot he doesn't like, run ballots through a tabulator multiple times, or tweak complex software to make it miscount or weigh votes differently.
How well does that scale? I saw some news articles about counties that tried switching to pure hand counts after 2020 and discovering it takes a lot of manpower - you need multiple counters so they can check each other, the counters get tired and make mistakes after thousands of ballots, etc. Machine tabulation followed by hand counts to spot-check seems like a reasonable middle ground.
Possibly we could alleviate these problems by not trying to get results finalized by election night. When does Taiwan announce election results?
Up to at least 47,567,752,, which is the size of the UK's electorate.
Yes, there are long delays... I agree that's a downside. On the upside, most of the multiple count-checkers are volunteers, at least in the US (not sure how they do it in Taiwan). Any party is allowed to have at least one challenger observing a count. Theoretically, a hand count would take no longer for a large county than a small one, assuming that the count is done on-site at each voting location, which would be staffed by its own counters and challengers. So for giant counties, there'd be stronger incentives for breaking up voting location districts into smaller & smaller units, to enable quicker counting at each. This used to work in the 1900s before tabulators, even with the US's increasing population over 200 years, so on paper at least it should work again.
I find this post misaimed. Most people are *already* in favour of the secret ballot. The load-bearing part of this argument is whether mail-in ballots can successfully remain anonymous or not; I think most people just assume yes. That's what you have to challenge if you disagree; not the importance or lack thereof of anonymity in itself.
That is the point its making, even if its misnamed, the issue of restoration of the secret ballot which most people haven't realized has been undermined. Most people can safely cast their vote secretly and misguidedly assume everyone else can to.
Sorry, its just that its frustrating seeing most of the public falling for the equivalent of this argument where the burden of proof is put in the wrong place (that site goes into it):
"Imagine a government-created online test that 10 million students take from home with a $500 reward for passing. The tests are unsupervised. The program manager — who has the same mindset of those in the 1950s who ignored what happens behind closed curtains — proclaims, “Of course no one cheated! Where is your evidence someone cheated?”.
Would you accept that no cheating would occur? It sounds absurd, doesn’t it? The manager might argue for the students’ innocence until proven guilty. But this is only relevant to individuals. If someone accuses John of cheating on his test, John is innocent until proven guilty.
We can’t accuse specific students without evidence. But we can challenge the validity of the entire system. The inability to prove any specific student cheated without evidence doesn’t prove the validity of a claim that no one cheated—an assertion that defies reason"
I suspect the vast vast majority of individuals do not value a vote at $500, which means there is far less incentive to "cheat" by taking someone else's vote, than in the scenario you are laying out.
Seriously, if this is happening at scale, why is there *no* instances of people saying "My spouse who abused me in X Y Z ways, and who I have now escaped from, also convinced me to vote the way (s)he wanted"?
Even without coercion: there are people that would pay $500 per vote to overturn an election, and like the drug trade black markets are hard to be aware of or stop.
In terms of the domestic violence issue, there are such stories, the site links to them and to data, limited as it is. Domestic violence is year round and yet that is difficult to study, vs. something happening once a year. People don't report actual violence, this is something they wouldn't even think to report.
This is like hearing from people in the 1950s who doubted domestic violence, human nature is a bit different than you seem to imagine.
A substantial minority of the populace is willing to consider political violence, are they really not going to coerce a vote? Again, this happened in our history and was widespread, and its unclear why anyone thinks human nature has magically changed to prevent it.
What would make you consider immigration to another country? One where they don't speak your native language? Would this answer change if you had kids?
I think it’s a bit odd to associate the national border with the linguistic border. For many people, one would be a border that the other would not. Moving between Toronto and Chicago is in many ways easier than moving between New York and San Juan, or Toronto and Quebec City.
In any case, I would far sooner move to a big cosmopolitan city nearly anywhere than to a small town without a cosmopolitan university! National borders are an issue, but they are easier to overcome than many cultural borders within a nation.
Well National border are very important from a bureaucratic standpoint. You can't just hop over the border and find a job and a place to live. You need a Visa, a work permit, etc.
The process to get those things can be long, expensive and uncertain - and those deter a lot of people from moving across national borders.
Math academic, so employment offer like Quiop, but I'd also have to be single (which I'm currently not)/have a partner willing to move. I'm confident in my lack of ability to mimic my mindset if I were a parent.
I'm an academic in the humanities, so my answer is "an employment offer." (I have kids.)
Really? how far would you move? Does your partner not working as well?
For example, would you move to Greece (Assuming you are not from Greece)?
Also, not so relevant to me :)
So far, I have moved to three different continents (other than the one where I was born) for education and employment. Since marrying and becoming a parent, I have lived in three countries (speaking three different languages, one of which was my own native language and the other is my partner's) on two continents.
As you note, the "two body problem" can be a serious issue for academic partners, although we have so far been lucky to be find jobs in the same cities. (Being able to do so is a major motivation for being willing to move countries.)
An oppressive and dictatorial regime in my home country, or its conquest by a foreign country. Have kids.
Same here. Would try to move even before that actually happened if the risk was high; just in case that the new government might close the borders immediately.
At this moment, I see no risk, but I insist that all family members have valid passports, even if we are not traveling anywhere. Because I assume that if something bad happened, the line to get a passport might suddenly become very long.
In the past: better living conditions, like better work / more money / better social life / living with my partner. Preferably to a country with a culture I have some affinity to.
Now: living with my partner.
With kids (currently don't have any): Social support network, and ease of living with kids. Probably wouldn't consider learning any really challenging languages for this.
Royalty and correlated riches.
Anything that would make this a rational idea would compel me to take up arms to fix whatever is going on in my native country, even if this were a mostly suicidal course.
I do have kids and they are now out of the house -- this answer might have been different when they were younger.
For a better life. Did that twice. Different languages. With kids. No regrets.
Really? were to? Did you immigrate with Kids?
Russia->Israel->Canada. Admittedly, it's been 3 decades+
Is that a common thing for Russian Israelis to do? I note that Putanumonit Jacob went Russia -> Israel -> USA
There's been lots of "book banning" recently, according to many headlines, although I can never fully grasp what "book banning" means according to the stories. It always seems to involve removing books from school libraries. Apparently there was a case in the 1970s which went all the way to the Supreme Court, which found for a group of students from Long Island who had sued their school, claiming that their civil rights had been violated by the removal of some books which had been deemed "obscene" or something by the school board. (I believe one of the books was Slaughterhouse-Five.)
But there has also been many cases in recent years where books deemed "offensive" have been removed from mandatory school reading lists. Those books tend to be old classics like Huckleberry Finn, which features the "n-word" prominently, which now offends people out of sheer spelling. Apparently removing those books from mandatory assignments isn't considered "book banning" because those books remain accessible to students in the school libraries.
I get that there is a difference between a book being required reading for a student vs. a book which is merely available in the library. However, I don't understand how the removal of books from a school library could violate a student's civil rights anymore than removing a book from a mandatory reading list does. Aren't school libraries rather choosy about which books they acquire in the first place? Do many school libraries stock old joke-books full of racist and homophobic jokes? Probably few do. So how is the removal of a book from a library qualitatively different from a refusal to stock a book in the first place?
I'm not in favor of actual book bans, which I consider to be when a government bans a book from publication or sale. But until every school is willing to stock Naked Lunch, The Anarchists Cookbook and Truly Tasteless Jokes: Volume II on its library shelves, it's hypocritical to call the removal of any book from its shelves "book banning".
Yeah, I think this is one of those "Neutral vs. Conservative" struggles, https://slatestarcodex.com/2017/05/01/neutral-vs-conservative-the-eternal-struggle/. Libraries are one of those institutions that have become very left-aligned, conservatives have noticed, and they're pushing back against it with the tools available to them, which are basically entirely "external" things like bans.
Both sides in practice are trying to control what the next generation reads (I think the Russell conjugation here is probably "you censor books, I protect children from harmful ideas"), but when a book is offensive to the left, they don't have to do anything as public and as gauche as banning it, they just don't shelve it. (Assuming it's too late to go harass the author on YA Twitter to make sure it isn't written in the first place)
Is one side worse than the other? Sure, probably! But the asymmetry of the situation makes it hard to make direct comparisons. The right-aligned stuff is necessarily out-in-the-open and reported on, while I think things like YA Twitter and the Dr. Seuss controversy are just the tip of the iceberg on the other side.
I work for an organization which advocates for freedom of speech in K-12 schools, so I am pretty up on this.
The Supreme Court case you reference was Board of Education, Island Trees Union Free School District No. 26 v. Pico, 457 U.S. 853 (1982), in which a plurality of the Court said that public schools cannot remove books from libraries because of hostility to the ideas expressed therein, but can remove them for other reasons, such as vulgarity. Decisions re what books to buy and what books to assign in class are different, in my view, because requiring a school to teach or buy a book that, for example, says "the free market is bad" limits their ability to send their preferred message that "the free market is good." Every dollar spent on the former, or time teaching the former, prevents the school from spending that dollar/time on the latter. In contrast, the presence on library book shelves of a book that says "the free market is bad" does not prevent the school from sending the message that "the free market is good," but rather merely competes with that message. Given longstanding recognition that 1) schools may try to inculcate certain values and ideas, but 2) that they cannot "prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion", W. Va. State Bd. of Educ. v. Barnette, 319 U.S. 624, 642 (1943), that distinction seems to me to make a lot of sense.
Note however that Pico was a plurality opinion, and hence is not technically binding, though many lower courts have adopted the reasoning of the plurality.
Also, you say that "here has also been many cases in recent years where books deemed "offensive" have been removed from mandatory school reading lists. Those books tend to be old classics like Huckleberry Finn, which features the "n-word" prominently, which now offends people out of sheer spelling." However, those cases are comparatively rare nowadays; over the last few years, the vast, vast majority of books that have been removed have been because of sexual material or profanity. See, eg, https://www.ala.org/advocacy/bbooks/frequentlychallengedbooks/top10 A distant second is what could very broadly be described as "CRT" or CRT-adjacent, though again that would be books in the curriculum, rather than libraries.
Finally, I certainly agree that using the term "book ban" to describe what is happening is not the height of intellectual honesty.
What would your take be on declining donations? Certainly there are space constraints in a library (they could not accept 10000 new books on communism for example), but with the promise of sufficient funding to expand/support their library services should they be obligated to accept any donation?
My view on removing books from school libraries is that as long as the book is allowed at school (if brought from home), it's not a hill Is be willing to die on. I think that parents probably have some degree of interest/rights in which books their children are allowed to read, and unlike at a public library they have less control at a school one. If a public library banned a book I'd be getting out my petitions.
Are you asking my legal take, or my policy take? My policy take is that the library should have as wide a range of ideas represented as possible, so they should accept the donation if it comes with funding to expand. My legal take is that, under current jurisprudence, the school is probably free to refuse the donation, just as a city can refuse the donation of a monument in a public park https://en.wikipedia.org/wiki/Pleasant_Grove_City_v._Summum
It is true that parents have some interest/rights in what their children can read, but the problems are 1) children have some rights independent of the rights of their parents; and 2) removing a book from a school library means that no child can read it, including children of parents who want their kids to have access. Which is why this is actually a rather complex issue. And suppose some woke school district wants to remove all books that support conservative ideas, or even, say, a complimentary biography of Ronald Reagan? Or all econ books by Milton Friedman, et al. I am sure the majority of parents would love that, but what about the parents who want their kids to have access to those ideas?
Thanks, I was curious about both. To your last point, this was my point about public libraries. To me the ideal solution would be that books would just be moved from the school library to a public library so that parents could still check it out for their students.
Re: students having rights. My middle school staged a peaceful protest, and I quickly learned that we indeed lived in a totalitarian state at school. I have since given up on the dream of students having rights; while I support other dreamers, I think that that dream is forever dead to me.
There are a number of fundamental problems when it comes to dealing with public education:
1. Children have different growth needs; some need a lot of firm guidance while others need the freedom to explore.
2. Parents have a massive investment in their children's future, and thus have a right and responsibility to act in their children's interest to shape their values and education.
3. Some parents are not good parents and do not act in their children's interests.
4. Most people don't have the resources necessary to chose an option other than public schools.
5. Public schools are taxpayer funded, and thus responsible to the governments which fund them.
6. While a good teacher will act in the interest of the students, teachers do not have the same level of investment in the children that parents do.
7. It's very hard for a bureaucracy to accurately tell who is a good parent or a good teacher.
8. Teachers and school administrators are not necessarily representative of the community at large.
9. Anything currently politically charged is that way because there are rational arguments on both sides.
10. Schools cannot teach everything. There is a limit to the educational time and space (such as the size of a school library).
Ultimately, if the school systems can't meet the needs of the population that is under-represented in education, the public will correct this via the legislature. Having the state legislature as a check on the school system is a bit brute force, but the legislature is elected by the public and the legislature's actions are mostly transparent. School boards are closer to the level of the public, but are much less transparent.
The long term fix would be some form of school choice.
A different but related question: how practical are these school censorship debates? With the Internet widely available, there are lots of older classics on standardebooks.org, for example. Also, lots of other Internet resources. Are school libraries still relevant or is this just an old-fashioned debate by people who don’t realize how much has changed?
I imagine that more practical barriers are not knowing stuff exists because adults don’t promote them and many kids being more interested in Minecraft and social media than reading books anyway. Also, maybe Internet access restrictions?
But what do I know, I’m old enough to remember carrying books around for entertainment like people nowadays carry a phone. It would be good to hear about it from people who were in school more recently.
School libraries are critical in that they are a space in which a child can access information without their parents being informed and having veto power over what can be learned. This makes them highly corrosive to any ideology which relies on information control to retain members, including most cults.
Moreover, many school assignments are given with the instruction that the student ought to consult either a specific book or their choice of books from the school library. If a student is told to research the history of the Spanish American War, but the school library cannot legally contain any books which discuss the history of American imperialism, the student is going to come away with a wildly incorrect picture of why the war was fought.
Perhaps you have very little familiarity with children or parenting?
As a parent you have fairly strong control over how and when your child can access the internet unless someone is actively invested in undermining that control.
Pretty easy to simply not get your kid a smartphone, or prevent a handful of social media apps from being installed, and thus cut off almost all influences you wouldn’t want.
Obviously this control attenuates with age, but by the time your child is a middle to late teenager, you’re probably less concerned with total control and more concerned about making sure that when inevitable encounters with the previously censored material occur, your child has already had enough context from you to make good choices about viewing it or fleeing from it.
The distinction is that the book bans being objected to forbid the school library from stocking the books even in the event that the school library wishes to do so. It's one thing if an individual school happens to not have any copies of the Quran; it would be another thing entirely if there is a legal ban prohibiting any school in the state from possessing a copy. Many of the laws (including Florida's HB7) are structured such that entire academic positions (ie. essentially any admission that systemic racism exists in modern day America) are banned from being expressed in any book in any Florida classroom or school library. That's obviously an unacceptable degree of censorship, well beyond a school simply declining to stock a particular book because they decide they don't want/need that particular title.
"(ie. essentially any admission that systemic racism exists in modern day America)"
Large numbers of Americans believe in similarly unfalsifiable claims like Jesus being the son of God but have (often grudgingly) accepted the fact that they aren't allowed to preach those ideas to a captive audience in public schools. You can too.
Wait - have they accepted the fact that it’s illegal for a school library to stock Bibles, just like it’s illegal for a school library to stock books about gay characters? I agree that teachers shouldn’t tell students to be gay or to be Christian, but to ban the book from even being in the *library* seems pretty extreme.
That is not a condition of the new Florida law. Its mindless keyword-ban makes no distinction between classrooms/libraries which present systematic racism in the manner that you describe vs those who present it as something to discuss/debate vs those who present it as a false claim.
In a society genuinely devoted to freedom of expression, the right response to shitty speech is more (different) speech. In that spirit the right lawmaking response to school systems presenting systematic racism as unarguable truth and forbidding the questioning of it, would be to insist that the subject be discussed/presented as one idea about which there are different viewpoints.
"In a society genuinely devoted to freedom of expression, the right response to shitty speech is more (different) speech. In that spirit the right lawmaking response to school systems presenting systematic racism as unarguable truth and forbidding the questioning of it, would be to insist that the subject be discussed/presented as one idea about which there are different viewpoints."
Or we decide there's only so much time in the school day and would rather spend it on other things.
Also, school discussions are not like discussions on this blog: students are forced by law to be there. They can't just go to another browser tab if they find that the "speech" isn't worth listening to.
Can someone point me to this keyword ban in the actual bill? (https://www.flsenate.gov/Session/Bill/2022/7/BillText/er/PDF)
The bill talks a lot about textbooks ("instructional materials") and curricula, but doesn't say much about library books.
That sounds like saying that the government should be allowed to pay people less than minimum wage.
More importantly, it’s saying that government has an important limitation, which is that it is not allowed to adopt any rules that empower its subdivisions to do things free from interference of higher levels.
We probably *should* treat employers and corporations as “private governments” that ought to be restricted by the constitution as well.
https://press.princeton.edu/books/hardcover/9780691176512/private-government
If you don't like it you can start your own tech company, but you can't start your own government.
Tech companies are not the government. They have no ability to compel or ban speech except on their property. No matter how big they get, you will never go to jail for violating Facebook's ToS. A government, on the other hand, can throw you in jail or fine you. They can compel your child to attend a school that only teaches what they approve of. So governments need far more scrutiny when they use their power to compel or ban speech.
I want to tell my friends something that insults a protected group ("the French stink"). Which of the following are acceptable?
A) I put messages in envelopes and the USPS will not send it.
B) I use bigger envelopes and FedEx / UPS will not send it.
C) I try to send it as a mail message and my ISP blocks it.
D) I try to put it on my webpage and my webhost takes down my site.
E) I try to put it on a cake and the bakery won't do business with me.
F) I try to pay for the above and my bank closes my account.
It's a book ban when the government declares that a specific book isn't allowed in a school library. Doesn't really matter whether this book previously was in the library or not.
If that's the case, then what's the word for the thing where the government declares that a specific book is not allowed in a private collection or on a privately-owned bookshore's shelves? Because that's something much worse than inventory management for school libraries being done by people you don't like, and I think we want to have a term that points specifically to the really bad thing.
Libraries exist, public and private. They have owners, and the generally accepted rule is that the owner gets to decide what books go in the library. If the functionary the owner hires to administer their library decides to stock the shelves with books the owner finds obscene or abhorrent, and the owner says "don't do that again or you're fired", I'd rather have a different term for that than "book banning".
The owner is the individual school. Not the state government. The state government is limiting the books that individual schools are allowed to acquire. There’s a reason that schools usually are run locally, rather than directly by the legislature.
The individual school is owned by the state government. Possibly by way of a local government, but those are wholly-owned subsidiaries of the state government. Public schools are not autonomous non-profit organizations.
They do complain: https://press.princeton.edu/books/hardcover/9780691176512/private-government
If corporations actually were as powerful as governments: had their own schools where people were mandated to study and the corporation decided what they are allowed to learn and what not, their own SWAT teams that would put you in solitary confinements or even execute you if you break user agreements and could do all the other stuff that goverment can - then yeah it would make all the sense to treat such corporations to the same standards.
But as for now we are talking about people being banned on a social media app and comparing it to a totalitarian state - this is just ridiculous. Don't get me wrong, social media bans for speech are still not a good thing. But they seem to be the least bad of current options.
RE governments: It matters what level of government one is discussing. Federal and State governments have broad powers. Cities, not so much. E.g. https://www.scstatehouse.gov/code/t05c007.php contains "The municipal governing body may fix fines and penalties for the violation of municipal ordinances and regulations not exceeding five hundred dollars or imprisonment not exceeding thirty days, or both." ( I don't know how often cities actually use the power to imprison over ordinance violations. I've mostly heard of fines. ) Come to think of it, my ISP has a bunch of weird charges it imposes, which are not terribly different from minor municipal fines.
Then it's not a ban, obviously.
The amount of authority school librarians are supposed to have is a separate question. There can be multiple different arrangements under civil society. And some of them can be more problematic than others. But it's a completely different level of problematic than outright govermantal ban.
There are categories of books that we know are de facto banned from public (government) schools because of the ideas in them. I'm thinking of the sorts of books that might get mentioned favorably in the comments on Steve Sailer's site.
Are those books banned? I don't see why only explicit bans should count as bans when the results of implicit bans are the exact same.
I agree. I think it's certainly worthwhile to have conversations on what takes up a school library's limited shelf space (and public libraries too), but I'm sick of the term "banning" being tossed around in this context.
"there have also been attempts made to pressure *private book stores*,"
I thought it wasn't censorship if it occurred through a "private company?"
"as well as law proposals to ban queer content in all-age media"
Where?
Wait, why is it incoherent to say that there should be a level of civic service that is immune to temporary fluctuations of elected officials? Most people believe that it’s a good thing, and President Garfield was even assassinated because he was part of the movement that supported this sort of good government movement limiting political control over broad-based government services.
Right. That seems to all count *against* removing books from libraries, because removing them is interfering with the choices of students and parents about what to read.
School libraries don’t typically have the power to fine and/or imprison citizens. Neither do universities. Most government agencies don’t.
What is your evidence these are the same groups of people other than that they may identify as liberal?
What is it a tell of? Usually the party making a claim needs to provide evidence to back it up. I was curious to know if trebuchet had actual evidence that the same people were doing both things or if it was just a general feeling they had. Well from their comment below it appears it was just a general feeling they have and they weren’t able to provide actual evidence.
Blog post on AGI - tl;dr I don't think the doom scenario is likely (mostly because, in my view, it rests on the orthogonality thesis, which is more of a series of assertions), but I think the alternative (a post-scarcity utopia) is almost equally scary.
Would love your thoughs!
https://arisc.medium.com/there-is-no-good-case-with-agi-792e4995ebb9
Sir, please consider me a friend. Agree with pretty much all of this especially orthogonality.
For your problems with utopia, we would basically all become aristocrats which has its dangers (i.e. we could all turn into very frivolous and vain people). But people always figure out something to do: there is no reason to climb a mountain, yet lots do it. I read a lot in spite of believing there is no such thing as a system that provides final answers. People still play chess professionally in spite of AI in theory obsoleting human chess players.
I expect even in post-scarcity there will be plenty of conflict of the interpersonal sort, as the attention and affection of others will remain a scarce resource (and no, AI can't fulfill that for most people, there is always the knowledge you are interacting with nothing when you interact with an AI).
Your thoughts on the orthogonality thesis deserve a wider audience in the rationalist community. But your mode of expression - which is combative and not charitable - will I think make this less likely.
Thank you (for the first sentence!). Re the second, you're right, I couldn't help it.
I would say it is worth recutting to be more persuasive to the rationalist community - they are a good audience! They would actively seek to engage with thoughtful criticism of orthogonality, politely expressed.
I don't want to come across as intransigent, but though I'd love to engage with the community (I'd love to see where I've gone wrong with my reasoning, if I have), I don't think I have the time to edit this - I have a few other posts I want to work on in the next few weeks.
Also, I'd suggest two things: first, I grant that I am unnecessarily combative, and parts of my posts are just my venting - but equally, rationalists' mode of expression _is_ almost purposefully incomprehensible, and turns people off. This needs to be addressed. Second, I hope that rationalists live up to their name and engage with arguments regardless of how they're expressed!
Also, I think the biggest problems with these sorts of criticisms is they don't offer an idea of what humans want. What do humans want for the future?
I think that's the crux of the issue - I want each individual to be able to think about this!
The fundamental issue is how we should engage with this ASI or ASI's. There is a scenario where humans use the super intelligent AI (ASI) to augment there own abilities. Instead of the AI doing the art for you, you create technologies which make you smarter or more creative and are able to create better art by yourself. This isn't limited to cognitive capabilities. You could intervene in human biology to modify temperamental dispositions. The general idea is to use ASI to improve human potential rather than recursively improving the ASI abilities. You bring up the issue of humans becoming addicted to virtual worlds or becoming bored or losing meaning. Well, we could use the ASI to develop tech which makes us more aversive to virtual experiences.
The second biggest problem here is that problems NEVER disappear. The development of AGI would likely raise new and profound questions about the nature of consciousness, intentionality, and mental states. Questions about the nature of time, causality, and the relationship between mind and matter would still be relevant.
The third problem of endless entertainment is a serious risk of causing a voluntary and unintended human extinction. The problem of these technologies being able to do art, entertainment, and other "humanities" work does not mean we should allow people to use them to do this. We could instead focus on fostering a culture that values personal development and real-world experiences, and setting boundaries for technology use. It's possible that this would lead to new cultures which voluntarily limit there engagement which the ASI.
A "deathless" world is not possible. The widespread use of advanced medical interventions would lead to new health problems, or technologies meant to extend life could be misused or weaponized, resulting in harm rather than benefit. Even with interventions from a superintelligence, it could be impossible to completely counteract the effects of entropy on biological systems, ultimately leading to deterioration and death.
We're living in a time where exploring new worlds and creating fresh cultures and ways of life mostly happens in the virtual or abstract realm. ASI comes in as a game-changer. We could use it to boost our creativity, speed up scientific breakthroughs, and offer innovative solutions to major global problems, ASI can kick off a whole new era of amazing possibilities and growth. It's like opening the doors to an even more exciting and adventurous human experience.
I think your imagination is a little constrained here! First off, any new questions that arise will be answered very quickly. That's what makes a deathless world possible.
I dunno. I think your imagination is too influenced by the idea of a being with great knowledge and power who also loves us. That's a story from Sunday school. lIt's not obvious to me that a superintelligence would love us, answer all our question quickly, make sure that new resources were distributed equably, etc. Superintelligence isn't going to be YOU, but with extra power, It's going to be more like an alien from a very advanced species. Stop sneaking helpfulness, fairness, love of our species etc. into our definition of superintelligent AI!
and it's not obvious there would be a single superintelligence. There would more likely be several, or many, competing with each other, with humanity cowering beneath, like sparrows hopping around below fighting eagles!
Also, for a superintelligence, doing away with humanity might not be a permanent irrevocable step, as we naturally assume. So this would mean less disincentive for it to make that decision.
If it had bottled DNA samples of ten thousand or a million humans, and all species of their intestinal bacteria, it could wipe us out any time it liked, secure in the knowledge it could recreate us, raising a new generation with synthetic robots indistinguishable from parents (as explored in several SF films).
So perhaps a superintelligence would rationally decide the World could do with a thousand year break from human depredations, to restore nature's abundance, before conjuring up Humanity Mark 2. Or maybe it would treat that as a trial run for an interstellar mission, where the same would be done on arrival, just as we have artificial Mars habitats in the desert on Earth as tests for the real thing!
It might even decide a "reset" like that would doing humanity a big favour in the long run, by eliminating the crushing weight of past human history and achievement, which would by then prevent anyone from ever making an original new thought, or theory, or song, or joke, etc, although I'm not sure quite how that would work if the new age is to be civilized, unless the superintelligence fabricated an elaborate pack of lies and distortions to account for the past.
So who knows, perhaps in a hundred thousand years there will be a human society where, unknown to them, the Evolutionists and the Creationists are both right! :-)
Yes, but some questions can't be answered even by a superintelligent AI. For example, the halting problem. There are an entire space of problems which are computationally irreducible. And if any of those problems are related to achieving mortality reduction then we end up in a world of death.
Fair enough, but I doubt that's the case.
Interesting read. I share your skepticism about the "utopia" outcome. The thought of scientific progress being made without any human understanding of that progress is a deeply depressing idea to me.
Side issue, but: MRI is definitely not telepathy. It's a very powerful magnet at very close range that can infer information about:
1. The location of different types of tissue.
2. Local changes in blood oxygen levels.
The examples of "mind reading" with MRI are very much party tricks that we shouldn't expect to scale much beyond what we already have.
Haven't read the article, but agree. I think people need work-- not backbreaking work, but work. I'm remembering something a depressed retired lawyer said: "People used to pay a lot of money to know what I thought. Now nobody cares what I think." In any case, I find it very hard to believe that everybody's going to have access to this utopia of no work, incredibly effective medical care, wild and rich forms of entertainment, and immortality. I think it's going to be the same old shit: The rich will have access to that stuff. I expect us to end up in something like the worlds in the earlier Gibson novels, with some clans so rich they own, like, the whole fucking moon, and most other people living in a scrappy world of urban sprawl, without much nature left, with access to various bits of tech that sound impressive to us, but that do not free them from being poor, lonesome, anxious, randos wandering through the sprawl.
I'm listening these days to a lot of podcasts, some of which are interviews with people developing novel uses of AI. There is now up and running a kind of AI that duplicates personality -- which in practice I suppose means behavioral tendencies, attitudes, mannerisms, and of course appearance and speech patterns. So you can pick some famous person to become your AI friend -- for instance Oprah, or various characters movies and video game. You then can access your AI friend whenever you like, and chat with them when you're lonely. Oh good, that's going to do a lot to help chronic isolation and the general fraying of the fabric of society! But that's not the worst. If you're wealthy and have speech and video samples of a deceased loved one, the company will make you an AI version, and you can then talk with your dead son, dead wife, dead mother or dead father, who will look at you with some approximation of the familiar twinkle in their eye, express some approximation of their familiar points of view using some approximation of their familiar speech cadences, curses, jokes, opinions, etc. Both of my parents are dead, and I really miss my mother and would love to have an hour to talk some things over with her. But having AI Mumsy on tap doesn't fill the bill. Jesus fucking Christ, I think it's obscene! Saw an article one time about somebody whose Jack Russell Terrior had died, so he had it taxidermied in a perky pose and it Dead Perky Dog was part of the living room decor. FML.
The guy describing the choose-a-personality AI was young, smart as hell, amiable and reasonable. But he seemsed utterly oblivious to the downside of the service he was offering. the peo[le who are developing AI are going to have an outsize influence on world events. I'm not sure it's an exaggeration to say they're going to run the world. They're not awful people. But they all have what a friend of mine calls engineer's disease. There's a lot they don't grasp about the complexities of being human and about what affects quality of life. They are kind in giving in that they truly want to produce great tech to help the world, but their ability to empathize and to grasp points of view different from their own is very limited.
I'm a recent (December 2020) retiree myself, but, unlike your depressed lawyer, I find that change purely positive. I do *NOT* want an endless series of bug reports to dominate my life. I am much happier choosing how *I* want to spend my time, thank you very much.
Catsambas's "but scarcity, death, and overcoming adversity are things that have defined and shaped humanity forever: they are the things that give our lives meaning." is, as far as I can tell, completely wrong-headed (or, alternatively, "meaning" is vastly overrated). One doesn't need any of those things to enjoy a sunset, or any of the other pleasures or comforts of life.
Yeah, I get that nobody needs an endless series of bug reports, but many people need to have things they're striving for, or tasks where they get appreciation from others (I think that's what the lawyer missed). I, for instance, never take vacations with a large lie-around-and-relax component. They make me feel aimless and blank. I like to do outdoor activities trips -- sea kayaking, hiking, etc. -- or go to NYC with a big itinerary of theater, etc. set up in advance. Do you do much sunset-watching, or do you have projects you're working on?
I watch sunsets, web surf, and grow crystals. I wouldn't call any of these projects, certainly not in the sense of "if I do X, and Y, then Z will be available and I can use Z and P to get Q, which is the point of it all"
'Any' of the other pleasures? What about everything else I wrote about, e.g. art and many forms of entertainment?
Entertainment generally does not need "scarcity, death, and overcoming adversity". If we actually got a friendly ASI which removed many of these constraints (scarcity never completely goes away - if nothing else, the observable universe has a finite amount of matter) some forms of art would become less relevant, but I consider that a small loss compared to the gains from removing many of the constraints on us.
Hypothetically, dead loved ones is just a starting point. The further utopia/dystopia is that infants start out with exposed to ai "people" who who are optimized for the infants, and later, friends as well as family.
Again, hypothetically, but good parents and friends learn and change. The better ai personalities should also do that.
Infants recognize their mother's skin by smell. AI is odorless. Though of course you can rub have some chemicals on the fucker and it will have B.O. too.
Nancy, how do you optimize AI people for infants? What is it are you trying to optimize? Are you trying to make them as calm, happy,neutral and smart as possible? How do we know that the parent's discontent with life as it is isn't useful the the infant, even if sensing that makes the infant less calm and happy? Think of the public figures you admire. Do you think most of them developed without absorbing some of their parent's frustration and discontent, along with their body odor? Do you think they would be better off with no deep connection with the smell of their parent's frustration and grief? And he parents -- do you think they're better off with no child to whom they tell the story pf the crap they've been through and what it's taught them? Are they better off being just egg and sperm donors?
I'm assuming a tech level which can also make good enough androids, including pheromones, so that automated childcare works.
Optimize for infants probably means calm and steady, with enthusiasm but not overloading the kid.
*Do* children need somewhat stressed parents? Maybe stressed enough but not too stressed? I assume that something reasonable is possible.
How about pretty good simulated parents for small children when the parents can't take care (too sick, perhaps?).
I'm assuming more that simulated people as an important part of life is something that people might try, and many might find that they prefer, rather than something that's absolutely better than the current world.
I think you're wrong to assume only the rich will have access to this kind of utopia - why would they? When there's literally near-infinite resources, it'll be impossible to hoard them - it'd only take one person to break ranks and share the wealth around.
"When there's literally near-infinite resources, it'll be impossible to hoard them - it'd only take one person to break ranks and share the wealth around."
Do we see Bezos walking around handing out thousand-dollar bills to lucky random strangers? The rich don't get rich by handing out money; they get rich and stay rich by penny-pinching, by making sure they know what is spent, where it's going, and what return they're getting for it. Right now, how many Indian peasant farmers are getting the wealth shared around by one person breaking ranks? What they're getting instead is "we'll back a WHO programme to sterilise you and your wife".
Why would a super-rich person hand over any of their wealth for no reason? Even the philanthropic get rewarded for it by way of tax breaks, fancy parties, their names on foundations, etc. Super-Bezos or Super-Zuckerberg can be rich, healthy, nearly-immortal and have all they desire, they don't need ordinary Joe from Oklahoma. If they want servants or company, they can get the AI to create better-than-human robots for that.
I really don't think you get what a post-scarcity world looks like. If Bezos gives $1k away (which he does, btw), that's $1k less for him. Post-scarcity, there's no longer a limited supply of resources to be divided.
Well. USA wealth is might be considered near infinite, depending on how strict your definition of near-infinite is. There sure as hell is a lot of it. If the money was distributed equally, every citizen would receive $3 or $4 hundred thousand dollars. But ain't nobody breaking ranks.
You might also consider the present situation, in which the tech companies have a tremendous amount of wealth and power. Congress does little to regulate them, and they are developing a technology that's going to change everything. In many ways, they are in charge of what happens in the world over the next century. What reason is there not to believe they are not going to hold on to a great deal of their wealth, power,secrets and access to extraordinary technology?
If you imagine that super-intelligent AI is basically God (except made out of shitty intricate little wires, etc., put together by low-empathy engineer type people) then I guess GodAi would come down and say "Let there be fairness!" But if we only assume AI is going to be god*like* in its knowledge and power then I think we all have to throw away our Sunday school versions of God and just think about a being of infinite knowledge and power, and drop all the frills from the old story about how it loves us blah blah blah. Honestly, why should it? Or maybe it will, but the way some people love their pets "Aw, look at his widdle face. He's mad because I didn't give him another treat, so cute, haha." Or maybe we will look like doodle bugs to it. Or sushi. Or turds.
I made an RPG setting inspired by Archipelago, all done except for pictures. It uses Traveller rules and has a 3D star map. Any TTRPG fans interested in playtesting?
I realized something about ambidexterity -- Scott used it as a baseline to compare with bisexuality etc. in recent posts.
Everyone uses their "off-hand" for some things. Forceful actions in one-hand, fine-dexterity in the other. Or some split.
Maybe there's semantic differences in how people understand and self-identify as "Ambidextrous". I myself write with my left hand but do plenty of things right-hand-dominant... do some people call that ambidextrous? And are certain types of people more likely to classify themselves as such?
Makes me wonder about musical instrument design. Guitar: Why pluck strings with right hand and fret the strings with the left? There are left handed guitars also though, so probably meaningful difference for hand dominance. Piano: Why are the bass notes on the left of the keyboard and the higher notes on the right? Are there any reversed pianos for lefties?
I do things with my left hand out of habit and attempt to balance myself, but my right hand is strictly better at doing those things too.
> Everyone uses their "off-hand" for some things. Forceful actions in one-hand, fine-dexterity in the other. Or some split.
This hasn't been my experience. My left hand is both stronger and more dexterous than my right. When I use my right hand it's usually because the task requires both hands, or because I'm using something that was designed for the right hand.
> I myself write with my left hand but do plenty of things right-hand-dominant... do some people call that ambidextrous?
"ambidextrous" usually means that you can do those things with either hand at least competently -- so unless you can also write OK with you right hand and switch-hit at baseball, batting right and writing left would not typically be considered ambidextrous.
Ah, the Lobster
"opinion on human challenge trials"
FWIW, Tom Scott had a nice recent short video on what a challenge trial is like from (approximately) a participant's viewpoint: https://www.youtube.com/watch?v=1v8u3ua6BPk "The people who get paid to get sick"
Would love to share my substack on culture in Japan and how it's taken on by the West. Writing from living between Japan and Australia
https://hiddenjapan.substack.com/
Wanted to see if I can assess people’s feelings about free will in a way that doesn’t degenerate into “well what do you mean by self, choice, natural, and the universe?”
Say you have a computer powerful enough to simulate the entire universe you’re in and you can look into it and see yourself a few seconds in advance. I suspect this isn’t possible but we’ll take it as a given.
Two questions:
Must anything be true of your feelings about the actions the computer shows you performing in order for them to come true?
Would your actions in this universe with the computer be any different than your actions in the universe without the computer?
Your hypothetical implies that the computer can solve the Halting Problem, which it can't. The computer can't possibly simulate the entire universe, including the effect that the computer has on the universe.
a) I'm not entirely sure about your meaning here. For me to do the action the computer shows me doing, it would probably have to be something I'd be fine with doing already. There's only a few seconds involved, so there's not much time for serious thought. I would hate to get in the habit of doing whatever it predicted, simply because I saw it. That's distinctly different from doing the thing it predicted without having seen it, because there's an extra bit of causality in the middle.
b) Is the simulation a simulation of a universe with a computer simulating a universe with a computer, etc., ad infinitum? Sorry, this is probably not an answer you wanted. :-) But practically, for me, I would absolutely test this out, almost the first thing. Do I raise my right hand or my left? Whatever the simulation shows, I'd do the other way. If that crashes the universe, so be it.
For A I’m right there with you. I think knowing you were going to see the output changes the output it can give you. And for B, I think the futures where you see the output is different than if you hadn’t because you took action based upon it.
Agreed that irl this probably causes an infinite recursive loop that crashes the whole thing. Also probably a lot of these involve people just choosing to look away or turn off the machine.
It would be amusing if I looked at the computer, saw myself raising a right hand, raised my left, and then looked closer to see that the simulation inside the simulation had me raising my left hand, and the simulation inside that one had me raising my right hand, and inside that my left, and inside that my right.
Try to do the YMCA.
Along similar lines, I came up with this thought experiment: suppose you and a functionally omniscient AI each have a deck of cards. You play a simple game where you and the AI both secretly select a card and then play them at the same time, 52 times until you exhaust both decks. The AI’s goal is to always play the same card as you. At the start of the game the AI pre-arranges their deck having already determined the exact order you’re going to play your cards in, so that each round the AI is simply flipping the next card off the stack. Without fail, the AI matches you card for card every time.
Now imagine in the middle of one of these games, the AI picks up the next card on the stack and shows you what it’s going to be before the next round. Would you still pick that card anyway? Personally, I wouldn’t. There does not exist any order the AI can arrange the cards that can be correct. Which means that even if events in the universe are deterministic, they can’t be determined.
So to answer your first question, in order for the computer’s predictions to come true it has to predict things that you want to do despite knowing you are predicted to do them. If the desire to be unpredictable ever outweighs the desire to take the predicted action, the computer will fail. So whether or not the computer will work comes down to whether or not you want free will.
Beautiful thought experiment.
As to the card flip, I imagine this is something like the theological position that God is in. I use God here in whatever the least offensive manner is you find but basically all knowing all powerful source of creation stuff.
From the inside of our lives we have free will. We see the future darkly and react to how we think it will be. From the outside it’s all atoms moving by laws. So does God show you that card knowing that you’ve accepted He is all knowing and your choice to do what He has revealed is your sign of submission? Or does God choose the card you won’t choose knowing that it will make you feel from the inside that it was your choice even if it was all destined?
I have had some thoughts about what you want to see influencing the mirror based on your fundamental beliefs about what you should see but your example is much better for this piece.
Would you be okay if I attributed this to you in a write up?
"Would you be okay if I attributed this to you in a write up?"
Sure, go right ahead
...isn't this just The Stanley Parable?
Yes and probably.
Not heard of it before but based on wiki, kind of?
I think the hope I had was that we could have a discussion about what you’d actually see on the computer screen and how that would have to relate to your feelings about it.
Also like your username. Triggered my secret name seeking instincts.
Well, if I personally were looking at it I would likely see myself doing nothing, and it would be accurate.
Once you get a mobile version of this computer, you see yourself running into a deer or a pedestrian, and so you slow down or swerve to avoid the outcome, because you feel strongly that it's to be avoided and know action has to be taken to avoid it. Weaker versions include going to a shop that can't sell something right now; you just won't go to that shop.
That doesn't really address determinism or free will, though. My take on free will is it's demanded by any universe that isn't a closed box. A prediction computer doesn't sway that one way or the other; predicting one box means it's another thing in the larger box, if there is a larger box.
I think in the case of you swerving to avoid the deer it would have to show you swerving and avoiding the deer and make you understand in the image through pure coincidence why you are doing that even if you start the swerve before you see the deer.
NasalJack had a great experiment on it above. I think the all powerful AI dealer showing you the card first knows you well enough to know you will flip over the same card even if you know it wants you to do so.
Determinism needs to be distinguished from predictability. A universe that unfolds deterministically is a universe that can be predicted by an omniscient being which can both capture a snapshot of all the causally relevant events, and have a perfect knowledge of the laws of physics.
The existence of such a predictor, known as a Laplace's demon is not a prerequisite for the actual existence of determinism, it is just a way of explaining the concept. It is not contradictory to assert that the universe is deterministic but unpredictable. But there is a relationship between determinism and predictability: predictability is the main evidence for determinism.
Nonetheless, determinism itself is the crux, in arguments for free will, and predictability only features indirectly as evidence for it. Determinism is the crux, because it removes the ability to have done otherwise, which seems to be important for moral responsibility; and also removes the ability to shape the future with present choices.
I'm not sure this scenario is actually intelligible, even if you swap out the computer for a genie who can magically show you yourself -- and your thoughts? -- a few seconds into the future. Nor do I understand what sort of answer you might be looking for with your first question. For the 2nd question though, I guess there are three possibilities:
1. You freeze and gaze at your future (frozen) self forever like Narcissus. I don't think this tells you much about free will except that even if you did have it prior to meeting the genie then you don't have it any more.
2. You do the exact same actions -- and have the exact same thoughts -- as the genie is showing you. If this is what happens -- like an infinite mirror except that it's driven by some layer further down than the original 'you' at the top (now reduced to a helpless, powerless marionette) -- then I guess it would be a strong piece of evidence against free will. But I don't think this is even intelligible, never mind possible, because you'd be standing on top of an infinite turtles-all-the-way-down recursion, while somehow not even seeing that or being aware of it. You'd have to see your future self in the infinite mirror and also *not* see you future self in the mirror but only your future self sans mirror. I guess you might be able to manage a flat out contradiction like that while eg in a dream state, since dreams can accommodate contradictions without breaking stride, but now we're talking about free will in dreams -- and we seem to have moved even further away from the original problem.
3. You do something different from what the genie is showing you. Well then, boom, free will I guess. Unless of course it's just a crap genie and the real genie who *could* have made this work is having the day off.
The problem is that I don't see any way of deciding between these three options. They all seem just about equally impossible, and if #1 and #3 are at least coherently imaginable, that doesn't make them any less impossible. (Like I can coherently imagine what it might be like to go back in time; but ti's still impossible for me to actually do so.)
So I'm not sure what this thought experiment was supposed to achieve? Did you have any particular answers to your questions in mind?
(Fwiw I think that 'free will' is just a nice story we tell ourselves. All that's really going on is that we do something and then we tell ourselves that we decided to do it.)
So yes probably the computer isn’t possible that’s why we have to take it as a given.
What you do in the future depends on your understanding of the future. I’ve never met someone into determinism who doesn’t believe there’s a difference between voluntary and involuntary action even if you nest that into a deterministic framework.
Because that’s a given element of the universe already, when you look at the computer screen to see what happens in the future it has to show you something you would always choose to do anyway even knowing that it’s coming in advance so that it can remain deterministic. Yes, in real life, nested simulations all the way down and a computer can’t support it.
This future is different than the futures without the computer because your knowledge of the future caused you to behave differently.
So however you define free will or determinism, that’s my idea of what happens.
There is clearly free will, if you presupposed MWI. It simply means choosing which Everett branches of the multiverse you wish to inhabit. Philosophies that disavow the meaningfulness of free will often presuppose a classical universe that is fully deterministic. Under MWI this is incorrect.
That's not how MWI works -- you don't have a non -physical soul that switches tracks between universes, you have a physical brain versions of which exist in every universe.
This is what I mean by definition games. I think you have free will in even a single universe that is fully deterministic because I don’t think those two things are at odds.
I do;'t see how that can fail to depend crucially on definitions.
I’m a compatabalist, or whoever it is spelled.
No, you got it wrong. You don't get to choose which branch you inhabit, or wish to. You inhabit all of them, unless you die in some. If you actually read what MWI proponents say about the issue (Sean Carroll is highly recommended), you will see that MWI has no bearing on the issue of free will whatsoever.
I should be more precise. You will no longer identify with those versions of you that choose otherwise, just as you don’t identify with those alternate versions of you who won the lottery or suffered a crippling car crash. While a version of you inhabits all branches, you only really care about the ones you identify with. Hence each choice is functionally the same as choosing which branch “you” wish to inhabit. I’ll look into SC though.
I guess my point is that you don't choose to inhabit a specific branch, causality goes the other way, sort of. The universe evolves deterministically according to the Schrodinger equation, including what can be emergently described as many-worlds branching, and it feels to the emergent you in each emergent branch like you made a choice to do something. For a hypothetical Laplace's demon (who knows the microstate of the universe and can watch it evolve) these emergent macroscopic concepts would be superfluous.
It's unclear from your description whether the simulated version of you also sees a prediction.
If your simulation doesn't see a prediction, then there's an important difference between you and the simulation of you, so it shouldn't be surprising if they behave differently. Even a fully-deterministic computer program can produce different results if you feed it different inputs.
If your simulation does see a prediction, then that prediction can't be generated simply by simulating physics, because now the computer is trying to predict its own output. Nothing can run a fully-detailed simulation of something more complicated than itself. So where did it come from?
You could "guess and check" by picking some prediction, simulating a version of you that sees that prediction, and then watching to see if the simulation fulfills the prediction it just saw. This is called a "fixed point" of a function; that is, an input that causes the output to be the same as the input. But there is no general rule saying that a function has to HAVE a fixed point; something as simple as f(x) = x+1 has none. So even if your computer can accurately predict what you'd do in response to seeing any given prediction, there might not be any where "what you'd do in response" is the same as the prediction you saw.
(It's also possible for a function to have more than one fixed point; i.e. there could be multiple prophecies that would all be self-fulfilling if you saw them. Neither existence nor uniqueness is guaranteed.)
All of which is to say that your thought-experiment isn't well-defined and may contain a logical contradiction, which would make any results you derive from it useless.
Totally agreed with you on why the computer probably can’t exist in the first place.
I don’t believe that makes the thought experiment useless.
Physically impossible is different from logically impossible. Logically impossible thought experiments are useless because, from a contradiction, you can prove anything. You therefore can't trust that the results are true or even internally consistent.
You also...don't seem to have answered any of my implied questions? You didn't say whether the simulated version of you is supposed to see its own prediction, and if so, what prediction it ought to see. I only said your experiment MAY contain a logical contradiction, but if you want me to reason about it further, you at least need to resolve the ambiguities I point out.
Supposing that the computer really has magically predicted the future, like a spice-vision from Dune or something, I think what it would mean is that when the time comes to take those actions, I would want to do them in the same sense as I do any other action I take.
My actions in the universe with the computer or spice-vision would be different because in the universe with the computer or spice-vision I looked at the computer or had the spice-vision, and that is an action.
Would you carry them out differently if they led to unwanted consequences? Could you?
If your understanding of the future is a cause of your present action then no because the computer must already take this into account in what it shows to you. It can’t show you something you don’t want to do.
In a truly deterministic universe, no. Physical determinism is not really compatible with "voluntarist" free will, where the actions you take cannot be explained by external factors (the corresponding physical process, in this case).
For free will to exist in a fully explicable universe, it would have to be on a more "intellectualist" model.
Isn’t the fact you see yourself in the future an external factor?
That’s the layer that the contradiction is at. If the mirror ever gave you cause to behave differently than it couldn’t be deterministic so because we know there’s a difference between voluntary and involuntary action it has to show you things that you would choose to do.
LarryBirdsMoustache, much like the man to whose face you are attached, you got it in one shot..
The universe can be deterministic, but since what you do depends on your mental model of what the future looks like, it would create a recursive loop where it can only show you future states that you want to enter.
It's been a while since I cracked them open, but I'll differ on this.
In a sense, that's certainly the subjective experience of the prescient - that they only have free will in areas that they haven't predicted. ("Absolute prediction is absolute death.") But they also do have choices - they can see many paths in the future and take actions that bring about their chosen future. And they have difficulty predicting random behavior, like sandstorms, or split-second decisions in a knife fight, or decisions based on tarot readings.
Part of the problem is that uncertainty and ignorance prevent ethics from dominating all choices, and once they **know** what the outcomes are, they're ethically obligated to take the actions that lead to the best outcomes. (See also, some of the psychological problems that people who get deeply into Effective Altruism have to deal with.) And there's a Great Filter in their future, and the actions that bypass it are horrifying, which leads Paul to attempt suicide-by-desert as a way to pass the buck to the next generation. It's a bit like if, in "Groundhog Day", Bill Murray's character felt obligated to constantly, repetitively, take the actions that maximized welfare for the people around him. No piano lessons, he could be saving lives instead. (The same life, over and over, every day.)
That's how I see it, anyway.
I think it's quite ironic that by trying to make feelings about free will more assessible you came up with a logical contradiction.
is "free will" in any meaningful sense still an open question?
thought folks nowadays are AI-optimists and predicting human or above human level silicon-run entities within decades.. do folks with those beliefs still believe "free will"?
Why would the existence of AI imply the non existence of FW? Are you defining FW as supernatural?
As said in the comment, it really depends on what you mean by free.
Do I think we can just change the world around us with a thought and break reality itself into a thousand pieces and that if we can’t then we live in a deterministic hellscape? No.
As always, it depends.
For me the question is solved to my satisfaction. Humans have free will in the sense of desicion making ability and do not have free will in the sense of transcending causality, regardless of determinism/indeterminism of our universe. And yes I do expect human level silicon-run entities within decades and I do expect them to have similar to humans free will. Though, calling such position "AI-optimism" gives me a sad chuckle.
hmm I'm confused, you seem to be saying "humans have free will" even if we assume for the sake of an argument the deterministic universe.
The universe is only deterministic if you assume there is a single universe which is classical in nature. But our best theories indicate reality is quantum in nature. Free will is just picking your branch!
I was pretty convinced several years ago by the series on free will on the Less Wrong site that the very concept of "free will" doesn't make sense.
https://www.lesswrong.com/tag/free-will
Yes. Actually, I mostly understand the concept from the lense of determinism. The less purely random stuff is going on in the universe - the better is our ability to predict possible outcomes and thus execute our decision making ability. So I'd say we are have more free will in deterministic universe.
It is logically impossible for a computer to simulate the whole universe. If it did, it would have to simulate itself, but then it would have to simulate itself simulating itself, however then it would have to simulate itself simulating itself... This lead to an infinite regress. How can any computational process within the universe simulate the very universe that gives rise to it? This seems logically contradictory and physically impossible. The simulator cannot simulate itself and create itself.
Agreed. Have to take it is as a given for that reason.
I would definitely say that your actions contingent on seeing a prediction of what you'll do will often be very different from your actions in another world where everything is the same except you don't see the prediction, if that's what you're asking.
Part of it! Other half is that the computer to remain deterministic would only be able to show you things that you wanted to do anyway.
I’ve noticed a lot of rationalists expressing anxiety or fear at the thought of AGI ruin. This doesn’t make sense to me, as someone who accepts MWI and who would have thought most rationalists do too. Am posting this in case it helps anyone suffering such anxiety or fear, but keen also for feedback.
TL;DR
Given (i) MWI, and (ii) that AGI ruin occurs with less than P = 1.0, then by quantum suicide/immortality logic, none of us should fear personally experiencing AGI ruin.
The key reasoning is as follows:
MWI remains our civilisation's best theory of the nature of reality.
Under MWI, since all phenomena are quantum mechanical in nature, it follows that every phenomenon that can occur under the laws of physics does occur in some part of the multiverse, ie in some universe.
Any phenomenon that involves obliteration of your consciousness, while it does occur in some measure of universes, is not something that can be subjectively experienced by you. You should therefore never expect to subjectively experience death.
AGI ruin entails your - and everyone's - death (AGI ruin scenarios that follow from premises such as the orthogonality thesis and instrumental convergence involve inadvertent disassembly of all humans).
AGI ruin occurs with less than P = 1.0. Or rather does not occur in all universes.
In universes where AGI ruin occurs, no humans will be there to experience it. AGI ruin scenarios do not readily admit of a situation where humans should continue to survive in a state of severe disadvantage or discomfort.
You should therefore not expect to subjectively experience AGI ruin. This alone should be sufficient to alleviate any personal fear or anxiety one might feel at the prospect of AGI ruin (whatever your views are on the measure of universes experiencing AGI ruin).
This reasoning is not new, and is I think a trivial extension of quantum suicide/quantum immortality thought experiments - so long as you accept that QM describes reality.
So what do you think. Should this alleviate a rationalist’s concerns? Is this an idea worth spreading?
Incidentally, should there be, for example, a moratorium against development of advanced AI, many of us should expect to personally experience living in a world deprived of the benefits of advanced AI - and to suffer from experiencing any ill that might otherwise have been averted by advanced AI. For those who place more weight on anticipated subjective experience, rather than, for example, negative outcomes in a measure of universes that will never be subjectively experienced by anyone, a moratorium would be dispreferred - which may justify a continued tech accelerationist outlook.
You seem to have rediscovered the quantum immortality argument. Go read the original paper on it, by the philosopher David Lewis, and reflect on what it really implies about your future experiences.
>MWI remains our civilisation's best theory of the nature of reality.
This isn't true?
I don't understand some of this argument, so maybe you could point to places where it's discussed?
For one thing, it seems to privilege consciousness as the fundamental of the universe. I don't see why that has to be the case; it seems like wishful thinking at best. Is this because of some technical argument about observers?
For another, we might not technically experience "death", but surely we can experience "dying"? If I experience dying horribly but then wake up in another me in another branch, then I'd believe this theory, but I haven't experienced that, and neither has anyone else that I'm aware of. If the argument is that we won't experience the dying, either, then wouldn't that percolate backward along our timeline, and the only timelines we experience are ones where we're immortal? (Leading to the conclusion that anyone who dies didn't have consciousness beforehand, leading to the conclusion that anyone you manage to successfully kill didn't have consciousness beforehand.)
Although perhaps that's memories, not consciousness? So maybe the question is what happens if I wake up one morning as a completely new person with a completely new set of memories - how can I tell? What is the "me" there?
I don't understand why people are so convinced that MWI is true. Yes, it is a natural conclusion from our existing laws of quantum mechanics; but our laws of quantum mechanics do not fully explain our universe. We know for a fact that our models are incomplete, and there is likely to be underlying physics of which our existing models are an emergent property. Why assume that the universe works based on an interpretation of an incorrect model?
To add to that, even if MWI is true then dying in one branch decreases your probability mass. If you care about all versions of yourself, it's better to avoid that.
Sure the theory of everything hasn't been discovered yet and we don't know what it is, but we do know that in lots of cases it can be approximated by quantum mechanics, just like quantum mechanics can be approximated by newtonian mechanics. Given that quantum mechanics is a good approximation in practucally every case we can observe, it seems likely that, for any given phenomenon we can't observe (such as separate world branches), quantum mechanics will probably also be a good approximation in that case too. While we know that the standard model isn't always correct, the situations where we know it isn't correct are ones involving gravity, extremely high energies, something to do with neutrinos I don't really understant, and probably other cases I haven't heard of, and I see no reason to believe that decoherence is one of them other than an intuition derived from our experience of a basically newtonian human-scale world, which is even less of a complete description of reality than the standard model is. To assume that MWI is *probably* false, rather than merely *possibly* feels to me like using our ignorance as an excuse to believe what you already want to believe for other reasons.
"Given (i) MWI, and (ii) that AGI ruin occurs with less than P = 1.0, then by quantum suicide/immortality logic, none of us should fear personally experiencing AGI ruin."
I don't care if there are 100 worlds out there where the guy pointing a gun at me right now doesn't shoot me in the face, I very much care if in this one world I get shot in the face. And if I do get shot in the face, I'm dead and whatever happens in the 100 other worlds is meaningless.
Go find a parent whose young child has died and cheerily tell them "Hey, under quantum immortality, your kid is alive somewhere else so stop being sad!" After you get punched in the face, console yourself that there are worlds where you didn't get punched in the face, so your pain right now isn't happening.
"You should therefore never expect to subjectively experience death."
That has to be the single stupidest sentence I have read all week. You can objectively experience death but not subjectively? You'll die but not notice you died? When dying, you will not have the feelings of dying? Unless you expect to drop dead immediately, when I suppose you might not subjectively experience it since you won't have long enough to experience anything, this is ridiculous.
The only sense I can make of it is "upon death, what we might call consciousness is immediately extinguished, so you cannot experience 'being dead', you can only experience 'dying'". Great, but I'm still dead either way.
>"Hey, under quantum immortality, your kid is alive somewhere else so stop being sad!"
It's not like i agree with Snakegown's theory but many religions do say very similar things when people die, so maybe it works. And QI is just a newer religion for physics nerds.
If he wants to re-invent reincarnation, he can do so but there are already religions with this as part of their structure. So it does seem like wasted effort.
It may be his religion, but it's not mine.
1)Wut? How about if you die after lying with a great block of collapsed cement building on your legs? You're in agony from you crushed legs, you're septic, you're dehydrated and alone. Most people's deaths aren't a sudden "lights out," which, I agree, need not be feared. Most deaths are preceded by a period of every-increasing disability and pain. Many have to endure that period with nobody who cares about them near at hand to comfort them. If that isn't worth dreading I don't know what is.
2) I'm reasonably OK with the prospect of dying myself. But I have a 20-something daughter who is on the verge of getting married, and full of joy at the thought of her coming years with her husband, building a house together, having babies, raising their children, going on camping trips with friends. The thought of the astonishment and grief on her face if the world comes apart and she realizes she can't have any of those things is absolutely unbearable.
This is not meant as a general purpose argument not to fear death or disability, but to assuage fear of AGI ruin. Posited AGI ruin scenarios are so total that you, your daughter and your future children have nothing to fear from them personally, since none of you will ever experience them. Positing an AGI ruin that you can experience, whether involving blocks of cement or otherwise, requires developing a specific, unlikely scenario to be worried about.
Yeah sure, assuming we're all snuffed out in the twinkling of an eye. But that's just one of many possible scenarios.
Even the scenarios that take longer to result in death will end (for you) when the time of death has passed. The you who suffers that death will identify with current you, but so will the you who exists in the non-AGI ruin universes. But the latter will not identify with the former. And the former will no longer matter. So “you” will survive AGI ruin (or rather AGI ruin will not occur for you), even if some other versions of you who you won’t care about will not survive AGI ruin.
Quantum immortality is a nightmare, not a get-out-of-jail-free. “Cannot experience your own death” doesn't mean what you _do_ experience will be pleasant.
Exactly, I don't think people understand how horrifying quantum immortality is. It doesn't just mean you will reactively experience bad things of unimaginable disutility. It also means you will pro-actively experience bad things of unimaginable disutility because of value drift or some other process which operates on the long-term. Nobody would take a pill which would make them do certain things, but quantum immortality means eventually because of probability (there is a non-zero chance) you will do those things.
You have free will though. Every choice you make confines “you” to futures that include that past choice having been made. Even though other versions exist who made a different choice. But you dont identify with those ones any longer, they are not “you”. Thus I wouldn’t worry about to the future versions of you that make a choice you would disagree with.
Free will makes no sense. if we have free will, then we should be able to choose to act against our own free will. But if we can do this, then we don't really have free will.
If our free actions are random, uncontrolled events, how can they be truly "free" and stem from our own willing and agency?
If every action requires a preceding cause that also requires a cause, how do we ultimately choose anything freely?
If our choices can be manipulated by directly manipulating our brains, this seems to challenge the hypothesis that there is some kind of free will beyond mere physical causality.
if our thoughts are strongly determined by evolution, how can we have free will?
There seem to be fewer things for a "free will" to do - neurons and brain events seem sufficient. But if there's no role for an agentive self to play, how can our will be free?
Even if we have free will now, in order to be ultimately responsible for our actions, don't we need to be responsible for becoming the kinds of beings we are - for developing our own character, values, and motives?
hi, would love to talk philosophy with you, if you can just drop the pseudo-science language (I think we can debate the core claims without that but lmk if disagreed)
Think the key claim you need physics for is
>You should therefore never expect to subjectively experience death (1)
which we can take for granted (and which I agree with).
Also seems most of the argument is really not much about agi but just death and what comes after so lemme ignore agi as well for the moment.
Seems to me you might be implying (1) implies that
(2) humans can't have preferences over duration of time till their death
(3) humans can't have preferences over what comes after their death
(maybe (humans can't have preferences) should be replaced with smth else but not sure what)
Think people would mostly disagree with (2) (metoo), but also bigger issue/concern is (3).
(3) I think is related to what's discussed here
https://www.overcomingbias.com/p/ai-fear-is-mostly-fear-of-future
I'm curious to hear what people feel about (3), humanity-consensus think is disagreement, philosophy-consensus and arguments for/against I'd be curious to learn about.
If you think there is only one universe, you may place great weight on whatever happens after your death, even if you won’t experience it. As many have through history. If you know you will survive in some branch, seems better to focus on optimising those branches.
"If you know you will survive in some branch, seems better to focus on optimising those branches"
How? How do I do that? In this branch, I will die some day. We know this for a true fact of reality. If I live to be eighty, I'll die eventually. How do I 'optimise the branch' where I live to be eighty-five? This is dream logic for dream worlds.
There is no way I can 'live' by leaping from branch to branch, never dying, because I live to be ninety in this one and then another version of me in another branch lives to be one hundred and so on for infinity. I die for good and all sometime, in some branch.
This argument also means that one's consciousness and one will never die.
Feature not a bug! I gather Everett himself understood this to be true
>So what do you think. Should this alleviate a rationalist’s concerns? Is this an idea worth spreading?
Absolutely.
It eased my mind.
But a true rationalist should have no concerns.
Excellent. Did this post help or had you already come to that conclusion? Am trying to figure out if it is worth spreading this idea.
I think it resonates with my growing sense that we worry too much and it’s not helpful.
Yeah, it’s like Communism that way. It seems Jack Spratt and his wife cannot get along. Never mind though because humanity’s hive -mind-child AGI will change that. It has the best chance.
I don't think quantum immortality is generally taken as seriously as you imply, even by those who accept MWI.
By the same logic, no-one should ever expect to experience dying in a car crash. But MWI should have no impact on whether you experience your brakes failing or not, or whether you experience falling asleep at the wheel and swerving into a truck, or even whether you collide with that truck and begin flying through the windscreen. By the time this principle could be applied you are already dead in all possible universes with near certainty.
This doesn't make sense. If there is even one universe where "death is avoided by epsilon - Jeffrey Soreff" then you will exist in that universe. If there are an infinite number of alternate possible universes then there will always be one in which you end up in. But this applies for first-person perspectives. From a third-persons perspective, people die all the time.
AGI ruin involves everyone dying, so there are no third persons.
Yeah, but from a first-persons perspective you cant die. So it means you end up in a scenario where everyone dies and you survive. These two viewpoints make no sense to me and I have no idea on how reconcile them.
There's no quantum immortality force stopping you from experiencing a world where AI has unleashed a deadly virus that has infected everyone on the planet, all of whom will die within days. The point of divergence between the worlds where you exist and the ones you don't can come after the point where you're already guaranteed to die in every universe.
To get around that, the claim you have to subscribe to is that something of ridiculous unlikelihood would save you - maybe all of the virus in your body simultaneously falls apart and you survive. And that these unlikely events would keep saving you.
Some say that quantum mechanics makes this inevitable. I don't see why it needs to personally, but I sure hope it's not. Because that guarantees that all of our subjective experiences end up in eternal torment as everything else breaks down around us. In this example, you're still the only survivor in a world ruled by AI. Not comforting.
I believe poincaré recurrence theorem says that anything that can happen, will happen.
Presumably this implies you can luck your way to immortality. Say you are 120 you are likely to die soon of something or other. But if you die of a heart attack on 2 June there's a world in which you survive that heart attack, and the next one, and the next one...
There is. The possibility is that the virus just happens to not infect you. Or the virus just happens to miss all of you healthy cells. Or you immune system just happens to create antibodies from a previous asymptomatic infection which is enough to trigger an immune response to the deadly virus.
One of the nice things about AGI ruin scenarios is that they are often quick and unanticipated deaths, so those of “you” destined to die would experience it only briefly, if at all. But even if there are bad scenarios, you will still survive in all the other worlds on the other side of the point of divergence.
"If there is even one universe where "death is avoided by epsilon - Jeffrey Soreff" then you will exist in that universe." Yes, and that existence is very likely to be highly unpleasant. Being _almost_ killed is generally a torturous fate. Which is why I don't think quantum immortality is particularly comforting.
Yup. You are looking at a bunch of unpleasant precursors to death, and I commented about unpleasant situations where death is avoided by epsilon, but in either case quantum immortality is (nearly) no help.
>MWI remains our civilisation's best theory of the nature of reality.
Curious how you arrived at this conclusion. I thought it was considered to be a fringe theory.
Yes by far the current "best theory" is "its not clear WTF is going on", not MWI, which has maybe IDK 5-10% support among the people who really deeply understand high end physics. and its metaphysics.
"Any phenomenon that involves obliteration of your consciousness, while it does occur in some measure of universes, is not something that can be subjectively experienced by you. You should therefore never expect to subjectively experience death."
To my mind, the problem with the quantum immortality argument (regardless of AGI) is that there are possibilities which are epsilon away from death and are very unpleasant.
The AGI ruin scenario is special because it posits death of all humans. Even if you don’t take comfort from QI re normal situations, it seems to work here at least. Esp if you personally do not expect to influence the likelihood of AGI ruin.
As it happens, I indeed don't expect to influence the likelihood of AGI ruin.
I see your point about the special nature of AGI ruin, but I think it is less binary than you posit. There are slivers of probability amplitude where e.g. portions of nanomachines that an ASI intends to use to kill us just barely fail, and leave the observer in question severely damaged but just barely alive. Quantum immortality will push some of the amplitude of the observer into these scenarios too.
I guess I'm skeptical of quantum immortality, even given MWI, because it rests on there being something really really special about the portion of the wavefunction where the observer survives. But that seems to rest on the observer's "survival" being a very crisp binary function of the microstate of the system. That seems implausible to me. Just exactly how many firing neurons does someone have to retain to count as "alive"? It seems more plausible to me that the observer's status is a smooth function of the microstate, and the importance of their being there as an observer fades out smoothly as the probability of their being a fully functioning human drops towards zero. Which brings us back to something much closer to the ordinary view where quantum immortality can be ignored.
Agree not binary. But if someone is anxious (not you it seems), then this idea should alleviate it - since they would have to go out of their way to devise a scenario (a sliver as you say) to be worried about.
Objective probability and subjective probability just aren't the same thing. QI has to be based on subjective probability.
Your subjective probability of witnessing anything whatsoever is 100%
Aside from currency, the one thing I thought blockchain would be useful for is for things like platform-agnostic libraries of content. For example, you might own a key for a game on the blockchain, and you could then play that game via any participating digital distribution platform (ie Steam clone) that had that game in its library. This would protect your game library in case Valve or whoever shuts down, it would allow you to purchase one copy of a game that would be playable on PC and mobile, it would allow better competition between storefronts, and you could then sell the game key on when you're done.
I get why existing platforms would not want to implement this unless under heavy pressure, but why have I not heard of any new platforms offering this service? Is blockchain not as useful for this as I imagine? Is it a bad idea for other reasons? Do consumers not actually care?
Blockchain tech has been very underwhelming in general considering the hype ~10 years ago. But at thw very least these kinds of use cases seem like they should have been exploited by now.
I mean, this is a version which ends with strictly worse sales for all companies which provide games to that platform (as they get $0 for every resale), so I would anticipate that any service which did this would basically have to be GOG clone, or actually own the games in question. They're definitely not getting triple A games.
It's arguably better for the consumer, but worse for both the storefront and the developer and the benefits for the consumer are honestly pretty low, so, no I wouldn't expect something like this to be of any use.
It would also, almost inevitably push game development in a direction I don't particularly like (away from single-player, story-focused, one-time games and towards ongoing DLC heavy, multiplayer and 'endlessly replayable' games, as you'd want John and his buddy to both buy it, not John to buy it and then 'sell' it to his buddy), so I'm glad there's not much chance of it happening.
I agree that the resales are the messiest part. The main model for digital game sales right now is to do progressively bigger discounts over time until you've theoretically captured everyone who would ever buy your game at the highest price they'd be willing to buy it. This captures the group that would traditionally have waited to buy second hand.
If resales were a thing this model would need to change. Games would probably have a higher upfront price. I would also expect stores to facilitate resales and take a cut for themselves and the developer. Users could exchange games outside of stores, but the extra friction involved in finding someone willing to buy, the knowledge that you're ripping the devs off, and the minimal savings involved would make this rare I think. If stores really wanted to discourage it they could refuse to honour resales that haven't gone through them or a recognised competitor, but it would be against the spirit of it all.
It'd be weird but not insurmountable I think.
I disagree that the benefits to the consumer are low. Or at least, I'd expect the benefits to be very compelling to like 5% of consumers, while most don't care and stay on traditional platforms. The target consumers would be the same ones that care about DRM-free, who are still concerned about Steam having a monopoly, and a bunch of crypto people.
It would be a little worse for existing storefronts, but I'd expect this to be implemented by new storefronts offering it as a point of difference. It's infinitely better for them, because they wouldn't exist otherwise.
It's pretty neutral for developers. Indies are better off being on as many storefronts as possible and benefit from sales on smaller ones because they take a fraction of the cut that steam does. The resales thing is the only problem that some might have, but as I said, I think that's surmountable. Most revenue comes from launch anyway, and your tail would come mostly from cuts from resales instead of dirt cheap sales.
I mostly agree about the direction it would push games. There are a few forces pushing them in that direction anyway though. I'm pretty ambivalent about the idea myself, just surprised it doesn't really exist.
I mean, all the emphasis I've seen is going the exact opposite direction of making it seamless to transfer between platforms, or resell games, or anything like that. Especially with things like Netflix cracking down on account sharing, the concern is that they've made it too easy to share, not easy enough.
On the specifics of resale, I think you'll have a very hard time with 'residuals' for resales. At the moment, people have basically decided to accept the fact that what you buy from Steam/etc. is a nontransferable personal license. They don't generally love it, but they don't think of it like they do a physical book. If you manage to get them to think of it like that, then they likely won't agree that 'residuals' are appropriate or selling in such a way as to avoid them is improper.
Now, the platform might be able to make some money (Ebay survives just fine, as far as a I know and paying to use their resale platform makes some sense).
But in the end, this idea (in my view) fails because the actual advantage it you tout 'transfer between platforms' only comes into existence if everyone adopts it, which means there's no competitive advantage to doing so. It might be better for the entire overall games marketplace/consumers (though I have my doubts) but it wouldn't be better for any individual retailer/developer, so there is no incentive to create/adopt it.
The trouble with that is that, even if Valve gave out a crypto token for everything in your library, there's no reason another storefront would want to honor them. Why would they? From the second store's perspective, that just means giving you a lot of games for free.
You also have a bit of a chicken and an egg problem: whoever does this first can't promise that the crypto tokens will ever be usable in another store, they would have to just hope that someone else hops on the bandwagon despite not having an incentive to do so.
I could see a company doing this as a way to attract new users, similar to Epic's free games - "We'll give you all your old games for free, just sign up and start using our service!" but that's going to be an expensive way to get users. And you don't really need a blockchain to do that, either - you could just look at Steam's API.
"Why would they? From the second store's perspective, that just means giving you a lot of games for free."
They're giving you a digital asset at essentially no cost to them, which you would not be inclined to buy from them anyway since you already own it. In exchange they have made it very likely that you will make new purchases on their storefront by reducing the cost of switching to nearly zero.
It's not that different to stores like Humble and GOG now offering Steam keys with purchases.
Developers would like it (speaking as a developer). We're not trying to sell you the game more than once, and we don't care what storefront you're on. In fact, smaller storefronts take half the cut that Steam does, so it's better for us if people switch. It would even make it easier to sell directly from our own websites where no-one takes a cut.
The main issues I see are the effects of the secondary market (mitigatable) and getting exisiting storefronts on board (very difficult, but not required to get value from this). But I'm not trying to lay out a foolproof business case, I'm just expressing surprise that I've not seen anyone trying this. To me this seems less dumb than basically any exisiting blockchain use cases.
DRM (Digital Restrictions Management) is fundamentally unsound for technical reasons. I do not think the block chain can be added to the mix to avoid single points of failure practically.
If you want to use the blockchain for resilience, this means that you may not depend on any central service interpreting it. While the block chain can establish a group consensus in the face of a minority of participants being adversarial, it can do little to help a single node whose traffic is controlled by an adversary. Passive verification would require parsing at least the block in which the purchase happened and then all subsequent blocks (to make sure that the product was not sold subsequently), and would vulnerable to attacks by someone just withholding the block of the sale and all subsequent blocks. To counter that, the DRM process could actually try to write on the block chain, verifying that someone is spending enough compute to calculate the next block. This still does not feel remotely feasible.
The other route would be that the copyright holder could commit to allowing anyone owning the appropriate token to download the game from bittorrent, and pre-commit to putting their product in the public domain if their blockchain verification website ever goes offline for more than 48 hours. This would be very resilient, but might not be a very good deal for the publishers.
In my model, the reason you can't resell used digital games is because neither stores nor game publishers want you to have that ability. A single upstart store can't buck the trend because the publishers wouldn't join so they'd have nothing to sell.
If they wanted to give you this ability, they wouldn't need blockchain, they could just do it. I'm not sure what problem you think blockchain solves.
You might think it's a way to tell what games a user has already purchased elsewhere, but stores already solve that problem when they want to. GOG had a feature for a while where they'd add certain games to your GOG account if you already had them on Steam. And Steam already gives game publishers free "Steam keys" that they can sell through other stores (which is effectively Steam giving you a free copy of games you purchased elsewhere), and Steam doesn't even take a cut, they just demand that you ALSO let customers purchase the game directly through Steam.
You might think that blockchain would mean that you don't have to trust Valve when they say John owns the game, but actually you still do, because in order to sell the game, Valve would need the right to add new copies of the game onto the blockchain. I suppose you could theoretically set it up so that each individual publisher adds copies to blockchain and then gives them to Steam to distribute, but that seems pretty similar in practice.
If you're worried that Steam might suddenly collapse and you'd lose the records of who owns what, then you just need an archive, not a whole blockchain. Steam could just periodically publish that data if anyone cared enough to archive it (it could even be digitally signed if you care).
Assuming that the stores and publishers actually wanted to do this, what part of it would they need blockchain in order to pull off?
I think frictionless transfer of ownership of Steam games would change the economics fundamentally.
At present, there is probably less than one play-through per primary sale, and the sticker price of games tends to drop by 25% per year after it is out, perhaps.
Assuming playing through a game takes four weeks, this would allow twelve people playing through the game in the first year alone for one single primary sale. There would still be a hefty premium on buying the game the day it comes out, bugs and all (just like there is a premium to buying a new car instead of a year-old ones), but afterwards the cost of owning a game for a month would just be a few percents of the sticker price.
Even for games people play long-term, frictionless transfers of ownership would change things. I have probably spend 100s of hours on FTL, but the truth is that I spend much less than 5% of my day playing it. A rental service renting out games so that each copy is played 50% of the time would drive down prices even more.
For video games, the costs of copying and distribution are basically zero (see bittorrent). I would assume that most of the costs actually involve creating the game. The sticker price for a frictionlessly saleable game would have to be much higher for it to be economically viable. I have no reason to think that I would spend less on games of this became the new equilibrium. (Do I sound like a bird arguing why his gilded cage is actually fine? I suppose I do. I am aware that "owning" a game on Steam is not the same as owning a physical object. Practically speaking, the chances of Steam going out of commission in any year are seem lower then me going out of commission, so I am mostly ok with it.)
I'm not actually sure what you're arguing.
I broadly agree that a system where you can sell used digital games would be very different from the current system. I'm saying the reason we don't have such a system is that game stores and game publishers do not believe it's in their interests, and that if they wanted to build such a system they could do it without blockchain.
What about the LBRY protocol? It is a blockchain-based platform-agnostic library of content, if I understand what you mean:
https://en.wikipedia.org/wiki/LBRY
The most populous current isntance of it is https://odysee.com/, but it is platform agnostic.
That's interesting, thanks. That is the sort of thing I'm talking about.
I believe there's some work being done (Similar to Zvi's work on Emergents) on some CCG's / game pieces that can be owned / tracked independently of the underlying game substrate - meaning someone else could create another game system that you could use your assets in, which is a kind of cool concept!
See, these kinds of use cases make less sense to me because no-one is going to develop a game based on some other games assets. So you're just left with some inherent support for a secondary market for as long as the game is active.
Zvi himself said in his last AI post (https://thezvi.substack.com/p/ai-12-the-quest-for-sane-regulations):
"I created an NFT-based game, started before the term ‘NFT’ was a thing. I did it because I saw a particular practical use case, the collectable card game and its cards, where the technology solved particular problems. This successfully let us get enough excitement and funding to build the game, but ultimately was a liability after the craze peaked. The core issue was that the problems the technology solved were not the problems people cared about - which was largely the fatal flaw, as I see it, in NFTs in general. They didn’t solve the problems people actually cared about, so that left them as objects of speculation. That only lasts so long."
Nobody is likely to base a game's design off assets that they don't have the legal rights to, but the execution of subtle themes on an idea is extremely common. Consider all the different D20 RPG's that exist. Or, for another example, there are at least a dozen home-brewed MTG variants that I've played around kitchen tables over the last few decades, from slight changes to make multiplayer better to entirely new games that recycle MTG cards. The creators of these variants couldn't really profit from them (WoTC's legal department would shut down any attempts), but they still went to the effort of creating and distributing the rulesets locally, so I'm not sure I buy your argument.
If I could take my digital cards from Zvi's game and use them in another game's engine (whose creators presumably operate under a broad license that allows them to do so, which I don't think is the case with the assets from Emergents), then they would have more value for me than being locked to that one game. The existence of software like Tabletop Simulator also strikes me as good evidence for this kind of ecosystem's potential to emerge.
The problem is that digital games take years of hard work to make, while MTG variants can be thought up in a single fun evening (I have done both, for what it's worth).
Most game developers don't want to spend their years piggybacking off someone else's IP. Some do - see the Black Mesa project - but the original game needs to be pretty damn popular for it to make sense for devs to spend their time working on a spin off instead of making their own projects.
Modding is the usual solution for this - it allows variants to be created without having to build something from scratch.
Why would any new platform benefit from tracking who owns stuff on the blockchain rather than tracking it on their own database? The only benefit to the blockchain is you don't have to trust the person/organization who controls the database. Obviously they trust themselves, so there would have to be a strong customer demand for them to give up the control.
But then even if they store the ownership data on the blockchain, what does that actually mean? Someone still has to honor that ownership by eg. providing the game for download to the supposed owner. If you trust them to do that, why not trust them to keep all the ownership info on their database too?
My assumption was that customer demand would be strong enough to make it a selling point for new platforms. The same way that DRM free was a point of differentiation for GOG and Humble that helped them become established.
DRM free is a selling point that customers value because DRM is a pain in the ass in specific ways (and even that's pretty marginal- many customers don't care at all). What tangible benefit does a customer get from an NFT saying they own the game? They still need to trust the distributor to make it available for them to download. Can you describe a situation where having an NFT makes the customer better off than just having an entry in the seller's database?
I listed the advantages in my original comment, but I'll expand on them. Here are the problems it would fix:
1. I own hundreds of games on Steam. If Steam shuts down for whatever reason than I would lose them all.
2. I am locked into Steam as a platform because that is where all my games are. There are competitors to Steam that I prefer, but I avoid them because of the hassle of splitting my library up.
3. I own games on Steam that I would like to play on mobile, but I would be required to purchase the game again. That's not always worth it.
4. I can't resell games I've purchased over Steam.
DRM free helps a bit with some of these but not much. Eg. unless I keep all of my DRM-free games downloaded at once my library is still at risk of disappearing if Humble or GOG go under.
In a world where your game library could be on the blockchain, Steam going under would be no problem - you would just sign up to another platform that honours blockchain games. The advantage to the new platform would be that they could take a cut of new transactions you make on their platform.
I think that's enough reason to justify that this is a feature that some people would want. And I think it's enough to say that it's unusual that there seems to be a complete absence of blockchain platforms of this form.
Somewhat similar to my question last week on the open thread: has anyone been pulled out of a depression through falling in love with someone? I've been writing about my experience with schizophrenia over on my Substack, and Friday's installment featured the thing that sparked my interest in working out and taking care of myself, as well as for life in general. And that thing was falling for a girl who was unavailable. I'm not a romantic, per se, but I think there is something to the whole Lex Fridman-esque belief about the power of love to move people to good in the world. Interested in what ACX readers think about love and the power it has to move people both individually and collectively to make things better.
Sort of, but it's not so much "pulling out of depression" in the sense of a chemical imbalance. One experience was starting to fall for someone, noticing it, asking them out, getting a polite no, and moving on to being friends. In that case, the thing that helped was mostly the realization that I could still have those feelings at all, which gave me hope for the future. Another experience was being around someone for a weekend who was clearly subtly attracted to me just as I was to them. Nothing happened for a whole lot of very good reasons, but it was still a nice wake-up that such things are still possible.
Being perpetually single in my early 20s left me feeling quite depressed, in a way that was almost instantly solved by actually getting a girlfriend, even though that relationship lasted mere weeks. One's self worth can be rather strongly affected by other people's demonstrated valuations of you.
The flip side that you describe seems like a plausible kick to break out of a negative rut, but I highly doubt it would be sufficient by itself rather than as the last piece of a puzzle.
Yeah my depression was, for about 10 years, linked entirely with being perpetually in love with people who were not in love with me. It wasn't until I was not depressed that people would actually want to be in a relationship with me.
Absolutely. I have a patient whose father died when he was a child, several decades ago. the man has in fact had quite a difficult life, mostly because of OCD and depression. He much prefers to discuss his life in an "understand yourself" way, rather than in a cognitive-behavioral let's break patterns way, and most of what he has to say about the problem of being him is that growing up without a father prevented his forming a feeling of having a coherent, adult self. So in the last 6 months the man has met and fallen in love with a women who is a good match for him in interests, a good sexual partner, and pretty, smart and kind. He's happy. And now he's saying to me, "OK, I lost my father -- but why does it follow that that ruined my life? I can just think of it as, it is what it is."
I entered a years-long depressive state in the aftermath of love. I eventually made my way out via building friendships with other people and a single remarkable dream that I had one night. I don't think, based on my experience, that it would have even be possible for me to have had experienced love during the depths of my depression, for the same reason that I wrote no poetry (but was astoundingly prolific both before and since).
I would love to hear what the dream was.
I wrote a history of how independent courts gained the power of judicial review in common law systems. It's a history focused on the institutional questions -- how do the courts internally discipline themselves, and how do they use this discipline to influence other branches, despite lacking the power of the sword or of the purse -- and so it's rather different than the standard case-focused histories of this which lawyers tend to write.
Link: https://cebk.substack.com/p/producing-the-body-part-two-of-three
This is my second (and last) time posting this in an open thread before the next classifieds thread, but my sister (trans-woman, 24 years old) has recently moved to Seattle and is striking out left and right trying to find a job. She graduated a year ago with a bachelor's degree in actuarial sciences, has some coding experience, and has a good chunk of capital to use on finding the right position (rent for multiple months while unemployed or in a poorly paid internship, headhunter fees, etc.).
Didn't mean to bury the lede like this, but does anyone have any advice? Specifically in Seattle, but Portland or the Bay Area would be similar enough that I'd be interested. Are there agencies or single contractors who specialize in job-hunting that you've worked with? Does your company have openings for someone with that kind of CV? Have you been in a similar position and found a strategy that worked well?
Feel free to respond in a comment below or email me at: 7o2wzrybd (at) mozmail.com
Coding bootcamps, if they're any good, should have connections to hiring tech companies for their graduates. Any bootcamp that wants to teach you at their own expense must genuinely believe that they'll land you in a job afterwards ( of course they'll keep a cut of your pay, but it may be worth it to keep the incentives aligned ). If the bootcamp wants you to pay them, take a critical look at how many of their graduates find work and make sure they'll actually put their network at your disposal. After your first job, people will take you more seriously, you'll have your own professional network, and it'll be easier to get your second job, even if you realize you want to do something besides coding.
what does she want to do?
actuary? tech?
as both are probably doable but both might require a bit of specialized effort investment to increase the chances (as presumably she tried out and didn't immediately get anything)
She's undecided, a little demotivated by the lack of success so far. Either one of these routes described above or below (coding - boot camps/actuary - actuarial exams) may be the best route forward. Thanks for the thoughts.
hope she has people/therapy/resources to allow her to not feel pressured over it. yeah the job market is tough.
but viewing it as being about getting whatever job instead of making a thoughtful decision about the next arc of her life that will probably last at least a few years and will have resume effects for way longer.. might not be wise.
not everybody is privileged, but this is the kinda thing where privilege, not having the crazy pressure as well as confidence that one can have reasonable success in whatever reasonable path they choose, is quite material for making quality decisions with significant impact on one's life.
Is she interested in performing actuarial work? My brother was an east coast actuary and the way he described it there is a formal track of actuarial exams and completion of a certain number of exams essentially guarantees employment.
That's great information to hear, the last Open Thread had this recommendation as well, so we'll dive more into the Actuarial exam track.
'''$500 Bounty/Prize Problem: Channel Capacity Using "Insensitive" Functions'''
https://www.alignmentforum.org/posts/YDSjSpD7yBoivMHay/usd500-bounty-prize-problem-channel-capacity-using
I’ve been thinking recently about the concept of the “deep state” through the lens of Franz Kafka. While it’s common to compare the current state of the world to Orwell’s 1984, I believe two of Kafka’s novels, The Castle and The Trial, contain themes which are more relevant to understanding the insidious power structures of society. Kafka imagines a world in which bureaucracy is a sort of Frankenstein, an unthinking malevolent force created by humans which has, without anyone really noticing, enslaved the humans. The law does not exist for man; man exists for the law.
One might compare the law to the rules of an algorithm abided, not by the pure logic of microchip circuitry, but by fallible humans. In Kafka’s universe, the chief weakness of man is stupidity in all of its manifestations: incompetence, immaturity, narrow-mindedness, pettiness, incomprehension. Kafka’s anti-hero, K, is not stupid. But everyone he encounters: his work assistants, cops, lawyers, clerks, officials of The Castle, magistrates, the mayor, schoolteachers, tavern workers, the townspeople... are drowning in stupid. The world is a bafflingly idiotic yet complicated place full of rules which, at every turn, work against an honest, earnest man who merely wants to do his job and stay free of prison.
Unlike with Orwell, Kafka’s novels are not dystopian warnings to the world. How can one warn the world that the world is too stupid to function according to its own rules? It isn’t the rules which are stupid; it is the people living by and enforcing them! It is the cops and lawyers, the judges and juries. And it isn’t a stupidity which one can be educated out of. A court magistrate may have a sophisticated understanding of the law, for instance, but is also petty and narrow-minded. Stupidity is a cancer that has spread throughout the organism.
I started thinking about this subject while pondering AI risk. It struck me that we’ve already created a monster that is much like an alien intelligence running our world which is indifferent to the plight of humanity. It is all our bureaucracies!
Could you help me understand your post a little better? You seem to mention the interpretation that Kafka was writing a critique of bureaucracy, then downplay it in favor of the interpretation that Kafka was writing about a (negative) experience of humanity in general, before offering your own original critique of bureaucracy. But critiquing bureaucracy fits better with the original interpretation, which you had seemed to be setting aside. Have I misunderstood?
Yeah, my post wasn't very clear.
First, Kafka isn't critiquing anything. He wrote novels that are essentially extended nightmares. The nightmares are about dealing with various hells of bureaucracy: in The Trial, Joseph K is arrested and doesn't know how to defend himself because he has no idea what crime he has been charged with. In The Castle, K is a land surveyor who has been called by The Castle to do a job but can't figure out what the job is; try as he might, he can't work his way through the bureaucracy of The Castle to get information pertaining to his job.
These are absurdist novels.
My argument, which is independent from what Kafka may have ever thought, is that these nightmares of bureaucracy do offer insight into the way the world really works.
Kafka's insight, as I see it, is that the weakness of bureaucracies is that they are run by humans, who mostly aren't competent for the task. In other words, the main problem with bureaucracies is not that the rules and regulations are themselves stupid, but that the people charged with enforcing them are. In Kafka's novels, at least, the people working inside a bureaucracy are absurdly stupid.
I then attempt to compare this world in which a rule-based system (a bureaucracy) is unaligned with humans because (ironically) it is run by humans--with a world run by an unaligned AI.
Is that clearer?
In my interpretation, the two biggest influences on Kafka's writing (fiction and journals) were his basic experience and a corollary of that experience.
- The basic experience was that his world didn't make sense.
- The corollary was that the right thing to do could not be established.
I take this as mostly Kafka's tacit knowledge, i.e. a pre-theoretical attitude, rather than coming from philosophy or theology. Kafka's attempts to theorize this attitude in his journals tend to be self-consciously absurd. (Contrast that with Samuel Beckett, who seems to have taken Jungian ideas seriously in an attempt to explain his own basic experience.)
Kafka was an administrative lawyer, so it's not surprising that he sometimes expressed his underlying attitude with fiction about law and bureaucracy; he was Jewish, which probably gave more valence to law in particular. But I think law and bureaucracy were basically symbols for him, so I don't separate law from bureaucracy from bureaucrats from protagonist when reading The Trial and The Castle. Nor do I see much difference between the bureaucratic fiction and the other fiction.
You may endorse the Walter Benjamin interpretation of Kafka as [sic]:
>not so much a failure as a comic figure. Kafka is man 'whose fate it is…is to be surrounded by clowns.'
(https://schlemielintheory.com/2014/04/25/literature-and-failure-on-walter-benjamin-and-howard-jacobsons-description-of-literature/).
Whereas I would modify that by saying that Kafka was a clown who knew he was a clown, and mostly did not want to be a clown, yet did not know how to be anything else. It seems to me like he amused himself with the (pleasant?) daydream of being a non-clown tormented by clowns strictly external to him. Superficially, that's The Trial and The Castle, among others.
So I largely agree with Michael Wood's view (and apparently Stanley Corngold's):
>In this sense it is surely a mistake, as Corngold suggests it is, to pit Kafka’s bureaucracy against its ‘hapless supplicant’. He is bureaucrat and supplicant, perhaps more bureaucrat than supplicant.
(https://www.lrb.co.uk/v30/n22/michael-wood/double-thought)
None of this is to disagree with your point that unaligned AGI might be more familiar to us than we think. This is an interesting idea and I hope you post more about it.
I wouldn't call it "stupidity", so much as conflicts of interest and tunnel vision. Rules are made that ignore some parts of reality, and contain incentives that aren't acknowledged. It's not a lack of intelligence per se, it's what that intelligence is focused on, and what that intelligence studiously ignores.
Is this meant to be reassuring? Bureaucracies are a slow cancer destroying all that is valuable. I expect AGI to be much faster (molasses are usually faster than bureaucracies, it's not a high bar), and not even theoretically containing components whose conscience might wake up and cause them to heroically defy the machine they are a part of.
No, it is not meant to be reassuring.
common counterargument is that "indifferent" is an overstatement, human institutions (in the free world to a larger extent) are to a decent extent shaped by aggregate human preferences and subject to various feedback loops making them more human-compatible
About the birth rate problem, it's odd that there aren't companies or businesses or even business models coordinated around trying to increase birth rates.
If you think of small fundamentalist religious cults as a business run by the cult leader then there are dozens and dozens of these business just in Utah!
Oh, but remember a few years back when Apple offered egg-freezing to its female employees? They will graciously permit you to have a family after they've squeezed all the use out of you and you're no good to them anymore, so you can go ahead and waste your time with babies:
https://www.theguardian.com/technology/2014/oct/15/apple-facebook-offer-freeze-eggs-female-employees
I am very damn cynical about that because companies don't want staff taking large chunks of time out, which is why maternity leave is generally a legal imposition by the state and not something voluntarily offered. A business working on increasing the birth rate would probably require its staff not to be going out on maternity/paternity leave too much, which would be an instance of saying one thing and doing another.
It's illegal. Human cloning is illegal. Slavery is illegal. Commercial surrogacy is illegal.
There are some things are not illegal yet, but be if commerce will start to use them.
Oh, freedom of association is illegal.
Who would be paying the companies? There are governments trying to increase birth rates, but so, far, I don't think they're working through companies.
Also, the governments are 'trying', but aren't willing to pay very much. We have fairly strong evidence that direct cash payments for babies help but they need to be much larger than governments typically offer (orders of magnitude larger, if you want to get TFR back above replacement)
They don't need "to pay much", rather they need to pay less.
Just stop paying retirement pensions, stop subsidizing women higher education (most of higher education degrees are useless, and womens are to larger degree) => TFR up with no extra dollar spent.
So only men go to college, no women. Great! And, uh, who are these college-aged women marrying to have those babies? Older men, I presume?
So now you've got a bunch of 18-22 (or older) men who have no chance of finding a permanent relationship until they're 30-40 and earning high salaries. That'll work out fine. As will having a lot of women out of the workforce, because it's companies and the economy that want women working.
Will the children be okay with the idea of taking care of Mom and Dad who will need their financial support in old age? Already you have families putting parents into nursing homes rather than bringing their elderly parents to live with them. And of course, if the potential parents decide that instead they both need to work and save, in order to live without pensions in old age, and having kids is an expense that means taking away money they could be saving.
Society has done its damnedest for decades to undermine child-bearing, stay-at-home housewifes, and the idea of 'too many people on this fragile planet'. You can't simply reverse course that easily to bring the birth rate back up, especially when the people you want having more babies are the exact people who don't want large families.
as Zvi has pointed out, they also aren't willing to undo any of the myriad of things they are currently doing which discourage having children either. (which would, potentially, have a much larger effect size than paying people.)
Fertility treatment clinics?
This is not a problem for any specific couple. Any couple that wants kids and has the resources for it, can have more kids.
So:
1) or they don't want kids - Business can't help.
2) They don't have resources for more kids - Business can't profit, so can't help.
So this is a jointed-action problem, where we all agree there should be more kids, but no one want to have more of them/support those who do by greater taxes, etc.
I've seen so, so many discussions of low births rates in these comments without anyone actually looking at stated reasons people don't have kids. I'm not saying you need to 100% trust exactly what people say, but it sure seems a decent place to start, and for many people, the reason they don't have kids / don't have more is housing and childcare.
There are cultural factors as well (having three kids wasn't a big deal for my two working parents because they would just kick us outside all day and we walked ourselves to school, both of which would get child services called on them these days), but simple inability to afford the kind of childhood people were given where the jobs are is a pretty straightforwardly understandable reason.
"where we all agree there should be more kids"
I, for one, don't agree. We had a vibrant culture when the population of the world was 2 billion. As technology stands right now, we've overrun the ability of the planet to absorb the CO2 our civilization emits. I'd rather stop deaths from aging - but we don't currently have the biomedical technology for that. As technology stands, I have no objection to a slow shrinkage back down to 2 billion. We can defer any questions about incentivizing births to that point.
And kindly remember, as of this moment, the population of the world is still _growing_, not shrinking.
You have a point. With current technology there aren't a lot of options. The population cannot keep exploding. We've overrun the ability of the planet to absorb the CO2 our civilization emits. Yes, with a slowly declining population for a century or so (setting aside AGI or other technology advances to simplify the discussion) the population will be mostly middle-aged during that period (presumably slightly younger when it stabilizes (itself a long discussion)).
yeah coz the idiots outlawed human ownership and destroyed the incentives
Its not odd if your ultimate goal is to depopulate the world to where you have just enough workers to undertake work AI can’t do whilst exterminating ‘excess’ consumers gobbling up resources you want...
"exterminating"? As far as I know, even in the most extreme case of South Korea, the depopulation is happening due to deaths from old age, not gas chambers. Now, I'd like to see aging solved like smallpox was solved, but it is a very hard technical problem and unlikely to happen soon.
I can't really imagine what revenue stream such a component could get?
Does anyone have experience with EDMR they are willing to share? The wikipedia page positions it as pseudoscience and I'm wondering where it falls on the spectrum between "pure snake oil" and "worth trying". I have a traumatic experience from two years ago that still jumps to mind multiple times a day. Thanks for any pointers!
I did EMDR to deal with negative thoughts, and I would say it worked. I believe it worked by creating a disturbance in the negative thought-pattern. I suspect "The Work" of Byron Katie works by creating an alternative to the negative thought pattern. I think CBT sucks. For example, I don't see how one can become less judgemental by judging one's thoughts as distorted. I've also found a mindfulness practice to be useful for dealing with negative thoughts. I welcome critique.
Yes. It appears to be working for me. My PTSD is not as bad as it used to be. I began treatment a few years ago, and saw improvement, then the covid-19 lockdowns hit and treatment stopped and I saw stagnation and a bit of backsliding, and now treatments have started again and I'm once again seeing improvement.
I have no solid idea why it works. It might as well be witchcraft, and I could easily imagine some poor old woman from a few hundred years ago getting burned at the stake for discovering this. But here's some thoughts.
First, one thing I noticed when I started EMDR is that I'd stopped dreaming at some point. Prazosin made the nightmares go away, but in the years after that, I don't think I had the sorts of dreams I'd had before. Maybe it's that my life had become static and circumscribed, and dreams help integrate new stuff, or maybe it's that my memory was damaged by both PTSD and also some treatment for depression. And it's probably not that I completely stopped dreaming, but rather that whatever dreams came were below the threshold of my notice. But after my first few sessions of EMDR, I began to have vivid, strange dreams, and I realized that I hadn't been having any for a long time. Now those vivid dreams have died down, but it was a distinct change. So on the model that dreams are processing of experience, perhaps that processing had become blocked somehow, and EMDR is deliberately re-starting it, like crank-starting an old motor, or priming a well pump. Perhaps the eye movement stuff has some connection with REM sleep, although I've also heard that it might simply be placebo.
Another idea is that EMDR is taking advantage of quirks of the way our mind-emulations run on meat brains. Current theory seems to be that every time we access a memory, we also modify it. There may be enormous emotional content stored along with certain memories, which are frequently accessed, and in ways that maintain or intensify the emotion. By carefully and slowly accessing those memories in a calm and safe situation, repeatedly, we can lower the emotional valence of those memories until they're merely normal. On this view, the eye movement might serve as a harmless distraction, a way to keep our conscious attention occupied while the rest of our brain re-lives the experience.
Or perhaps the whole thing is slowly retraining the amygdala to relax and not freak out whenever the memories are encountered. I've noticed that I've completely lost normal adrenaline responses to things like assault, severe bodily injury, rollercoasters, and that type of thing. However, something that reminds me of something that reminds me of the bad stuff can flip me out, even if I'm just walking in a park on a sunny day. So maybe part of what's going on is that I'm slowly retraining my brain so that the fight-flight-freeze thing doesn't kick in when it's not appropriate. (Yeah, yeah, "define 'appropriate'".) I do wonder whether the normal adrenaline response will come back, but that's not a problem for today.
One thing I should mention is that EMDR can have very strong effects on my general mood over the week or so following a session. The first time I did any serious work with EMDR, I had a high like the first time I tried amphetamines (5mg generic Adderall), and like the amphetamine high it faded quickly and was not repeated. That time I came out of the session feeling extremely fatigued, such that I went home and took a nap for 4 hours, then woke up feeling amazing. (This is odd, because I almost never nap, because I normally feel horrible and groggy for the rest of the day.) Sometimes I've come out of sessions feeling lousy for the next week. Other times, when we did lighter work, I've come out feeling happy. So be prepared. If you're working, you might want to schedule these for the end of the day, maybe even Friday afternoon if possible.
Also, I personally would not describe it as "trance", at least not the way either of the therapists I've seen have done it. At least, not any more than I'd call a normal PTSD flashback a "trance". But then, I've never been in any state I'd use the word "trance" to descrive. Other people's subjective experiences may differ, of course. :-)
EMDR is a form of trance that works really well for a single trauma or a few isolated ones, at least for a large fraction of people. It works a lot better with a good practitioner. Also, consider ART instead of EMDR (https://acceleratedresolutiontherapy.com/), it seems to work faster for many people, and, anecdotally, also works for CPTSD, where EMDR ends up pulling on too many traumatic threads to be really useful. Again, YMMV, if it feels like something is off after you tried it, best reevaluate the next steps.
Psychologist here. I have not had EMDR, but know many people who have. Most say it helped then. In my opinion, the active ingredient in EMDR is the remembering and recounting of the trauma the person does while being guided through the eye movements. I think the eye movements mostly provide a theory of why the process works, and a task to do to that helps the person endure the process of reliving trauma. However, so long as EMDRit works for trauma, it's fine to do. The only kind of EMDR I think is truly not a good idea is EMDR as treatment for negative events that are not traumas (traumas are sudden, overwhelming, dangerous and disastrous events that most people never even experience: seeing deaths in combat, being in a major fire or automibile accident, being raped, etc.). I do not think EMDR is an effective treatment for phobias, bad but not traumatic memories (funerals, break-ups), bad experiences that were not an incident but a bad circumstance that stretched out for a long period of time (being broke, hating yourself, suffering a long depression, having a job with an asshole boss).
My partner is currently doing EMDR as a patient. She emphasizes that it is different from and more intense than CBT, which is the other form of therapy with which she has the most experience. Where the latter is about interrupting an emotional process with an intellectual one, EMDR involves direct confrontation of trauma on an emotional level. It will make you feel worse before it makes you feel better, but if you can afford that then EMDR is very effective.
It's definitely not snake oil. It very effectively produced a dramatically altered state of conscious, when I tried it. It's a much more intense experience than I'm used to having in a therapists office, and it was comparable to other heavily mind altering experiences I've had. I get the impression set and setting are probably just as important here as they are with drugs, rituals, etc; I didn't particularly trust my therapist at the time so it was not a useful experience for me at all, but I don't think that's a problem with the treatment and I could absolutely see how it could be effective.
Worth trying.
A few times over the last few years, commenters have put "AI" and "Moloch" together (e.g. from a cursory search: Maxwell IV and Mr. Doolittle). I'm curious if others, seeing the rather impressive evolution of AI-safety discourse (including AI-not-kill-everyoneism, AI-not-make-everything-worseism) would like to revisit this comparison.
I.e. if you go back and read Ginsberg's original and Scott's commentary on it, do you now more than before get an "aww shit advances in machine learning is gonna accelerate this" feeling and if so do you have thoughts to share on it?
Or, of course, thoughts to the contrary?
Notably, if you reread the original Meditations on Moloch, you'll note that Scott stresses that Moloch hasn't won. Humans again and again defy the explicit incentives in order to preserve that which is good and beautiful in the world. Unfortunately, an artificial intelligence won't have a human soul at its core prompting it to do good and defy its explicit instructions, so we'd better get a lot better at crafting instructions that don't destroy all value. when optimised.
Well, you see people worrying "making the AI too powerful too soon might kill all humans" and the companies... not really saying out loud, but thinking "but if we don't make the AI as powerful as possible, as soon as possible, our competitors will take over our market share".
That's Moloch in action: increase the probability of human extinction by 1% as a sacrifice, get a temporary advantage against your economical competitors. Repeat, if necessary.
I have a somewhat specific ask: I’m going to be living in Washington DC for the next year or so, and during that time, I would love to meet people who are also interested in reforming US college tuition wholesale--especially if you’re already involved in something along those lines. I specify DC because it’s where I am and I assume it would involve something on the policy side, but am certainly open to talking to anyone interested in possible solutions, regardless of where you’re from. I’m especially interested in people with backgrounds in insurance (bonus points for health insurance), antitrust, or university administration. If you are interested, my email is harrisongfriedman@gmail.com.
Will you release the initial voting scores for all the entries (including unsuccessful) at some point?
I would also like some sort of feedback mechanism for non-finalists. Thanks
There's no way to remember everything in a book, but you can do things to remember more. The best way to consolidate things in memory is to use them. At the end of sections (chapters or whatever), put the book down & *do* something with the info. If it's material that lends itself to problems, make and solve some problems. Or draw a diagram of material that was presented just in words (draw it without looking back over the section, then afterwards look back to check its accuracy.) If it's, say, a history book, think over what you read and ask yourself questions like, "what thing, if it had been different, would have completely changed the impact of this event?" Make a few notes on the answer. Also try to have some conversations with people about the book.
Why are you trying to memorize the books? Is it just the information or is it the actual verbiage?
I haven't used spaced repetition for anything, but had moderate success with mental pictures. Does Anki let you draw pictures? Draw some pictures that evoke the sentence you're trying to remember.
Two suggestions of highly dubious efficacy:
1) Use Readwise. It' a paid subscription service to manage notes and highlights in whatever you read. Book highlights you take are automatically presented to you in an Anki-style review, and if you like the highlight, you can turn it into a proper Anki-style card à la Cloze Deletion or similar. Disadvantages include the monthly fee and that the whole review process is only minimally customizable, in contrast to Anki.
2) I've heard about a bunch of attempts to use GPT3 or GPT4 to auto-generate Anki cards. I don't know if any of them have come to fruition, or work well, or have decent user interfaces, etc.